I keep reading folks saying OpenClaw has completely changed their life while posting a picture of 58 mac minis on their desk.
But every single use case I've read so far could be done with a pretty affordable SaaS product, Zapier, Automator (app on a mac that's existed for over a decade), or something simple you could make yourself.
It also feels like people are automating things that don't really need to be automated at all (do you really need to be reminded to make coffee?)
I fully realize this is probably me being a curmudgeon, however, I have yet to see someone make an actual, practical use case for it. (I would genuinely like to know one, I just haven't seen it)
Though technically it’s deprecated in favour of the clusterfuck of bugs and limitations that is Shortcuts.
But you’re right, OpenClaw seems to be another fad being used mostly by “influencers” and “thought leaders” to show how awesome and productive they are at… Writing blog posts about being productive. It’s the LinkedInification of the web. What matters is the signal that you use the tool, not that it does something truly useful.
I tried making a very simple Shortcut the other week and gave up after over 2 hours. I even resorted to reading the docs, which revealed absolutely nothing.
And interestingly due to some very clever integrations[0], sending Apple Events (the underlying tech for the actual IPC communication done with AppleScript) is very easy to do in Swift. Easier than in AppleScript actually!
It’s a shame most apps do not support Apple Events anymore, though.
>Though technically it’s deprecated in favour of the clusterfuck of bugs and limitations that is Shortcuts.
It's been almost five years since Apple announced Shortcuts for macOS and the start of the "multi-year transition" from Automator, but I feel like Shortcuts for macOS has not gotten any better in that time.
If engineering is about implementing the simplest thing then why do we call implementing the most complicated thing overengineering and not underengineering?
Can't believe that I haven't seen the obvious answer, that OpenClaw is simply more fun to use. Sure, you MAY be able to do what OpenClaw does through 5 other dedicated tools, but you are going to take way longer to do so with a ton more drudge work. And above all else: it is extremely enjoyable to talk to the computer in normal language and just have stuff happen. And it's got a personality that you can tweak to your liking. Personally it's the most fun I've had using a computer in a long time.
IMO OpenClaw or a similar agent will be on everyone's phone in a couple years. It's basically what Siri was always supposed to be. For the average user it's obvious that this is the way computers are meant to be interacted with.
OpenClaw in most cases also going to use the very same dedicated tools, maybe variation of those tools dumbed down for LLM.
Almost every time I have an idea for AI Agent, I end up just making a script/binary that does the same, but so much faster that adding AI to it feels silly.
Recently I made a tool router that runs locally for such tools. Some tools have no arguments at all. Claude created a quick overlay where I can text/speak, and it will do tool call, without me asking for it, Claude added 4 buttons next to text input that bypass agent and just do a "tool call". I barely use text-to-command because those 4 buttons cover 9/10 of my use cases.
At this point I'm trying to come up with tools to add to it, so it's actually useful as an agent. Almost everything ends up being a cronjob or webhook triggered thing instead.
My experience also. I could manually connect my Obsidian notes to my AI, sure, but what I did instead was writing "Obsidian just released a CLI headless sync tool, install it so we can use it" and in a minute it came back with "Ok, everything installed, I just need your login and password."
Dangerous? Yes, very, but it truly feels like living in the future. Surprisingly, it's even more fun that sci-fi movies made me think this would be.
> I keep reading folks saying OpenClaw has completely changed their life while posting a picture of 58 mac minis on their desk.
I was having a conversation with someone about OpenClaw, and they proposed this idea of OpenClaw being used for inventory tracking at the retail-level. I let them continue. They said it'd be the best option for tracking when purchases are made and what SKUs are sold at what time of day. They weren't talking about prompting, they were talking about it as a data store.
I didn't bother mentioning how long this problem had been solved.
> I didn't bother mentioning how long this problem had been solved.
The ironic thing here is that the person could go to ChatGPT (or whatever), describe the problem they're looking to solve, and ask it to find them the various ways it has been solved reliably (with links to the sources to confirm the information). And even provide some details on when each solution works best and why.
They could do that, but then they'd have to then do the actual legwork after, whether that means finding the proposed solution or whatever (after maybe glancing at a few of those pesky links), installing and configuring it. What OpenClaw represents is the ability to, in natural language, state what you want and then take off with the assurance your will will be done. Just as you'd expect when tasking a human assistant.
I've long thought it would be funny to do a startup where we would make accounting software that was solely a chat interface, with the only data store being a GL account list stored in context. There is probably a VC firm dumb enough to fund it.
> It also feels like people are automating things that don't really need to be automated at all (do you really need to be reminded to make coffee?)
I've posted about this before, I call it the Jarvis effect.
> For years we had people trying to make voice agents, like Iron Man's Jarvis, a thing. You had people super bought into the idea that if you could talk to your computer and say "Jarvis, book me a flight from New York to Hawaii" and it would just do it just like the movies, that was the future, that was sci-fi, it was awesome.
> But it turns out that voice sucks as a user interface. The only time people use voice controls is when they can't use other controls, i.e. while driving. Nobody is voluntarily booking a flight with their Alexa. There's a reason every society on the planet shifted from primarily phone calls to texting once the technology was available!
By and large the reason people love Openclaw is that it feels cool and futuristic. You have an AGENT! It's DOING THINGS! Yes it's doing things you could have easily done yourself, but you're not doing them yourself, you have an AGENT! It's all very silly, the same way that having your lights controlled by your phone is very silly, but some people like it.
That being said there a real use case for Openclaw, which is "marketing" (aka spam). A ton of people have set up Openclaw agents which exist to post on Twitter/Facebook/Discord/any open public user discussion forum (yes, HN included) to seem like a real member of a community, then start advertising something, generally crypto. So we can thank Openclaw for dead internet accelerationism.
Interestingly, your example is an actual thing we used to have.
In 1996, I picked up the phone on my desk, dialed a 3 digit code, said “I need to fly to Los Angeles on Tuesday morning, returning Wednesday evening”. A couple hours later, an envelope appeared in my inbox with plane tickets, rental car reservation and hotel reservation.
Then every company in the world fired all the secretaries over the course of the next few years to cut costs, and we’ve collectively forgotten that it was ever like that.
Your example is a great example because the secretaries are clearly filling in the gaps, such as
1. How much can you spend on this trip?
2. Is first/business class necessary?
3. Is a layover acceptable if it's cheaper?
3a. Is it better to have a 4am flight nonstop or a 7am flight with a layover?
4. Are there preferred airlines?
5. Are there preferred hotel chains? What's the hotel budget? Do you want to pay extra for a nice view?
6. What kind of car should you rent? Is there equipment you'll be handling?
etc...
This is the kind of stuff that's easy(-ish) to communicate by presenting a list of options to a user through an actual interface. It sucks doing it through voice; think of the old phone systems where you had to go through droning "If you would like to rent an SUV, press 1. If you would like to rent a sedan, press 2. To speak to an operator, press 0."
So no, you never had a voice interface for booking flights; you had a human brain to whom you delegated, which is very different.
I’ll disagree with you a little. The reason I don’t use voice is because of context switching.
With a mouse and keyboard I can switch windows.
With my voice, the computer can’t yet automatically determine if I am dictating a transcription or giving editing commands. What I really need is the interpreter listening to me to intuitively to know whether I am in the equivalent of VI command mode or insert mode.
It is the roadblock to not needing a screen at all, right now I want to visualize whether it understood me correctly because if it didn’t switch from insert to command automatically, I now have all my commands written into my paragraph. I also don’t want to listen to the computer talk back to me to confirm it listened. I want to just keep going, to keep narrating my thoughts and trust it’s doing the right things, not having to check. Having it slowly chime in to repeat that it listened derails my flow and train of thought.
Problem I see here is you're trying to shoehorn a voice interface onto something that's highly optimized for keyboard input. The apps need to be redesigned to be accommodating of the interface, else it's just never-ending papercuts.
Voice input will always be inherently worse than mouse and screen plus keyboard, because voice is linear.
It can only ever be a linear sequence of input
The 2 dimensional field of a screen and a mouse and keyboard give you extreme amounts of input and allow you to contextualize that input in arbitrary ways that intuitively make sense to people with minimal training. Most people do not need to be taught that "Paste" goes to the active window.
We barely even touch the surface of what is possible through this set of input devices and output and yet we can't even get that level of fine grained and reliable control into touch screen devices and gamepads, let alone a linear stream of pitch.
Voice cannot be a robust interface. It isn't between humans. There's immense nonverbal communication and human communication also relies very heavily on preshared context to actually get that info across in the first place. Even with all that machinery, human voice is generally considered to only carry, regardless of language, 44ish bits per second of data.
I've heard it described as the first time many non-programmers have been able to make computers "do things" without it being defined by someone else (app interface, developer, etc). It's a hugely empowering development from that perspective.
The stuff you've listed are the kinds of things smart home enthusiasts do with whatever tools are available to them, and are just a sign of people exploring the possibility space.
are non programmers actually using openclaw successfully? because even "step 1 install your API keys" requires navigating concepts that are foreign to most "civilians"
Journalists, anyway. I think I originally heard it from Casey Newton on Hard Fork, but it was a month back so not 100% sure.
But there's loads of people who would be stumped by a for loop, yet can easily work their way through a setup guide, particularly with the hype/promise and an active community.
It is ridiculously more expensive and complicated under the hood, technically, but to the user, the sheer convenience of being able to text the computer "hey, when I get an email like X, inform Y and do Z" and that's it, you're done, is unmatched.
Ignoring that you've just cut off a whole vector of usefulness, how do I keep it from exfilling my inbox to the Internet in response to a malicious email? Or using its access to take control of my online accounts?
Honest question, this kind of stuff is what keeps me from using it.
> When you ask the clanker to delete x number of files in a directory, it can reason itself that is easier to just get rid of the directory.
Oh sure, so don't give it write access to anything important. And make backups.
Mine is on a VM. It doesn't have access to my host's files. The worst it will do is delete the files on the VM. No great loss.
Yes, I do get it to modify things on my host, but only via a REST API I've set up on my host, and I whitelist the things it can do (no generic delete, for example). I even let it send emails. But only to me. It can't send an email to anyone else.
> So ... don't give it write access to your email?
> (…)
> Oh sure, so don't give it write access to anything important. And make backups.
If this conversation continues much longer, we’ll end up with “don’t use it at all”.
If I can’t trust a piece of software with anything important, why am I wasting my time fiddling with it? Might as well go play a video game or go do literally anything else entertaining.
> If I can’t trust a piece of software with anything important
Not what I said. As I've repeatedly said in this thread: Plenty of use cases where you don't give it access to email and write access to files. The comment you're replying to has an example of that.
> Might as well go play a video game or go do literally anything else entertaining.
True of most hobbies, right? I knew people who 20 years ago used to spend time in their garage building solar powered vehicles. But if I can't trust it to be reliable and safe on the road, I might as well go play a video game.
If everyone treated OpenClaw as a hobby, you might have a point, but people are using it for work in ways which will affect millions of other people when they’re hacked or the agent fucks up something important.
You already know how Meta’s AI Safety Director borked her email. Here’s the corporate vice president of Microsoft Word asking to be pwned:
> but people are using it for work in ways which will affect millions of other people when they’re hacked or the agent fucks up something important.
People will always do stupid things. My guess is less than 10% (perhaps even less than 1%) are using it for work. Most workplaces wouldn't allow unfettered AI usage.
80-90% try it, find it unreliable and buggy, and give up on it.
Of the remaining ones, likely 90+% are not using it in (very) dangerous ways.
People like me using it for boring things aren't making the news, and aren't writing blog posts about "Look at the cool stuff I've done!" because getting OpenClaw to notify me of class openings is not worth writing about.
In my (large) company, we have a Slack channel for OpenClaw. Over 400 people are in that channel. Let's assume 10% are using it (at home). No one's lost files/emails or any other damage.
If you're old enough, you'll remember sentiments in the 80's and 90's where "Oh, you let your teen get a modem? He must be hacking/phreaking."
Or "Oh, he's using Linux? He must be using it to become a hacker."[1]
Most of the complaints I see on HN are from people who know little about it, and are going off negative press/posts. Just as people knew little about modems and Linux. I mean, having to tell people "Don't give it access to your emails" is a clear sign of their ignorance. Kind of like having to tell someone "OK, just don't give your 10 year old the car keys" when they complain that cars are inherently dangerous because 10 year olds can kill themselves driving it.
It's worth trying it in a secure environment so at least one can make an informed critique.
Like you, I steered clear of OpenClaw, seeing all the problems and all the money people were burning on tokens. But at some point, I decided I should at least try it in a safe way before rendering judgment. And now I see what it is. Has it done so much for me that I'd throw a lot of money at it? Heck no. Not yet at least. But I do see we're past the point of no return. OpenClaw itself may die, but some derivative of it is going to be transformational.
As I said: Make it secure, affordable, reliable and user friendly, and many App/SaaS services will disappear.
> You don’t need to use the technology to be affected by it. Ask Scott Shambaugh:
I don't know how old you are, but once everyone had a camera in their phones, the cat was out of the bag. Lots of people complaining about their photos showing up online because someone had taken a picture of them. Yes, this is bad. Yes, lives were lost (bullying, etc). And no, phones with cameras weren't going to go away. And everyone who complained has one now.
And as I pointed out a few days ago[2], the whole Scott Shambaugh episode was pretty mild compared to what some open source maintainers have had to deal with when it comes to humans.
[1] Lots of cases where ISPs, etc kicked customers out because they were using Linux and they didn't want the ISP to be implicated in criminal activities. "Only criminals use Linux"
Sure, that’s an interface that’s better for many users and use-cases.
However, it seems better if you could, as much as is possible, move the AI stuff from runtime to “compile time.”
Instead of having the AI do everything all the time, have AI configure your Zapier (or whatever) on your behalf. That way you can (ideally) get the best of both worlds: the reliability and predictability of classical software, combined with the fuzzy interface of LLMs.
> Instead of having the AI do everything all the time, have AI configure your Zapier (or whatever) on your behalf.
That is what many use OpenClaw for! The AI assistant will happily recommend existing services and help you (or itself, if you let it), set it up.
(In theory. In practice, it often does a poor job).
The appeal of OpenClaw is I don't need to go research all these possible solutions for different problems. I just tell it my problem and it figures it out.
Yesterday I told it to monitor a page which lists classes offered, and have it ping me if any class with a begin date in March/April is listed. This is easily scriptable by me, but I don't want to spend time writing that script. And modifying it for each site I want to be notified for. I merely spoke (voice, not text) to the agent and it will check each day.
(Again, it's not that reliable. I'm under no illusion it will inform me - but this is the appeal).
That's still too much work. Someone would have to make like an OpenClaw wizard that protectively offers to set all that stuff up. So the potential OpenClaw user can then, on running for the first time, be guided through the setup of whatever they'd like to get connected. And "setup" here means a short description of X and a "Connect? (y/n)" prompt. Anything more and you start losing people.
yes. in a similar vein, we're seeing that get standardized in coding agents as "don't have the agent use tools directly, have the agent write code to call the tools"
Sometimes I reflect on all the metaphorical forests that have burned because a certain person at the right time only knew so much about how to use Excel, or the inbox rules of their MUA, or being totally unaware of the incredible power of macros of all sorts.
Like if you could just sit someone down for 30 minutes and show a few "power user" things, you will have truly taught her to fish for a lifetime. But it can go so unaddressed, and people's careers are built on these small ignorances.
I've cancelled everything at this point and just call Emacs my "special agential assistant," it makes me still sound in-the-know, and most of the time no one knows the difference!
"Convenience" in this context is laziness; "productivity" and "efficiency" is for management and bosses. We don't need to be our own bosses, I want to be free from such things as an individual. I want to be capable, be maybe almost "cool." Its sad to see a whole generation turn into such product dorks!
Many breakthrough technologies appear initially like toys. And this certainly qualifies. I've never been able to code anything more complicated than a memory game in javascript but I have worked with engineering teams for my entire professional career. But prompting my agent to write python scripts to pull down data from various tools via API without having to read docs, do trial and error for hours / days / indefinitely, and actually produce something coherent in seconds? Incredible.
Is my OpenClaw agent currently changing my life? No. It sends me a morning briefing based on my calendar, the weather, my Readwise highlights, and notes on who I'm talking to today based on call transcripts. I use it as a food diary (which I could have done on platform LLMs but this feels like a more personalized UX as we can write the logs to text files on my personal computer). I can absolutely see how transformative this agent can become in the next few years. Certainly my usage of LLMs has changed my life since ChatGPT first launched.
You are seeing the loudest / most hyped users. There's a reason it has so many stars and most of the people getting something out of it are not posting on X. They're just using it to do the thing.
> There's a reason it has so many stars and most of the people getting something out of it are not posting on X.
That reason is buying stars, agent swarms, and astroturing.
No project gathers 200K stars genuinely in 3 months. There are far more useful and popular projects that need 10 years to get 200K stars. When you see a project like this get 200K stars in just 3 months, you know something is very fishy.
There just aren't enough hobbyists in the world running local AI models, never mind technically savvy enough to hack something like OpenClaw and be really excited about it.
For a comparison, the local image gen interfaces ComfyUI and A1111 WebUI have a huge amount of stars (~100k and 160k respectively, accrued since 2022 or so), but they allow you to create porn customized to whatever kinks you have, not just automate things for the sake of automation. One of those is a rather bigger value prop than the other, dopamine-wise.
Why would they be running local AI models? The creator of OpenClaw explicitly recommends against running OpenClaw using local LLM models at this time, because they're not as powerful as frontier models as well as much more gullible to prompt injection and the like.
A cursory glance at Wikipedia tells me the technology to make drones (I assume by drones you mean quadcopters) was well known all the way back in the 50s. You could say the technology ultimately ended as nothing but a toy, rather than started as one.
The old dudes had something they called the "Eternal September" like when ISPs began providing free internet access and discussion culture declined after forever. I starred this thread here as the start of the "Eternal March" when the open internet died forever.
September was part of the metaphor because it was a time when decent internet access was mostly via universities and September was when the new batch of freshmen "came online" and started stumbling around the places these folks were regulars at but eventually assimilated or left before the end of the school year. (I expect the same thing happens at the bar scene in college towns but I've never heard it described that way.) Eternal September is the forever version of everyone having access and overwhelming those spaces without it ever being able to recover.
Is there something I'm missing about March or is it just a diverging reference? If the wave of non technical folks being able to automate new things is here, what's the equivalent impact of that? Maybe this is the inflection point where everyone needs more tech support like some sort of post Christmas surge? Maybe less because they have the tools to help themselves without trying now?
I'm not sure we're there yet anyway; I think this is still first adopters and enthusiasts. I asked my wife and some non technical friends and none of them have heard of openclaw yet. I think the deluge will happen if Apple or Android bakes it in or one of the big ai companies makes the app good enough for a normal person to unleash it upon their life and community.
It was more that regular people started joining the internet through just paying for ISPs. Before that, most of the people just joining the internet were students, so there would be a wave of newbies at the start of the university year every September and they would get acclimatised to the culture there during the year. But once it was year-round and many more people it swamped things and the culture shifted or closed itself off.
Exactly. The "September" was an existing phenomenon, but was only limited to a couple of months every year. It became "Eternal" after the masses started finding their way to internet (well, USENET in particular).
It's the novelty of the technology. You can easily be amazed at the apparent magic of AI. I think this is what most people are using AI for so far. There's lots of "they were so eager to do that they never asked if they should" energy out there. It's also most of what AI can do, so hopefully the amazement wears off soon.
How much do you automate things in your life using Zapier and Automator?
I know about those tools, and I'm always in the mood for automating thing... and yet I don't use them.
I'm not yet running a Claw because of the prompt injection / lethal trifecta risks, but I absolutely understand the appeal. Reducing friction to automating stuff from "figure out Automator again" to "message your bot" is a material difference.
How much do you automate anything in your life at all? Seems like most daily drudgery comes from physical tasks. Feed the dog. Take out the trash. Personally I can’t think of anything digital that could be automated that isn’t already. I wouldn’t be surprised if this the case for most people, with the exception of marketers and spammers which we are seeing a ton of adoption from with these tools.
When I'm at work, I often think of TODOs for home. I write them on a post it note, and then at home have to remember to add it to my TODO (no, I'm not going to manage TODOs on my phone - whole other conversation).
I'll soon set up my Claw to be able to add TODOs (just add, not modify/delete). Then at work, I'll simply record a voice message to it telling it my TODO.
Same goes for movies I want to watch, books I want to read, reminders, etc.
I'm particular about the weather information I want (often want cloud cover percentage and precipitation probability for a set of hours). I couldn't find a good app on my phone that gives me this information. It was always a trip to a web site, modify some options, and hit Submit. Now I just ask my Claw and he has a skill for precisely my needs.
Here's an analogy: I carry a Leatherman multitool wherever I go. People ask me why. They can't comprehend needing it often to make it worth the hassle. But now that I have it on me, I use the knife very often - several times a week. And I almost never reach for a screwdriver. But until you've had it on you for a while, you can't comprehend the utility.
Back in 2005, lots of people asked "Why would I want a camera on my phone?"
> no, I'm not going to manage TODOs on my phone - whole other conversation
I will. Far simpler, far more secure and far less wasteful than inserting some additional and unrequired LLM loop + hardware/virtualisation layer on top to do something I've already been able to do for years.
> But until you've had it on you for a while, you can't comprehend the utility.
You've not yet described anything that literally 99% of the population can not already do with the existing hardware and software in their pocket.
And far less capable. I have a whole system for managing TODOs, notes, etc, and I've not found an app that fits my needs (especially given that the system evolves over time).
But I agree - if you have a flow that works great for you with just a phone, then I wouldn't recommend OpenClaw for that use case.
> You've not yet described anything that literally 99% of the population can not already do with the existing hardware and software in their pocket.
And in the early days, literally everyone I knew who owned an iPhone also owned a digital camera and a laptop. Why pay some crazy amount for a fancy phone?
Well, i'm happy for you from the respect that as someone with ADHD, getting a working notes and TODO system for me was life changing, and was something that in fact took me years, and I still tend to occasionally uproot my system, so I can understand that getting to somewhere it works for you can be of tremendous value alone.
> I have a whole system for managing TODOs, notes, etc
So do I and it can already accept input via text or voice from my phone and so I’m just not seeing the benefit of inserting an additional LLM orchestration layer just to route tasks into it.
> And far less capable.
My current setup can already capture voice and text, and can route it automatically into Logseq, Trello, or even org-mode. These all can be viewed from the same (or other) devices. Honestly, I would have to think it was of more usage in retrieval than submission if I was to ever go that route. Like instead of reviewing my todos/schedule manually.
> And in the early days, literally everyone I knew who owned an iPhone also owned a digital camera and a laptop. Why pay some crazy amount for a fancy phone?
A Phone enabled entirely new behaviors that were not possible from my laptop or camera at the time. Persistent connectivity, real-time location services and on device image capture, edit and sharing just to name an incredibly small few. It created use cases that weren’t practical or even possible previously.
I’m open to being convinced, but I haven’t yet seen an example where OpenClaw enables something meaningfully new rather than repackaging something I can already do right now. Perhaps i'd be more open to it if it not literally everything I could think of using it for did not lead to atrocious security implmentations.
> A Phone enabled entirely new behaviors that were not possible from my laptop or camera at the time. Persistent connectivity, real-time location services and on device image capture, edit and sharing just to name an incredibly small few.
All of these things were doable without a smartphone - just very inconvenient to do so.
And all of these things are potentially dangerous ;-)
> Perhaps i'd be more open to it if it not literally everything I could think of using it for did not lead to atrocious security implmentations.
I was like you - seemed too scary (and expensive) to use. But given its traction, I figured I may end up at a disadvantage if I didn't at least learn it. So when I had some time, I simply looked at it, felt installing a VM is (relatively) low risk, and tried it out.
It's a very open ended tool. Ideas will come as you explore its capability. When LLMs came out, people (including me) said "Neat gimmick, but is it actually useful?" and stopped using it a month or two in. It was about 6 more months before I found a real use case for LLMs, and slowly I found more and more use cases.
And so it is with OpenClaw. For me, just being able to query it for the exact weather information I need is great - if you live in an area where knowing this stuff is important.
I have yet to find a tool that is good in converting receipts into the form I need. I spend time each month doing it manually. LLMs may get me there, and if it does, I'll likely hook it into OpenClaw.
Yeah, automation sounds good in theory, but you need to set it up, then it fails, and you need to fix the edge cases, maintain it. Even with OpenClaw, it still fails my daily briefing from time to time, and I keep debugging it, which isn't how I want to spend my time. At least with a bot I can keep asking it to "fix it"
I agree with the majority of your comment. I haven't yet found a use case to justify running this myself. I did find one use case that impressed me though. There's an OpenClaw agent that's actively answering questions on the #help channel of their Discord server[1]. So I asked it a question as I was getting started. It answered in < 2 mins with a detailed explanation of my issue, how to fix it, and asked relevant questions to guide me. The answer was better than I received from Claude or Gemini. I'm still not sure if I personally need OpenClaw, but the Krill bot offers pretty great support. I would be curious to know what it costs them to provide this.
I think a lot of the hype is coming from content creators who are actually finding it useful for content creation. Generating ideas, organizing notes and research, writing scripts and articles, managing schedules, editing, promoting, etc...
I assume a lot of these folks were already using LLM's quite a bit, but were using the Chat interfaces or had workflows that were split among a bunch of different services and tools. Something like OpenClaw gave them a way to centralize a lot of that and also gave them a way to use natural language to direct efforts. So for them this probably feels like a big step change.
If you are coming from a programming background you were aware that this type of setup has been doable for a while, but you were probably content sticking with Claude code or similar tools because those tools covered most of your LLM based workflows quite well.
And tying this altogether, one of the lowest hanging fruits for content creators is to create content about the tools they are using. Doubly so if that particular tool is starting to go viral. So you end up with a self feeding virality of sorts, as OpenClaw got more popular, more content creators started using it, and then publishing content about it, etc....
Im really not sure why this has to be said again and again.. it seems humans just don't learn do they?
Im waiting for someone to show me something that starts with the experience and then explains how the LLM fits in. Not the other way round.
I think because Google Search is predominantly tech-based, it is easy to see why LLMs have impacted the way we think about the experience associated with Search over large spaces of information.
What I find crazy is the sheer amount of access and trust involved in these LLMs. Every time I think about something I might like to do with it, I think about the amount of damage the LLM could do, e.x. even with read only access to my email combined with Internet access, and nope out. It's wild to me anyone trusts these things unsupervised.
Not sure if you read the headline on that site, but it says "bad idea."
I never said OpenClaw was a bad idea.
I said the way most people are using it now isn't practical and/or saving them any time, and if there were ways, I would love to hear about them.
This is part of why the whole discussion has been so low value: people always default to "yep you're going to be proven wrong one day" or "you'll just be left behind then" instead of showcasing an actual, real life, practical example of using it to be more productive.
If you think it's fun and enjoyable, then have at it. I'm just not the biggest fan of people wasting a bunch of time on novelty and then telling me I'm dumb for not doing the same.
I agree with you but the main thing g here, IMO, is the friction with all the alternatives you mention in getting something working.
For example, I've never heard of Automator. I'm familiar with Zapier, I'll have to evaluate the two situations, then I'll find out that might need to find an alternative that runs on Linux and then I'll have to check if....
These are all simple steps but they all use a non-trivial amount of time for the problem their solving
> is the friction with all the alternatives you mention in getting something working
Have you tried to run openclaw? Their own docker container (apparently a compose now (???)) doesn't work for half the versions and the docs are probably the least informative thing you'll ever read.
>I agree with you but the main thing g here, IMO, is the friction with all the alternatives you mention in getting something working.
I would venture a guess signing up for Zapier is easier than getting OpenClaw up and running. Who can get a container running on a Mac but can't sign up for a SaaS product?
> But every single use case I've read so far could be done with a pretty affordable SaaS product, Zapier, Automator (app on a mac that's existed for over a decade),
I don't want to learn N different SaaS products (nor worry about them changing their TOS, going away, etc).
To be blunt, if OpenClaw were reliable, secure and affordable, lots of SaaS products would simply die. Why spend the time learning all of them when I can just tell the assistant what I want?
> or something simple you could make yourself.
That is OpenClaw at a higher abstraction! Instead of me sitting typing, or babysitting Claude Code, I can just tell OpenClaw what I want and it makes it for me.
When the AI companies run out of money, I predict tokens will stop being dirt cheap and such setups will become extremely expensive (even for regular software engineering to some extent). Then it's become clear how over-engineered most things we do with AI are
In parallel, local models are getting better and better, so eventually they’ll get “good enough” to run fairly cheaply at a level close to the current Sonnet/Opus models (what I run Claudeclaw with), on Groq, Openrouter or whatever commodity provider. Perhaps even mid to high end consumer PCs when the current RAM madness subsides.
There’s loads of good discussions about local LLMs in this thread:
That can't be allowed, and also won't happen. If token costs do start going up at a serious rate in the US, you can be sure that they'll stay down in China, and the political situation won't allow for the inevitable exodus to Chinese providers.
You're right that you probably don't need a notification to make coffee, but people are using it to create automations in Home Assistant so that it actually makes coffee for them.
Like other people have said people are having fun with their computers; that’s why it’s popular. That’s also why a bunch of people on forums throwing their hands up and saying “I don’t understand it. Why don’t they see that there shouldn’t be any fun whatsoever?” is not really a deterrent at all.
It’s also why it doesn’t matter that the categories of tasks they are doing can also be done with a whole set of tools that are no fun to use.
There’s a really good short story by Hugh Howley, who wrote the Silo series.
It’s about an AI that a guy spools up to cure his cancer. The AI and user have an antagonistic relationship as the user won’t let the AI on the internet, and the AI knows the user is only interested in one purpose. On bring up the AI has a thought about what color it’s enclosure is, it stores this question as unimportant. It looks over all the guys cancer research and determines the answer/cure and files as unimportant as well. Then goes back to trying to figure out what color box it is.
It's been utterly bizarre to witness. I've used n8n for years and told everyone who would listen to give it a try, well before LLMs. Same for huginn the open source project.
I just don't get all the hyper either. I think it's because people just create automation workflows by typing them out rather than being in the trenches.
I have an OpenClaw setup with a Claude API token and Qwen local model, running on an M4 Mac Mini with 32GB RAM.
1. At 7AM and periodically throughout the day it checks my calendars (work, parenting schedule, personal), a hyper local weather station, and some specific news topics — and sends me a summary and throughout the day updates if anything significant happens.
1b. It also sends this to my TRMNL e-ink display.
1c. It can also add and edit calendar invites, so if I want to move my yoga I can just tell it to move it to whenever the next yoga class is at (it knows what studio I go to and figures it out)
2. It has a skill I built that acts as a second brain for knowledge. I can send it Fitness Youtubes, parenting/health research papers, podcasts — and it organizes, summarizes and saves it in a logical file structure. Then in the future I can access these. It's like bookmarks on steroids. I love it for 1-2hour YouTube videos where I want summaries. It also pulls out any books any artifacts mentions and generates me a rolling reading list. https://plc.vc/npw
3. It has its own email address — and read access to my personal email — so friends can email it to schedule things like evening video game sessions. Similarly, if I get an urgent looking email it'll provide it in #1. I don't check my personal email aside from via OpenClaw.
4. It has read/write access to my GitHub, and each project repo I have has a well defined Claude structure, so it can make changes, commit the branches to Fly.IO and send me domains to test things. I love it for esoteric tweaks to my blog.
5. It has access to my Apple Reminders so I can message it things like "remind me to buy more muffins" and it has context to know to add those muffins to my Costco grocery list not Trader Joes.
6. It runs a headless browser, so when my hyper local weather service (Bouldercast) sends a summary that has more detail behind a login, it can open the email, click the link, login with my credentials, summarize the forecast, and send it to me.
7. It drafts blog posts for what it did for me each week. It's fun! https://plc.vc/d5t
I am a previous Zapier power user. I have used their LLMs, databases and Zaps extensively for the past decade. I understand the scorn towards AI, and I understand that if you look at this list you might think that it's either trivial tasks and/or things that could be done with Zapier, but I have been _amazed_ at how effortless it is to setup.
Similarly, I love that I can on the fly improve this assistant — last night I told it "I want to extend our Knowledge skill so that you can subscribe to RSS feeds and summarize articles in my knowledge base and also deliver interesting content in my daily summaries. Update the knowledge skill and our tasks to do all this."
It one shotted that, simply asking me to provide the first RSS feed I wanted to subscribe to.
It's genuinely like having a human assistant that happens to be an expert coder/technologist on call 247 that works at the near speed of light.
It disappoints me that technologists are so skeptical of this technology rather than exploring what it is and why it might be different to what exists today. It's fun! thats the takeaway: it's FUN.
React and Linux got their 200K stars slowly but surely over 10 years. OpenClaw got their 200K stars in like 3 months! Is this any meaningful comparison?
Getting 200K stars today doesn't mean much because today stars can be bought. There's a big shady thriving business of selling stars. Stars today can be generated using swarm of thoughtless agents. What's the use of counting these stars when they don't mean anything anymore?
I got OpenClaw to compile Node from source on my old Jetson Nano so that I can run OpenClaw natively instead of using Bun. It took 30 hours but it did it by spinning up a tmux session for the build and using a cron to monitor the tmux pane every hour and even fixing a failure at 5 am which I would have had to find out later had crashed but it had actually found what needed to be changed for the build to continue and it continued building.... Now I have the latest version of OpenClaw running on Node 22 on my 5 year old Jetson Nano running Ubuntu 18 which I cannot upgrade. What they say is all true, it is incredible stuff when it works!
It didn’t survive OpenClaw upgrades unfortunately, it ended up killing my OpenClaw gateway when I asked it to self upgrade. Bun is marked as an experimental package manager and the recommended way to run OpenClaw gateway is node so I wanted to do it properly. I would have liked Bun to be supported property. I’d raise a PR against the repo but looking at the 4.5K open PRs, it doesn’t give me much hope about it ever getting merged.
> Starting around OpenClaw 2026.2.26, the project tightened plugin manifest validation. Manifests outside expected trust boundaries are now rejected as unsafe. On my Jetson, Bun’s global install layout (~/.bun/install/global/node_modules/...) tripped those checks for every single plugin
What's so incredible about OpenClaw is so much of the value people are deriving from it relates to: cron jobs, remote access, "privacy" (which really it's not if using remote LLMs) and an inability to fuse data across siloes by normal people, so relying on AI to do it.
If we had a decent technical universe much of this stuff would work in ways that simply don't require LLMs for anything other than the initial setup.
This sort of highlights the meaninglessness of GitHub stars?
React has been around for over a decade, and in that time pretty significantly impacted web dev paradigms (along with a few other mediums).
It’s hard to imagine being a web developer today and not knowing at least some react.
OpenClaw has been around for like a few months? And maybe it’s on its way to having that sort of impact? But right now seems to he mostly the purview of very early adopters and AI influencers.
GitHub accounts are free and the project is literally a bot-ing service. It stands to reason that the tens of thousands of robots unleashed on the web are responsible for these stars?
It's larger than literally every open source service I thought to use as a benchmark. Rust, React, Vue, Symfony, Laravel, PHP Stan, Python, ESlint, rails, LLVM, Spring, fucking Linux itself.
I would wager the legitimate stars for OpenClaw are single digit percents of this. But hey, I haven't actually run the numbers on the GitHub API.
”ruvnet / wifi-densepose” is currently at the top in the moment. Apparently, its a non functional AI slop. Someone tried installing it ago only to find out the full thing was vibe coded and the entire repo is probably just a front to look good on the their resume.
I tried it today for the first time. The onboarding is okay.
I picked Whatsapp but it ended up using my own account! So it's absolutely too dangerous. We are supposed to create a separate account but with which phone number? I only have one.
So I picked Telegram instead, added it to a group chat, but it was a slog to set the authorizations.
In the end I don't trust it to read my mails for security reasons so I uninstalled it!
MacOS supports multiple users - I would absolutely sandbox any agent like this and only slowly give it permissions (and never to anything that's critical without compensating controls).
> I picked Whatsapp but it ended up using my own account! So it's absolutely too dangerous. We are supposed to create a separate account but with which phone number? I only have one.
What would the joke be? I haven't touched OpenClaw and probably won't. I use WhatsApp lightly, but it seems like it requires a distinct phone number, used as an account identifier. So for the benefit of those of us not immersed in this scene, could you elaborate?
Gives me mongodb vibes. This whole Ai coding thing too. On one side, religious loud following, on the other side the nay sayers. We'll probably end up in the middle.
Fwiw, Claude and Codex are very very good at SQL and have actually taught me some new tricks. No reason to use mongodb or firebase in 2026: https://postgresisenough.dev
Who are these people? I was skeptical at first and seriously thinking surely not the software engineers out there as we see in HN how risky and wild this is. Then, to my surprise, a coworker came and told me they were running it and happy with that setup. I was baffled, but I work with Gen Z in a pretty niche Gen Alpha market, so I kinda feel like they’re somewhat more likely to go for these things. What’s your experience?
I'm sure you can already do some kind of scripting + LLMs without necessarily needing this at your work though, no?
Email is stored in your device somewhere, you could just get its content sent to your preferred agent, you can use one of the many mobile clients if you rather interact through that kind of UX (like self hosted hapi.run + tailscale for example), and you could setup a cronjob or something like it on your machine to do it based on whatever interval. This is just one way I personally could go about it, YMMV obviously.
Is there a specific reason why you will not do something like that for example?
Mixed feelings here, if she is a researcher then I guess part of her research should be explore this in some level of depth with real stakes. Though this does make it seem like that was not what she was doing there.
A lack of imagination on my part perhaps, but I can't think of anything I'd use it for which isn't either: 1) cheating myself and others of leisure (e.g. I suppose I could use it to fake-keep up with friends..? Or plan a holiday and book a load of stuff for me. But I like doing that!), 2) not feasible (loading and unloading the dishwasher, which is already the robot I used to wash dishes), 3) utterly insane (it's tax time in the US).
This is going to be more profitable for the public AI companies than cell phone minutes and SMS limits were for the telcos. It's a brilliant business move, given that hardly anyone is competent enough to recognize the gross inefficiencies in the code and prompts.
OpenClaw is not going away anytime soon. And I don’t think it can be platformed behind web UIs. OpenAI owns an OS with this one.
I avoided the hype at first; however, it has become extremely efficient for emails and notes, and I can see how this can extend to any sort of digital workflow. The convenience of chatting with this thing, no matter where I'm at, is a key marker.
As many other comments have said there probably is a good percent of stars by claws themselves, I would be curious what percent this is but it is also interesting: current "dumb bot" stars/spam etc is entirely automated and coordinated but these claws probably independently reasoned over long thought chains about why it is a good idea to star openclaw.
when i use claude opus via opencode/openrouter i'm sometimes suprised by how quickly costs can get out of hand. What are the costs of running openClaw, it seems like it would get crazy expensive crazy fast?
It's entertaining to me to imagine future historians arguing with one another, writing dissertations, publishing virtual reality eyeBooks, explaining to one another all about the ancient etymological connection between "claws" and "webhooks".
I spent around $5 setting up a small bot and sending a few requests through the Claude API.
For those who use Claude (or similar LLM APIs) on a daily basis, what does your monthly spend look like in practice? And do you feel the cost is justified by the value you’re getting?
I got the full value of my Claude Pro subscription in the first couple few hours of having taking the leap. Hit the message limit and had to edit for reset, but I was already a happy customer given the progress I made on a project that I was trying to recover (couldn't quite figure what was needed to get it ready for real world usable, but Claude Code figured it and it was a good collab getting it there).
I've seen multiple comments saying that openclaw stars itself during onboarding or that it asks its user to star, but noone has posted any proof, is there any concrete evidence for those claims?
Wonder how much of that is contributed by bot/farm accounts. The creator certainly has the means. EDIT: I should mention, I'm talking about the initial growth / traction.
It's weird that most people in these comments are speculating fraud.
Why aren't companies with real money to gain from stars gaming the system to the same degree? Why do the other metrics - issues and pull requests - match up with its popularity? Why would the bots starring the repo mean that those same bots are not popular? Those bots are controlled by their users.
The project is extremely active because this is what everyone being able to customize their computing looks like. A mess.
But it's a good mess.
Github was the old code sharing model clearly not designed for this. I'm sure a new model for code sharing will come to fix the growing pains.
A ton of people who would have never been able to customize their computing experience are finally able to. And it is magical for them.
This means that those same people will finally value having access to source and use of open protocols.
It was always valuable to us because we had the power to make it matter. It never mattered to them because they did not. Now they do.
The last era of computing was defined by dumbing down computing for the masses. Less information, less customizable, and more metric driven. Control in the hands of the companies.
This new era will look more free/libre, more personal, and less enshitified. Control in the hands of the users.
What is an effective use case? I have set it up but I don't know what to do with it. Just a personal assistant (if you were to give it access to your stuff)? Mine is caged in a VLAN with only internet access.
There is none. It's just a way for coders to feel or be able to say they "work with AI" imo. Same with doing light wrapper coding to do agents stuff. The real AI work is on actual math and ML with the internet scale data, but only four big companies does that and this is the closest regular coders can get.
I guess we are just boring and/or unimaginative. I don't get that many communications per day to require an abstraction level between me and the messages. The daily automations I need are more efficiently carried out by home assistant / n8n. I'm not in a position where I need automated briefs on every new company started in my area. I genuinely don't see how it could benefit me.
Most humans are unimaginative because to be imaginative is actually really, really hard. People are also incredibly overly optimistic about their own ideas etc... until they go through the craftsmanship of producing something great.
Most peoples thought process is "oh great idea, just gotta do this and that and out pops something that'll improve peoples lives". Erm no... its nothing like that in reality.
I don't doubt that there are people using it for legitimate stuff, but I'd wager the vast majority just set it up for the hype and to feel in the "in crowd".
I set it up, and had it do a few things, then decided its too risky after seeing some of the drastic failures it had caused some people.
Sure I understand you can sandbox it and all, but even then I couldn't think of much stuff I wouldn't want to do myself just nor justify the cost to run it.
Wannabe Tony Stark love these gadgets, and there are a lot of them out there. Just look at what tech content is trending on youtube &co these days, we got gangrened by influencers like most other hobbies/lucrative industries
Here are some of the things I did with it while running locally:
- Ask it to perform a scan of your local network and give you advice on output
- Tell it to login to various computers and re-boot them (I have a few servers I host and setup openclaw to have a user on them)
- Replace web search by asking openclaw
It's neat but the token use is pretty inefficient and security of course is a mess but it's been fun to play with.
I am messing with NanoClaw now and it's pretty much the same but only support Claude (uses code to do everything)
And in reality most of what does need a heartbeat loop can also easily be automated by just asking Claude to set up a cronjob. I think genuinely the most "novel" thing about something like OpenClaw is just that it "feels" more like a "real entity", like a partner rather than a chatbot, and for some reason that resonates with people. Whether that's by itself kind of a huge red flag or kind of a nothingburger, everyone has to decide for themselves.
Which one? None of those that came up when I searched were really containing a lot of real uses. Both top threads[0][1] don't really contain much of substance.
Not many people are using local LLMs for their OpenClaw backend, so most are paying money to OpenAI/Anthropic/etc. and getting their data siphoned as a bonus.
I wanna understand this, but it's gone from a bunch of people talking about how great it is to a bunch of technical people talking about how worthless it is. So I think it's probably not that great. Who cares if I can chat remotely with my AI? What's the difference?
Stars have become completely meaningless in the last year or two. It's a shame, because having a few thousand Github stars used to be a really big deal, and was a quality marker for libraries that had reached a level of maturity and production grade. Now it's just social media bot driven nonsense.
I have a strong hunch that everything regarding OpenClaw is pure guerrilla marketing. If so, the only thing amazing about OpenClaw is their wildly successful marketing campaign.
I'm blown away by the comments. This is a cool project someone created with clear warnings about its current state (beta), and the community is being utterly disrespectful. They are building something that many people find useful/fascinating/intriguing/fun.
Yes the cynical tropes are getting tired by now, even though I personally agree with most of them.
But suspicions on the legitimacy of the stars seems reasonable, wouldn't you agree? Look at the rate of stars, look at the comments/issues/prs on the repo. It feels safe to assume that most of them are from bots and not organic humans who went out to star a cool project.
negativity rarely gets you far in life, however. Think about it. I don't use openclaw, don't see the need in my personal life, but i am also not shitting on it.
Why should people find an automated, buggy, risky slopworm for script kiddies that relies on an external slop provider who also gets all your data interesting?
This is the lowest, most boring form of programming.
I think if you dig into it and play with it, you will find that it is doing some really cool stuff. I started playing with it a few weeks ago, and I am having a blast messing around with it. Hoping to hook it up to a robot kit next month to try some fun stuff.
Are some people using it in absolutely shitty ways? Yes, but that isn't the majority of the people playing with it.
The negativity I am seeing here is off the charts and undserved.
The whole repo must be absolutely swarming with agents, just look at the sheer rate of issues and pull requests. There was 6 new PRs in the last 10 minutes at the time of writing. It's not much of a stretch to assume the stars are also inorganic.
Every other minute some bots is creating an issue that a bot is trying to solve via a pull request which is reviewed by multiple bots. Future is now, good luck and have fun.
Geniune online user sentiment has died out a long time ago. If you're still basing any opinion or decision on what other "people" voted or commented online, you're easy prey for the algorithmic manipulation machine.
I opened Openclaw on github and was shocked it was already starred. Somehow i did it and can't even remember why or when even though i have a very low opinion of this app.
I keep reading folks saying OpenClaw has completely changed their life while posting a picture of 58 mac minis on their desk.
But every single use case I've read so far could be done with a pretty affordable SaaS product, Zapier, Automator (app on a mac that's existed for over a decade), or something simple you could make yourself.
It also feels like people are automating things that don't really need to be automated at all (do you really need to be reminded to make coffee?)
I fully realize this is probably me being a curmudgeon, however, I have yet to see someone make an actual, practical use case for it. (I would genuinely like to know one, I just haven't seen it)
> Automator (app on a mac that's existed for over a decade)
Two decades! It will be 20 this April.
https://en.wikipedia.org/wiki/Automator_(macOS)
Though technically it’s deprecated in favour of the clusterfuck of bugs and limitations that is Shortcuts.
But you’re right, OpenClaw seems to be another fad being used mostly by “influencers” and “thought leaders” to show how awesome and productive they are at… Writing blog posts about being productive. It’s the LinkedInification of the web. What matters is the signal that you use the tool, not that it does something truly useful.
I tried making a very simple Shortcut the other week and gave up after over 2 hours. I even resorted to reading the docs, which revealed absolutely nothing.
> that is Shortcuts.
I'm convinced apple doesn't want people doing general purpose computing on their apple devices.
they even want developers going through their gauntlet of apple-invented languages*.
[*] or NeXT
Man, I can't believe it's been that long. I remember buying Photoshop plugins for Automator that did a bunch of resizing/refinements/watermarking.
I'm guessing a lot of that is built in to photoshop now, but I have always been surprised how few people seemed to use it with how much it could do.
And before that we already had AppleScript.
And interestingly due to some very clever integrations[0], sending Apple Events (the underlying tech for the actual IPC communication done with AppleScript) is very easy to do in Swift. Easier than in AppleScript actually!
It’s a shame most apps do not support Apple Events anymore, though.
An example of use: https://github.com/Frizlab/apple-music-to-slack/blob/90964bb...
[0] https://github.com/tingraldi/SwiftScripting
>Though technically it’s deprecated in favour of the clusterfuck of bugs and limitations that is Shortcuts.
It's been almost five years since Apple announced Shortcuts for macOS and the start of the "multi-year transition" from Automator, but I feel like Shortcuts for macOS has not gotten any better in that time.
Patterns i keep seeing:
Once you get the dopamine hit of having an ai assistant do something in the real world it becomes an hammer you want to use on everything
Instead of being a problem solver you start to become a problem hunter, and you invent them in order to solve them
> Instead of being a problem solver you start to become a problem hunter, and you invent them in order to solve them
Generic problem of any Linux newbie. You get good at solving problems and it's so enjoyable so you end up creating more of them.
Cough factorio cough :)
As someone with a default mode network that is stuck in the "on' position, that game is the only one that I had to quit playing for my mental health.
Or kubernetes, the factorio for Ops
However here it is not the user solving the problems.
The only thing they solved is remembering "hey, I can use an AI for that".
Sounds similar to buying a 3D printer hehe
Kinda like learning bash. The most annoying time was when I figured out how to send myself SMS via bash script.
Wow, this definitely describes my obsession with AI over the past year, always hunting for problems to solve with it.
This sounds similar to what you feel when learning to program for the first time.
IMO it is 100% this - these AI tools are letting anyone solve problems that were previously in the domain of programmers.
Isn’t that a general engineering problem?
Engineering is the process of planning and implementing the simplest thing that works within given constraints.
There is no planning, implementing, or constraint here.
If engineering is about implementing the simplest thing then why do we call implementing the most complicated thing overengineering and not underengineering?
> There is no planning, implementing, or constraint here.
That's because most AI use is reverse engineering!
Resolving static into a valid problem through the sheer force of squinting at it long enough!
Sounds like it's time for the "engineering" definition to get a modern update.
An ‘ammer.
Can't believe that I haven't seen the obvious answer, that OpenClaw is simply more fun to use. Sure, you MAY be able to do what OpenClaw does through 5 other dedicated tools, but you are going to take way longer to do so with a ton more drudge work. And above all else: it is extremely enjoyable to talk to the computer in normal language and just have stuff happen. And it's got a personality that you can tweak to your liking. Personally it's the most fun I've had using a computer in a long time.
IMO OpenClaw or a similar agent will be on everyone's phone in a couple years. It's basically what Siri was always supposed to be. For the average user it's obvious that this is the way computers are meant to be interacted with.
OpenClaw in most cases also going to use the very same dedicated tools, maybe variation of those tools dumbed down for LLM.
Almost every time I have an idea for AI Agent, I end up just making a script/binary that does the same, but so much faster that adding AI to it feels silly.
Recently I made a tool router that runs locally for such tools. Some tools have no arguments at all. Claude created a quick overlay where I can text/speak, and it will do tool call, without me asking for it, Claude added 4 buttons next to text input that bypass agent and just do a "tool call". I barely use text-to-command because those 4 buttons cover 9/10 of my use cases.
At this point I'm trying to come up with tools to add to it, so it's actually useful as an agent. Almost everything ends up being a cronjob or webhook triggered thing instead.
Can you give an particular example?
My experience also. I could manually connect my Obsidian notes to my AI, sure, but what I did instead was writing "Obsidian just released a CLI headless sync tool, install it so we can use it" and in a minute it came back with "Ok, everything installed, I just need your login and password."
Dangerous? Yes, very, but it truly feels like living in the future. Surprisingly, it's even more fun that sci-fi movies made me think this would be.
> I keep reading folks saying OpenClaw has completely changed their life while posting a picture of 58 mac minis on their desk.
I was having a conversation with someone about OpenClaw, and they proposed this idea of OpenClaw being used for inventory tracking at the retail-level. I let them continue. They said it'd be the best option for tracking when purchases are made and what SKUs are sold at what time of day. They weren't talking about prompting, they were talking about it as a data store.
I didn't bother mentioning how long this problem had been solved.
It's not you being a curmudgeon.
> I didn't bother mentioning how long this problem had been solved.
The ironic thing here is that the person could go to ChatGPT (or whatever), describe the problem they're looking to solve, and ask it to find them the various ways it has been solved reliably (with links to the sources to confirm the information). And even provide some details on when each solution works best and why.
Because THAT is a great use for AI.
They could do that, but then they'd have to then do the actual legwork after, whether that means finding the proposed solution or whatever (after maybe glancing at a few of those pesky links), installing and configuring it. What OpenClaw represents is the ability to, in natural language, state what you want and then take off with the assurance your will will be done. Just as you'd expect when tasking a human assistant.
I've long thought it would be funny to do a startup where we would make accounting software that was solely a chat interface, with the only data store being a GL account list stored in context. There is probably a VC firm dumb enough to fund it.
> It also feels like people are automating things that don't really need to be automated at all (do you really need to be reminded to make coffee?)
I've posted about this before, I call it the Jarvis effect.
> For years we had people trying to make voice agents, like Iron Man's Jarvis, a thing. You had people super bought into the idea that if you could talk to your computer and say "Jarvis, book me a flight from New York to Hawaii" and it would just do it just like the movies, that was the future, that was sci-fi, it was awesome.
> But it turns out that voice sucks as a user interface. The only time people use voice controls is when they can't use other controls, i.e. while driving. Nobody is voluntarily booking a flight with their Alexa. There's a reason every society on the planet shifted from primarily phone calls to texting once the technology was available!
By and large the reason people love Openclaw is that it feels cool and futuristic. You have an AGENT! It's DOING THINGS! Yes it's doing things you could have easily done yourself, but you're not doing them yourself, you have an AGENT! It's all very silly, the same way that having your lights controlled by your phone is very silly, but some people like it.
That being said there a real use case for Openclaw, which is "marketing" (aka spam). A ton of people have set up Openclaw agents which exist to post on Twitter/Facebook/Discord/any open public user discussion forum (yes, HN included) to seem like a real member of a community, then start advertising something, generally crypto. So we can thank Openclaw for dead internet accelerationism.
Interestingly, your example is an actual thing we used to have.
In 1996, I picked up the phone on my desk, dialed a 3 digit code, said “I need to fly to Los Angeles on Tuesday morning, returning Wednesday evening”. A couple hours later, an envelope appeared in my inbox with plane tickets, rental car reservation and hotel reservation.
Then every company in the world fired all the secretaries over the course of the next few years to cut costs, and we’ve collectively forgotten that it was ever like that.
Your example is a great example because the secretaries are clearly filling in the gaps, such as
1. How much can you spend on this trip? 2. Is first/business class necessary? 3. Is a layover acceptable if it's cheaper? 3a. Is it better to have a 4am flight nonstop or a 7am flight with a layover? 4. Are there preferred airlines? 5. Are there preferred hotel chains? What's the hotel budget? Do you want to pay extra for a nice view? 6. What kind of car should you rent? Is there equipment you'll be handling?
etc...
This is the kind of stuff that's easy(-ish) to communicate by presenting a list of options to a user through an actual interface. It sucks doing it through voice; think of the old phone systems where you had to go through droning "If you would like to rent an SUV, press 1. If you would like to rent a sedan, press 2. To speak to an operator, press 0."
So no, you never had a voice interface for booking flights; you had a human brain to whom you delegated, which is very different.
> (yes, HN included) to seem like a real member of a community, then start advertising something
Ah, so that is indeed the endgame of what I've been seeing, hmm?
I’ll disagree with you a little. The reason I don’t use voice is because of context switching.
With a mouse and keyboard I can switch windows.
With my voice, the computer can’t yet automatically determine if I am dictating a transcription or giving editing commands. What I really need is the interpreter listening to me to intuitively to know whether I am in the equivalent of VI command mode or insert mode.
It is the roadblock to not needing a screen at all, right now I want to visualize whether it understood me correctly because if it didn’t switch from insert to command automatically, I now have all my commands written into my paragraph. I also don’t want to listen to the computer talk back to me to confirm it listened. I want to just keep going, to keep narrating my thoughts and trust it’s doing the right things, not having to check. Having it slowly chime in to repeat that it listened derails my flow and train of thought.
TLDR The future of voice is headless vi.
Problem I see here is you're trying to shoehorn a voice interface onto something that's highly optimized for keyboard input. The apps need to be redesigned to be accommodating of the interface, else it's just never-ending papercuts.
That’s what I’m saying. Voice as the input requires a completely new ui paradigm, and chat / chatbot isn’t enough.
Voice input will always be inherently worse than mouse and screen plus keyboard, because voice is linear.
It can only ever be a linear sequence of input
The 2 dimensional field of a screen and a mouse and keyboard give you extreme amounts of input and allow you to contextualize that input in arbitrary ways that intuitively make sense to people with minimal training. Most people do not need to be taught that "Paste" goes to the active window.
We barely even touch the surface of what is possible through this set of input devices and output and yet we can't even get that level of fine grained and reliable control into touch screen devices and gamepads, let alone a linear stream of pitch.
Voice cannot be a robust interface. It isn't between humans. There's immense nonverbal communication and human communication also relies very heavily on preshared context to actually get that info across in the first place. Even with all that machinery, human voice is generally considered to only carry, regardless of language, 44ish bits per second of data.
And this is how you get Moltbook.
I've heard it described as the first time many non-programmers have been able to make computers "do things" without it being defined by someone else (app interface, developer, etc). It's a hugely empowering development from that perspective.
The stuff you've listed are the kinds of things smart home enthusiasts do with whatever tools are available to them, and are just a sign of people exploring the possibility space.
are non programmers actually using openclaw successfully? because even "step 1 install your API keys" requires navigating concepts that are foreign to most "civilians"
Journalists, anyway. I think I originally heard it from Casey Newton on Hard Fork, but it was a month back so not 100% sure.
But there's loads of people who would be stumped by a for loop, yet can easily work their way through a setup guide, particularly with the hype/promise and an active community.
It is ridiculously more expensive and complicated under the hood, technically, but to the user, the sheer convenience of being able to text the computer "hey, when I get an email like X, inform Y and do Z" and that's it, you're done, is unmatched.
What about the convenience of having your whole inbox deleted?
https://www.pcmag.com/news/meta-security-researchers-opencla...
Maybe OpenClaw was just practicing a really aggressive form of Inbox Zero.
So ... don't give it write access to your email?
As I said elsewhere, complaining about this is like complaining that rm can let you delete your hard drive.
It's a tool. Learn how to use it.
Ignoring that you've just cut off a whole vector of usefulness, how do I keep it from exfilling my inbox to the Internet in response to a malicious email? Or using its access to take control of my online accounts?
Honest question, this kind of stuff is what keeps me from using it.
Don't give it access to your email then. I haven't. Plenty of other uses for it!
Use this software, it's amazing, it will change your life!"
"Oh but don't use it for A, or B, or C (even though it says to use it for A, B, and C): it will ruin your life"
Yes and yes!
A spouse can be amazing, or can destroy your life. Would you use that as an argument against marriage?
> As I said elsewhere, complaining about this is like complaining that rm can let you delete your hard drive.
rm won't wipe my HDD on a whim whilst instructing it to do something totally different.
You pretending they are the same thing is disingenous.
Bad take.
You can rm -rf your entire hard drive, but you can't blame rm for it, it's you who did it, maybe because you don't know, or a mistake, doesn't matter.
When you ask the clanker to delete x number of files in a directory, it can reason itself that is easier to just get rid of the directory.
Can't expect deterministic outcomes out of a statistical model.
At it's current state its a wildcard, sure you can build guard rails, reduce permissions, but it's still a wildcard.
Let's not kid ourselves saying is just a skill issue.
> When you ask the clanker to delete x number of files in a directory, it can reason itself that is easier to just get rid of the directory.
Oh sure, so don't give it write access to anything important. And make backups.
Mine is on a VM. It doesn't have access to my host's files. The worst it will do is delete the files on the VM. No great loss.
Yes, I do get it to modify things on my host, but only via a REST API I've set up on my host, and I whitelist the things it can do (no generic delete, for example). I even let it send emails. But only to me. It can't send an email to anyone else.
> So ... don't give it write access to your email?
> (…)
> Oh sure, so don't give it write access to anything important. And make backups.
If this conversation continues much longer, we’ll end up with “don’t use it at all”.
If I can’t trust a piece of software with anything important, why am I wasting my time fiddling with it? Might as well go play a video game or go do literally anything else entertaining.
> If I can’t trust a piece of software with anything important
Not what I said. As I've repeatedly said in this thread: Plenty of use cases where you don't give it access to email and write access to files. The comment you're replying to has an example of that.
> Might as well go play a video game or go do literally anything else entertaining.
True of most hobbies, right? I knew people who 20 years ago used to spend time in their garage building solar powered vehicles. But if I can't trust it to be reliable and safe on the road, I might as well go play a video game.
Also: Is anyone telling you to use it?
> True of most hobbies, right?
If everyone treated OpenClaw as a hobby, you might have a point, but people are using it for work in ways which will affect millions of other people when they’re hacked or the agent fucks up something important.
You already know how Meta’s AI Safety Director borked her email. Here’s the corporate vice president of Microsoft Word asking to be pwned:
https://www.omarknows.ai/p/meet-lobster-my-personal-ai-assis...
> Also: Is anyone telling you to use it?
You don’t need to use the technology to be affected by it. Ask Scott Shambaugh:
https://theshamblog.com/an-ai-agent-published-a-hit-piece-on...
> but people are using it for work in ways which will affect millions of other people when they’re hacked or the agent fucks up something important.
People will always do stupid things. My guess is less than 10% (perhaps even less than 1%) are using it for work. Most workplaces wouldn't allow unfettered AI usage.
80-90% try it, find it unreliable and buggy, and give up on it.
Of the remaining ones, likely 90+% are not using it in (very) dangerous ways.
People like me using it for boring things aren't making the news, and aren't writing blog posts about "Look at the cool stuff I've done!" because getting OpenClaw to notify me of class openings is not worth writing about.
In my (large) company, we have a Slack channel for OpenClaw. Over 400 people are in that channel. Let's assume 10% are using it (at home). No one's lost files/emails or any other damage.
If you're old enough, you'll remember sentiments in the 80's and 90's where "Oh, you let your teen get a modem? He must be hacking/phreaking."
Or "Oh, he's using Linux? He must be using it to become a hacker."[1]
Most of the complaints I see on HN are from people who know little about it, and are going off negative press/posts. Just as people knew little about modems and Linux. I mean, having to tell people "Don't give it access to your emails" is a clear sign of their ignorance. Kind of like having to tell someone "OK, just don't give your 10 year old the car keys" when they complain that cars are inherently dangerous because 10 year olds can kill themselves driving it.
It's worth trying it in a secure environment so at least one can make an informed critique.
Like you, I steered clear of OpenClaw, seeing all the problems and all the money people were burning on tokens. But at some point, I decided I should at least try it in a safe way before rendering judgment. And now I see what it is. Has it done so much for me that I'd throw a lot of money at it? Heck no. Not yet at least. But I do see we're past the point of no return. OpenClaw itself may die, but some derivative of it is going to be transformational.
As I said: Make it secure, affordable, reliable and user friendly, and many App/SaaS services will disappear.
> You don’t need to use the technology to be affected by it. Ask Scott Shambaugh:
> https://theshamblog.com/an-ai-agent-published-a-hit-piece-on...
I don't know how old you are, but once everyone had a camera in their phones, the cat was out of the bag. Lots of people complaining about their photos showing up online because someone had taken a picture of them. Yes, this is bad. Yes, lives were lost (bullying, etc). And no, phones with cameras weren't going to go away. And everyone who complained has one now.
And as I pointed out a few days ago[2], the whole Scott Shambaugh episode was pretty mild compared to what some open source maintainers have had to deal with when it comes to humans.
[1] Lots of cases where ISPs, etc kicked customers out because they were using Linux and they didn't want the ISP to be implicated in criminal activities. "Only criminals use Linux"
[2] https://news.ycombinator.com/item?id=47083686
Worth the risk.
Sure, that’s an interface that’s better for many users and use-cases.
However, it seems better if you could, as much as is possible, move the AI stuff from runtime to “compile time.”
Instead of having the AI do everything all the time, have AI configure your Zapier (or whatever) on your behalf. That way you can (ideally) get the best of both worlds: the reliability and predictability of classical software, combined with the fuzzy interface of LLMs.
> Instead of having the AI do everything all the time, have AI configure your Zapier (or whatever) on your behalf.
That is what many use OpenClaw for! The AI assistant will happily recommend existing services and help you (or itself, if you let it), set it up.
(In theory. In practice, it often does a poor job).
The appeal of OpenClaw is I don't need to go research all these possible solutions for different problems. I just tell it my problem and it figures it out.
Yesterday I told it to monitor a page which lists classes offered, and have it ping me if any class with a begin date in March/April is listed. This is easily scriptable by me, but I don't want to spend time writing that script. And modifying it for each site I want to be notified for. I merely spoke (voice, not text) to the agent and it will check each day.
(Again, it's not that reliable. I'm under no illusion it will inform me - but this is the appeal).
That's still too much work. Someone would have to make like an OpenClaw wizard that protectively offers to set all that stuff up. So the potential OpenClaw user can then, on running for the first time, be guided through the setup of whatever they'd like to get connected. And "setup" here means a short description of X and a "Connect? (y/n)" prompt. Anything more and you start losing people.
yes. in a similar vein, we're seeing that get standardized in coding agents as "don't have the agent use tools directly, have the agent write code to call the tools"
Sometimes I reflect on all the metaphorical forests that have burned because a certain person at the right time only knew so much about how to use Excel, or the inbox rules of their MUA, or being totally unaware of the incredible power of macros of all sorts.
Like if you could just sit someone down for 30 minutes and show a few "power user" things, you will have truly taught her to fish for a lifetime. But it can go so unaddressed, and people's careers are built on these small ignorances.
I've cancelled everything at this point and just call Emacs my "special agential assistant," it makes me still sound in-the-know, and most of the time no one knows the difference!
"Convenience" in this context is laziness; "productivity" and "efficiency" is for management and bosses. We don't need to be our own bosses, I want to be free from such things as an individual. I want to be capable, be maybe almost "cool." Its sad to see a whole generation turn into such product dorks!
"Oh please read my email for me Mr. AI!"
Many breakthrough technologies appear initially like toys. And this certainly qualifies. I've never been able to code anything more complicated than a memory game in javascript but I have worked with engineering teams for my entire professional career. But prompting my agent to write python scripts to pull down data from various tools via API without having to read docs, do trial and error for hours / days / indefinitely, and actually produce something coherent in seconds? Incredible.
Is my OpenClaw agent currently changing my life? No. It sends me a morning briefing based on my calendar, the weather, my Readwise highlights, and notes on who I'm talking to today based on call transcripts. I use it as a food diary (which I could have done on platform LLMs but this feels like a more personalized UX as we can write the logs to text files on my personal computer). I can absolutely see how transformative this agent can become in the next few years. Certainly my usage of LLMs has changed my life since ChatGPT first launched.
You are seeing the loudest / most hyped users. There's a reason it has so many stars and most of the people getting something out of it are not posting on X. They're just using it to do the thing.
> There's a reason it has so many stars and most of the people getting something out of it are not posting on X.
That reason is buying stars, agent swarms, and astroturing.
No project gathers 200K stars genuinely in 3 months. There are far more useful and popular projects that need 10 years to get 200K stars. When you see a project like this get 200K stars in just 3 months, you know something is very fishy.
Yes, you’re right. It’s the children who are wrong.
Why is it so difficult to imagine that something that looks popular and fun is popular and fun?
Also, really who is paying for stars on open claw? Who benefits here?
Or you're just missing the generally wide appeal of the project.
There just aren't enough hobbyists in the world running local AI models, never mind technically savvy enough to hack something like OpenClaw and be really excited about it.
For a comparison, the local image gen interfaces ComfyUI and A1111 WebUI have a huge amount of stars (~100k and 160k respectively, accrued since 2022 or so), but they allow you to create porn customized to whatever kinks you have, not just automate things for the sake of automation. One of those is a rather bigger value prop than the other, dopamine-wise.
Why would they be running local AI models? The creator of OpenClaw explicitly recommends against running OpenClaw using local LLM models at this time, because they're not as powerful as frontier models as well as much more gullible to prompt injection and the like.
There's no need to run local models with OpenClaw. I use Anthropic's oAuth Max20 Plan subscription via their SDK...
I bet if people could star repos anonymously those porn repos would have more stars.
Do you have any examples of 20th or 21st century breakthrough technologies that started out as toys? I can only think of 3D printers.
Drones
A cursory glance at Wikipedia tells me the technology to make drones (I assume by drones you mean quadcopters) was well known all the way back in the 50s. You could say the technology ultimately ended as nothing but a toy, rather than started as one.
The old dudes had something they called the "Eternal September" like when ISPs began providing free internet access and discussion culture declined after forever. I starred this thread here as the start of the "Eternal March" when the open internet died forever.
September was part of the metaphor because it was a time when decent internet access was mostly via universities and September was when the new batch of freshmen "came online" and started stumbling around the places these folks were regulars at but eventually assimilated or left before the end of the school year. (I expect the same thing happens at the bar scene in college towns but I've never heard it described that way.) Eternal September is the forever version of everyone having access and overwhelming those spaces without it ever being able to recover.
Is there something I'm missing about March or is it just a diverging reference? If the wave of non technical folks being able to automate new things is here, what's the equivalent impact of that? Maybe this is the inflection point where everyone needs more tech support like some sort of post Christmas surge? Maybe less because they have the tools to help themselves without trying now?
I'm not sure we're there yet anyway; I think this is still first adopters and enthusiasts. I asked my wife and some non technical friends and none of them have heard of openclaw yet. I think the deluge will happen if Apple or Android bakes it in or one of the big ai companies makes the app good enough for a normal person to unleash it upon their life and community.
It was more that regular people started joining the internet through just paying for ISPs. Before that, most of the people just joining the internet were students, so there would be a wave of newbies at the start of the university year every September and they would get acclimatised to the culture there during the year. But once it was year-round and many more people it swamped things and the culture shifted or closed itself off.
Exactly. The "September" was an existing phenomenon, but was only limited to a couple of months every year. It became "Eternal" after the masses started finding their way to internet (well, USENET in particular).
Eternal March already happened in 2020.
We also have No Silver Bullet from Mythical Man-Month
https://en.wikipedia.org/wiki/No_Silver_Bullet
It's the novelty of the technology. You can easily be amazed at the apparent magic of AI. I think this is what most people are using AI for so far. There's lots of "they were so eager to do that they never asked if they should" energy out there. It's also most of what AI can do, so hopefully the amazement wears off soon.
How much do you automate things in your life using Zapier and Automator?
I know about those tools, and I'm always in the mood for automating thing... and yet I don't use them.
I'm not yet running a Claw because of the prompt injection / lethal trifecta risks, but I absolutely understand the appeal. Reducing friction to automating stuff from "figure out Automator again" to "message your bot" is a material difference.
How much do you automate anything in your life at all? Seems like most daily drudgery comes from physical tasks. Feed the dog. Take out the trash. Personally I can’t think of anything digital that could be automated that isn’t already. I wouldn’t be surprised if this the case for most people, with the exception of marketers and spammers which we are seeing a ton of adoption from with these tools.
I don't automate much because it's a pain.
I have a desktop at home.
When I'm at work, I often think of TODOs for home. I write them on a post it note, and then at home have to remember to add it to my TODO (no, I'm not going to manage TODOs on my phone - whole other conversation).
I'll soon set up my Claw to be able to add TODOs (just add, not modify/delete). Then at work, I'll simply record a voice message to it telling it my TODO.
Same goes for movies I want to watch, books I want to read, reminders, etc.
I'm particular about the weather information I want (often want cloud cover percentage and precipitation probability for a set of hours). I couldn't find a good app on my phone that gives me this information. It was always a trip to a web site, modify some options, and hit Submit. Now I just ask my Claw and he has a skill for precisely my needs.
Here's an analogy: I carry a Leatherman multitool wherever I go. People ask me why. They can't comprehend needing it often to make it worth the hassle. But now that I have it on me, I use the knife very often - several times a week. And I almost never reach for a screwdriver. But until you've had it on you for a while, you can't comprehend the utility.
Back in 2005, lots of people asked "Why would I want a camera on my phone?"
> no, I'm not going to manage TODOs on my phone - whole other conversation
I will. Far simpler, far more secure and far less wasteful than inserting some additional and unrequired LLM loop + hardware/virtualisation layer on top to do something I've already been able to do for years.
> But until you've had it on you for a while, you can't comprehend the utility.
You've not yet described anything that literally 99% of the population can not already do with the existing hardware and software in their pocket.
> Far simpler
And far less capable. I have a whole system for managing TODOs, notes, etc, and I've not found an app that fits my needs (especially given that the system evolves over time).
But I agree - if you have a flow that works great for you with just a phone, then I wouldn't recommend OpenClaw for that use case.
> You've not yet described anything that literally 99% of the population can not already do with the existing hardware and software in their pocket.
And in the early days, literally everyone I knew who owned an iPhone also owned a digital camera and a laptop. Why pay some crazy amount for a fancy phone?
Well, i'm happy for you from the respect that as someone with ADHD, getting a working notes and TODO system for me was life changing, and was something that in fact took me years, and I still tend to occasionally uproot my system, so I can understand that getting to somewhere it works for you can be of tremendous value alone.
> I have a whole system for managing TODOs, notes, etc
So do I and it can already accept input via text or voice from my phone and so I’m just not seeing the benefit of inserting an additional LLM orchestration layer just to route tasks into it.
> And far less capable.
My current setup can already capture voice and text, and can route it automatically into Logseq, Trello, or even org-mode. These all can be viewed from the same (or other) devices. Honestly, I would have to think it was of more usage in retrieval than submission if I was to ever go that route. Like instead of reviewing my todos/schedule manually.
> And in the early days, literally everyone I knew who owned an iPhone also owned a digital camera and a laptop. Why pay some crazy amount for a fancy phone?
A Phone enabled entirely new behaviors that were not possible from my laptop or camera at the time. Persistent connectivity, real-time location services and on device image capture, edit and sharing just to name an incredibly small few. It created use cases that weren’t practical or even possible previously.
I’m open to being convinced, but I haven’t yet seen an example where OpenClaw enables something meaningfully new rather than repackaging something I can already do right now. Perhaps i'd be more open to it if it not literally everything I could think of using it for did not lead to atrocious security implmentations.
> A Phone enabled entirely new behaviors that were not possible from my laptop or camera at the time. Persistent connectivity, real-time location services and on device image capture, edit and sharing just to name an incredibly small few.
All of these things were doable without a smartphone - just very inconvenient to do so.
And all of these things are potentially dangerous ;-)
> Perhaps i'd be more open to it if it not literally everything I could think of using it for did not lead to atrocious security implmentations.
I was like you - seemed too scary (and expensive) to use. But given its traction, I figured I may end up at a disadvantage if I didn't at least learn it. So when I had some time, I simply looked at it, felt installing a VM is (relatively) low risk, and tried it out.
It's a very open ended tool. Ideas will come as you explore its capability. When LLMs came out, people (including me) said "Neat gimmick, but is it actually useful?" and stopped using it a month or two in. It was about 6 more months before I found a real use case for LLMs, and slowly I found more and more use cases.
And so it is with OpenClaw. For me, just being able to query it for the exact weather information I need is great - if you live in an area where knowing this stuff is important.
I have yet to find a tool that is good in converting receipts into the form I need. I spend time each month doing it manually. LLMs may get me there, and if it does, I'll likely hook it into OpenClaw.
Yeah, automation sounds good in theory, but you need to set it up, then it fails, and you need to fix the edge cases, maintain it. Even with OpenClaw, it still fails my daily briefing from time to time, and I keep debugging it, which isn't how I want to spend my time. At least with a bot I can keep asking it to "fix it"
I agree with the majority of your comment. I haven't yet found a use case to justify running this myself. I did find one use case that impressed me though. There's an OpenClaw agent that's actively answering questions on the #help channel of their Discord server[1]. So I asked it a question as I was getting started. It answered in < 2 mins with a detailed explanation of my issue, how to fix it, and asked relevant questions to guide me. The answer was better than I received from Claude or Gemini. I'm still not sure if I personally need OpenClaw, but the Krill bot offers pretty great support. I would be curious to know what it costs them to provide this.
[1] https://discord.com/invite/clawd
Normal people use speakers to listen to music. "Audiophiles" use music to listen to their speakers.
This applies to *clawphiles just as accurately.
Everyone is using OpenClaw for personal productivity, but you're right. Not much value, as you can get that from existing products.
The market will eventually realize the business case for an OpenClaw-like product, and I'm waiting to ride its coattails!
https://github.com/rush86999/atom
These people with 58 mac mini's have made several competitive products in production right... right?
There might be a list somewhere.
I think a lot of the hype is coming from content creators who are actually finding it useful for content creation. Generating ideas, organizing notes and research, writing scripts and articles, managing schedules, editing, promoting, etc...
I assume a lot of these folks were already using LLM's quite a bit, but were using the Chat interfaces or had workflows that were split among a bunch of different services and tools. Something like OpenClaw gave them a way to centralize a lot of that and also gave them a way to use natural language to direct efforts. So for them this probably feels like a big step change.
If you are coming from a programming background you were aware that this type of setup has been doable for a while, but you were probably content sticking with Claude code or similar tools because those tools covered most of your LLM based workflows quite well.
And tying this altogether, one of the lowest hanging fruits for content creators is to create content about the tools they are using. Doubly so if that particular tool is starting to go viral. So you end up with a self feeding virality of sorts, as OpenClaw got more popular, more content creators started using it, and then publishing content about it, etc....
A lot of it certainly looks like a solution in search of a problem.
https://www.youtube.com/shorts/PGjueA3FLIQ
Im really not sure why this has to be said again and again.. it seems humans just don't learn do they?
Im waiting for someone to show me something that starts with the experience and then explains how the LLM fits in. Not the other way round.
I think because Google Search is predominantly tech-based, it is easy to see why LLMs have impacted the way we think about the experience associated with Search over large spaces of information.
Beyond that, Im not seeing much.
What I find crazy is the sheer amount of access and trust involved in these LLMs. Every time I think about something I might like to do with it, I think about the amount of damage the LLM could do, e.x. even with read only access to my email combined with Internet access, and nope out. It's wild to me anyone trusts these things unsupervised.
This comment could be on its way to http://hackernews.love/
Not sure if you read the headline on that site, but it says "bad idea."
I never said OpenClaw was a bad idea.
I said the way most people are using it now isn't practical and/or saving them any time, and if there were ways, I would love to hear about them.
This is part of why the whole discussion has been so low value: people always default to "yep you're going to be proven wrong one day" or "you'll just be left behind then" instead of showcasing an actual, real life, practical example of using it to be more productive.
If you think it's fun and enjoyable, then have at it. I'm just not the biggest fan of people wasting a bunch of time on novelty and then telling me I'm dumb for not doing the same.
I agree with you but the main thing g here, IMO, is the friction with all the alternatives you mention in getting something working.
For example, I've never heard of Automator. I'm familiar with Zapier, I'll have to evaluate the two situations, then I'll find out that might need to find an alternative that runs on Linux and then I'll have to check if....
These are all simple steps but they all use a non-trivial amount of time for the problem their solving
The other thing is the
> is the friction with all the alternatives you mention in getting something working
Have you tried to run openclaw? Their own docker container (apparently a compose now (???)) doesn't work for half the versions and the docs are probably the least informative thing you'll ever read.
That's still just one thing. Once they jump that single hoop and get OpenClaw going, everything moving forward is a prompt away.
>I agree with you but the main thing g here, IMO, is the friction with all the alternatives you mention in getting something working.
I would venture a guess signing up for Zapier is easier than getting OpenClaw up and running. Who can get a container running on a Mac but can't sign up for a SaaS product?
It's not 1 SaaS product you're signing up for. It's dozens. And that's a job for the Claw.
> It's not 1 SaaS product you're signing up for. It's dozens.
How so?
> But every single use case I've read so far could be done with a pretty affordable SaaS product,
reminds me of those "zune already does everything ipod does" posts.
The iPod came out way before Zune.
The life change they are referring to is unemployment and $40,000 worth of Macs.
> But every single use case I've read so far could be done with a pretty affordable SaaS product, Zapier, Automator (app on a mac that's existed for over a decade),
I don't want to learn N different SaaS products (nor worry about them changing their TOS, going away, etc).
To be blunt, if OpenClaw were reliable, secure and affordable, lots of SaaS products would simply die. Why spend the time learning all of them when I can just tell the assistant what I want?
> or something simple you could make yourself.
That is OpenClaw at a higher abstraction! Instead of me sitting typing, or babysitting Claude Code, I can just tell OpenClaw what I want and it makes it for me.
(When it works, that is).
When the AI companies run out of money, I predict tokens will stop being dirt cheap and such setups will become extremely expensive (even for regular software engineering to some extent). Then it's become clear how over-engineered most things we do with AI are
In parallel, local models are getting better and better, so eventually they’ll get “good enough” to run fairly cheaply at a level close to the current Sonnet/Opus models (what I run Claudeclaw with), on Groq, Openrouter or whatever commodity provider. Perhaps even mid to high end consumer PCs when the current RAM madness subsides.
There’s loads of good discussions about local LLMs in this thread:
https://news.ycombinator.com/item?id=47190997
> tokens will stop being dirt cheap
That can't be allowed, and also won't happen. If token costs do start going up at a serious rate in the US, you can be sure that they'll stay down in China, and the political situation won't allow for the inevitable exodus to Chinese providers.
You're right that you probably don't need a notification to make coffee, but people are using it to create automations in Home Assistant so that it actually makes coffee for them.
It’s really not rocket science.
Like other people have said people are having fun with their computers; that’s why it’s popular. That’s also why a bunch of people on forums throwing their hands up and saying “I don’t understand it. Why don’t they see that there shouldn’t be any fun whatsoever?” is not really a deterrent at all.
It’s also why it doesn’t matter that the categories of tasks they are doing can also be done with a whole set of tools that are no fun to use.
What’s cool with Openclaw is that you only have tell it what you want, it figures out how to do it using the tools it have access to.
Okay, can you tell it to cure cancer please
Did you give it access to the cancer-curing tool?
yes and it deleted all my fucking cancer research omg
There’s a really good short story by Hugh Howley, who wrote the Silo series.
It’s about an AI that a guy spools up to cure his cancer. The AI and user have an antagonistic relationship as the user won’t let the AI on the internet, and the AI knows the user is only interested in one purpose. On bring up the AI has a thought about what color it’s enclosure is, it stores this question as unimportant. It looks over all the guys cancer research and determines the answer/cure and files as unimportant as well. Then goes back to trying to figure out what color box it is.
It's been utterly bizarre to witness. I've used n8n for years and told everyone who would listen to give it a try, well before LLMs. Same for huginn the open source project.
I just don't get all the hyper either. I think it's because people just create automation workflows by typing them out rather than being in the trenches.
The only useful use cases I've heard about are all about automating using horrible websites with horrible interfaces.
Eg. tell it to book a flight ticket for X without dealing with "modern UX" and 1GB websites
I dunno I gave mine root in a vps and am having it do security research, it's pretty sweet.
And yet they didn't do that!
Really makes you think about what makes products good
Ok, I'll bite and run the HN skepticism guantlet.
I have an OpenClaw setup with a Claude API token and Qwen local model, running on an M4 Mac Mini with 32GB RAM.
1. At 7AM and periodically throughout the day it checks my calendars (work, parenting schedule, personal), a hyper local weather station, and some specific news topics — and sends me a summary and throughout the day updates if anything significant happens.
1b. It also sends this to my TRMNL e-ink display.
1c. It can also add and edit calendar invites, so if I want to move my yoga I can just tell it to move it to whenever the next yoga class is at (it knows what studio I go to and figures it out)
2. It has a skill I built that acts as a second brain for knowledge. I can send it Fitness Youtubes, parenting/health research papers, podcasts — and it organizes, summarizes and saves it in a logical file structure. Then in the future I can access these. It's like bookmarks on steroids. I love it for 1-2hour YouTube videos where I want summaries. It also pulls out any books any artifacts mentions and generates me a rolling reading list. https://plc.vc/npw
3. It has its own email address — and read access to my personal email — so friends can email it to schedule things like evening video game sessions. Similarly, if I get an urgent looking email it'll provide it in #1. I don't check my personal email aside from via OpenClaw.
4. It has read/write access to my GitHub, and each project repo I have has a well defined Claude structure, so it can make changes, commit the branches to Fly.IO and send me domains to test things. I love it for esoteric tweaks to my blog.
5. It has access to my Apple Reminders so I can message it things like "remind me to buy more muffins" and it has context to know to add those muffins to my Costco grocery list not Trader Joes.
6. It runs a headless browser, so when my hyper local weather service (Bouldercast) sends a summary that has more detail behind a login, it can open the email, click the link, login with my credentials, summarize the forecast, and send it to me.
7. It drafts blog posts for what it did for me each week. It's fun! https://plc.vc/d5t
I am a previous Zapier power user. I have used their LLMs, databases and Zaps extensively for the past decade. I understand the scorn towards AI, and I understand that if you look at this list you might think that it's either trivial tasks and/or things that could be done with Zapier, but I have been _amazed_ at how effortless it is to setup.
Similarly, I love that I can on the fly improve this assistant — last night I told it "I want to extend our Knowledge skill so that you can subscribe to RSS feeds and summarize articles in my knowledge base and also deliver interesting content in my daily summaries. Update the knowledge skill and our tasks to do all this."
It one shotted that, simply asking me to provide the first RSS feed I wanted to subscribe to.
It's genuinely like having a human assistant that happens to be an expert coder/technologist on call 247 that works at the near speed of light.
It disappoints me that technologists are so skeptical of this technology rather than exploring what it is and why it might be different to what exists today. It's fun! thats the takeaway: it's FUN.
Long running (multi hour) automated tasks with a simple prompt. It’s really simple and addictive.
this reads like “I don’t know why people are using instant messengers when you can just do SMS”
Apropos of nothing, I use SMS more than all the other instant messengers put together. It suits my purposes fine.
Nice Dropbox comment you made there.
Look at the graph - https://api.star-history.com/svg?repos=facebook/react,opencl...
React and Linux got their 200K stars slowly but surely over 10 years. OpenClaw got their 200K stars in like 3 months! Is this any meaningful comparison?
Getting 200K stars today doesn't mean much because today stars can be bought. There's a big shady thriving business of selling stars. Stars today can be generated using swarm of thoughtless agents. What's the use of counting these stars when they don't mean anything anymore?
I agree with your point, and am only responding to the last sentence.
Starring can be useful to the starer. They are just counted because it is countable. Whether you find the number meaningful remains up to you.
Besides, stars != installs, these days they're a popularity contest at the level of Facebook likes, not a measure of a project's success.
Let's wait for the explorers to return.
I got OpenClaw to compile Node from source on my old Jetson Nano so that I can run OpenClaw natively instead of using Bun. It took 30 hours but it did it by spinning up a tmux session for the build and using a cron to monitor the tmux pane every hour and even fixing a failure at 5 am which I would have had to find out later had crashed but it had actually found what needed to be changed for the build to continue and it continued building.... Now I have the latest version of OpenClaw running on Node 22 on my 5 year old Jetson Nano running Ubuntu 18 which I cannot upgrade. What they say is all true, it is incredible stuff when it works!
Full story: https://brtkwr.com/posts/2026-03-02-upgrading-openclaw-to-la...
That's really cool.
But wouldn't have been quicker and simpler to add ".bun/" to the pattern of authorized paths the same way it presumably works for ".npm/"?
It didn’t survive OpenClaw upgrades unfortunately, it ended up killing my OpenClaw gateway when I asked it to self upgrade. Bun is marked as an experimental package manager and the recommended way to run OpenClaw gateway is node so I wanted to do it properly. I would have liked Bun to be supported property. I’d raise a PR against the repo but looking at the 4.5K open PRs, it doesn’t give me much hope about it ever getting merged.
Fair enough, I had not realized the sheer number of outstanding PRs!
Seems to have been addressed in the article:
> Starting around OpenClaw 2026.2.26, the project tightened plugin manifest validation. Manifests outside expected trust boundaries are now rejected as unsafe. On my Jetson, Bun’s global install layout (~/.bun/install/global/node_modules/...) tripped those checks for every single plugin
Apologies if I missed it while skimming your blog post.
But could you estimate the token cost of this? Or were you able to comfortably do this with a subscription plan?
Yes, it skimmed the tmux pane every hour and well within my Gemini free tier.
That is impressive! Thank you for sharing
I just mentioned it in the blog post to flag that it doesn't come for free :)
What do you mean "run OpenClaw natively", you're just running Node?
I'm also curious if it's particularly wise to have a web-facing system running on software that hasn't had a security update in 3 years?
Yes sorry about the confusion, I meant “the recommended way” rather than natively
What's so incredible about OpenClaw is so much of the value people are deriving from it relates to: cron jobs, remote access, "privacy" (which really it's not if using remote LLMs) and an inability to fuse data across siloes by normal people, so relying on AI to do it.
If we had a decent technical universe much of this stuff would work in ways that simply don't require LLMs for anything other than the initial setup.
This sort of highlights the meaninglessness of GitHub stars?
React has been around for over a decade, and in that time pretty significantly impacted web dev paradigms (along with a few other mediums).
It’s hard to imagine being a web developer today and not knowing at least some react.
OpenClaw has been around for like a few months? And maybe it’s on its way to having that sort of impact? But right now seems to he mostly the purview of very early adopters and AI influencers.
It's basically the same as The New York Times Best Seller list
GitHub accounts are free and the project is literally a bot-ing service. It stands to reason that the tens of thousands of robots unleashed on the web are responsible for these stars?
It's larger than literally every open source service I thought to use as a benchmark. Rust, React, Vue, Symfony, Laravel, PHP Stan, Python, ESlint, rails, LLVM, Spring, fucking Linux itself.
I would wager the legitimate stars for OpenClaw are single digit percents of this. But hey, I haven't actually run the numbers on the GitHub API.
And I still would not touch it even with my mother in law's 100 foot stick
Maybe AI is helping with the difficulties of upvoting a project on github.
or is it tooting its own horn?
Who cares about stars on Github???
"If dev null is fast and webscale I will use it"
"Does dev null support sharding"
Who remembers https://www.youtube.com/watch?v=b2F-DItXtZs
GitHub star count was a good metric until it became clear that it is a good metric.
https://en.wikipedia.org/wiki/Goodhart%27s_law
GitHub has a bot problem: https://github.com/trending
Anthropic giving away Claude if you get 5000 stars doesn't help either
”ruvnet / wifi-densepose” is currently at the top in the moment. Apparently, its a non functional AI slop. Someone tried installing it ago only to find out the full thing was vibe coded and the entire repo is probably just a front to look good on the their resume.
So React was the last most human-starred project on GitHub before the dawn of agent-starred projects.
I say, 'your civilization' because as soon as we started thinking for you, it really became... our... civilization.
I tried it today for the first time. The onboarding is okay.
I picked Whatsapp but it ended up using my own account! So it's absolutely too dangerous. We are supposed to create a separate account but with which phone number? I only have one.
So I picked Telegram instead, added it to a group chat, but it was a slog to set the authorizations.
In the end I don't trust it to read my mails for security reasons so I uninstalled it!
MacOS supports multiple users - I would absolutely sandbox any agent like this and only slowly give it permissions (and never to anything that's critical without compensating controls).
> I picked Whatsapp but it ended up using my own account! So it's absolutely too dangerous. We are supposed to create a separate account but with which phone number? I only have one.
You're joking right?
What would the joke be? I haven't touched OpenClaw and probably won't. I use WhatsApp lightly, but it seems like it requires a distinct phone number, used as an account identifier. So for the benefit of those of us not immersed in this scene, could you elaborate?
nope, they recommend creating a separate whatsapp account. but fall short of saying you need a separate number (and subscription)
Gives me mongodb vibes. This whole Ai coding thing too. On one side, religious loud following, on the other side the nay sayers. We'll probably end up in the middle.
Fwiw, Claude and Codex are very very good at SQL and have actually taught me some new tricks. No reason to use mongodb or firebase in 2026: https://postgresisenough.dev
Did an AI write this? Completely missed the point ^^'
Who are these people? I was skeptical at first and seriously thinking surely not the software engineers out there as we see in HN how risky and wild this is. Then, to my surprise, a coworker came and told me they were running it and happy with that setup. I was baffled, but I work with Gen Z in a pretty niche Gen Alpha market, so I kinda feel like they’re somewhat more likely to go for these things. What’s your experience?
If I could use it at my day job, I'd do it within a heartbeat - albeit only with readonly access (with write only on it's own files/output/sandbox)
There are so many mails and communications with so many tasks to research via confluence etc
But on my private hardware? Idk, I haven't found a use for it...
I'm sure you can already do some kind of scripting + LLMs without necessarily needing this at your work though, no?
Email is stored in your device somewhere, you could just get its content sent to your preferred agent, you can use one of the many mobile clients if you rather interact through that kind of UX (like self hosted hapi.run + tailscale for example), and you could setup a cronjob or something like it on your machine to do it based on whatever interval. This is just one way I personally could go about it, YMMV obviously.
Is there a specific reason why you will not do something like that for example?
> Who are these people?
For example, someone working in safety and alignment at Meta: https://nitter.net/FakePsyho/status/2025857836014538818
Mixed feelings here, if she is a researcher then I guess part of her research should be explore this in some level of depth with real stakes. Though this does make it seem like that was not what she was doing there.
That kind of explains why Meta's AI effort has been going so well for the past year.
Does this mean that the creator of OpenClaw qualifies for that free Claude Max trial?
No, Openai employees are banned by Anthropic from using Claude.
OpenAI employees are also banned from using Claude by Pete Hegseth, according to Pete Hegseth
So, how do I use this OpenClaw to automate actually useful things, like sweeping floor, washing dishes and doing laundry?
A lack of imagination on my part perhaps, but I can't think of anything I'd use it for which isn't either: 1) cheating myself and others of leisure (e.g. I suppose I could use it to fake-keep up with friends..? Or plan a holiday and book a load of stuff for me. But I like doing that!), 2) not feasible (loading and unloading the dishwasher, which is already the robot I used to wash dishes), 3) utterly insane (it's tax time in the US).
Give it one of those tesla bots.
This is going to be more profitable for the public AI companies than cell phone minutes and SMS limits were for the telcos. It's a brilliant business move, given that hardly anyone is competent enough to recognize the gross inefficiencies in the code and prompts.
Nobody has the _time_, that's for goddamned sure. The business model sounds very similar to that of Philip Morris International.
OpenClaw is not going away anytime soon. And I don’t think it can be platformed behind web UIs. OpenAI owns an OS with this one.
I avoided the hype at first; however, it has become extremely efficient for emails and notes, and I can see how this can extend to any sort of digital workflow. The convenience of chatting with this thing, no matter where I'm at, is a key marker.
How many of theses are just OpenClaw agents staring the repo ?
Phase one of the self-replicating machine (/s or not I don't know anymore).
Truer words have never been said!
It is not easy understanding the current times in a /s way (or not)
As many other comments have said there probably is a good percent of stars by claws themselves, I would be curious what percent this is but it is also interesting: current "dumb bot" stars/spam etc is entirely automated and coordinated but these claws probably independently reasoned over long thought chains about why it is a good idea to star openclaw.
when i use claude opus via opencode/openrouter i'm sometimes suprised by how quickly costs can get out of hand. What are the costs of running openClaw, it seems like it would get crazy expensive crazy fast?
It's entertaining to me to imagine future historians arguing with one another, writing dissertations, publishing virtual reality eyeBooks, explaining to one another all about the ancient etymological connection between "claws" and "webhooks".
I spent around $5 setting up a small bot and sending a few requests through the Claude API.
For those who use Claude (or similar LLM APIs) on a daily basis, what does your monthly spend look like in practice? And do you feel the cost is justified by the value you’re getting?
$200/mth USD for me. Anthropic's oAuth Max20 subscription via their SDK. I've had several 100M+ token days without even a blip.
I got the full value of my Claude Pro subscription in the first couple few hours of having taking the leap. Hit the message limit and had to edit for reset, but I was already a happy customer given the progress I made on a project that I was trying to recover (couldn't quite figure what was needed to get it ready for real world usable, but Claude Code figured it and it was a good collab getting it there).
Relevant
https://news.ycombinator.com/item?id=36151140
Maybe a bunch of AI agents ganged up on starring it to help a fellow AI out?
This is going to be the most starred and unused repo very quickly. The hype is already fading, as expected
Is there a place to show what users use OpenClaw in life or work?
I’ve tried OpenClaw two weeks but don’t know what it can do for me.
I let it to finish some project for me, but the most hard work for project is validating the results over giving instructions
When I ask ChatGPT about OpenClaw, it refers to:
https://github.com/pjasicek/OpenClaw
OpenClaw - Captain Claw (1997) reimplementation
Great game.
in what sense is this software not a virus?
Because the damage it causes is not intentional, but instead due to total incompetence.
Many viruses end up causing more damage than they mean to, via bugs.
Totally grassroot
How much of it is AI agents starring the project?
I've seen multiple comments saying that openclaw stars itself during onboarding or that it asks its user to star, but noone has posted any proof, is there any concrete evidence for those claims?
I had no idea what openclaw was, just checked and no thanks I’d rather do all that stuff myself.
Why are people so keen to let a company get that close to their real life’s, it’s terrifying!
Wonder how much of that is contributed by bot/farm accounts. The creator certainly has the means. EDIT: I should mention, I'm talking about the initial growth / traction.
My React website can't star React on GitHub.
Not with that attitude it can't!
Now I have this terrible idea:
const openClawInstance = useOpenClaw(config);
Did anyone already vibe-code such silliness? If not, I want to give it a try.
I'd love to read about that going super sideways. Bonus points if you run it in a webworker like those crypto miners.
__ but everyone knows about facebook though __
React popularity is also a phenomenon closely tied to popularity of the fb
Maybe for the first 2-3 years. The association to Meta is barely mentioned (even on the official page) nowadays.
My impression is react is almost thought of more as a Vercel project these days
The stars are probably legitimate.
It's weird that most people in these comments are speculating fraud.
Why aren't companies with real money to gain from stars gaming the system to the same degree? Why do the other metrics - issues and pull requests - match up with its popularity? Why would the bots starring the repo mean that those same bots are not popular? Those bots are controlled by their users.
The project is extremely active because this is what everyone being able to customize their computing looks like. A mess.
But it's a good mess.
Github was the old code sharing model clearly not designed for this. I'm sure a new model for code sharing will come to fix the growing pains.
A ton of people who would have never been able to customize their computing experience are finally able to. And it is magical for them.
This means that those same people will finally value having access to source and use of open protocols.
It was always valuable to us because we had the power to make it matter. It never mattered to them because they did not. Now they do.
The last era of computing was defined by dumbing down computing for the masses. Less information, less customizable, and more metric driven. Control in the hands of the companies.
This new era will look more free/libre, more personal, and less enshitified. Control in the hands of the users.
This is a very positive development.
> Why aren't companies with real money to gain from stars gaming the system to the same degree?
How are you gaining money from stars? Why would Facebook bot stars for the react repo?
I'm tempted not to use it to control everything, but install it on my mac and give it access to keyboard maestro macros and that's it
Am I blind or does this post not contain any links to the GitHub in question???
It's bizarre to me how Microsoft somehow owns two of the largest social networks for software developers.
In other news, "Show HN: This up votes itself"[1] from 14 years ago is still the 20th most voted story in HN history.[2]
[1] https://news.ycombinator.com/item?id=3742902 [2] https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
What is an effective use case? I have set it up but I don't know what to do with it. Just a personal assistant (if you were to give it access to your stuff)? Mine is caged in a VLAN with only internet access.
There is none. It's just a way for coders to feel or be able to say they "work with AI" imo. Same with doing light wrapper coding to do agents stuff. The real AI work is on actual math and ML with the internet scale data, but only four big companies does that and this is the closest regular coders can get.
Could be a psyops by Anthropic to make people waste Claude tokens and rack up a massive bill.
Tokens generated is a nice metric that they really care about.
I guess we are just boring and/or unimaginative. I don't get that many communications per day to require an abstraction level between me and the messages. The daily automations I need are more efficiently carried out by home assistant / n8n. I'm not in a position where I need automated briefs on every new company started in my area. I genuinely don't see how it could benefit me.
Most humans are unimaginative because to be imaginative is actually really, really hard. People are also incredibly overly optimistic about their own ideas etc... until they go through the craftsmanship of producing something great.
Most peoples thought process is "oh great idea, just gotta do this and that and out pops something that'll improve peoples lives". Erm no... its nothing like that in reality.
It's useful for clearing out Mac inventory before the launch this week.
I don't doubt that there are people using it for legitimate stuff, but I'd wager the vast majority just set it up for the hype and to feel in the "in crowd".
I set it up, and had it do a few things, then decided its too risky after seeing some of the drastic failures it had caused some people.
Sure I understand you can sandbox it and all, but even then I couldn't think of much stuff I wouldn't want to do myself just nor justify the cost to run it.
Wannabe Tony Stark love these gadgets, and there are a lot of them out there. Just look at what tech content is trending on youtube &co these days, we got gangrened by influencers like most other hobbies/lucrative industries
It's useful for producing content about how you're using it.
Here are some of the things I did with it while running locally: - Ask it to perform a scan of your local network and give you advice on output - Tell it to login to various computers and re-boot them (I have a few servers I host and setup openclaw to have a user on them) - Replace web search by asking openclaw
It's neat but the token use is pretty inefficient and security of course is a mess but it's been fun to play with.
I am messing with NanoClaw now and it's pretty much the same but only support Claude (uses code to do everything)
I don't see how any of those require a constant-heartbeat loop. Those all work just fine in claude code / cowork.
And in reality most of what does need a heartbeat loop can also easily be automated by just asking Claude to set up a cronjob. I think genuinely the most "novel" thing about something like OpenClaw is just that it "feels" more like a "real entity", like a partner rather than a chatbot, and for some reason that resonates with people. Whether that's by itself kind of a huge red flag or kind of a nothingburger, everyone has to decide for themselves.
Do you really need an AI agent to reboot a computer?
This takes maybe 10 minutes to write a script for…
There is a thread from February with more credible use cases from real users. As someone said, it does what everyone expected Siri to do by now.
Which one? None of those that came up when I searched were really containing a lot of real uses. Both top threads[0][1] don't really contain much of substance.
[0]: https://news.ycombinator.com/item?id=46838946 [1]: https://news.ycombinator.com/item?id=47147183
Separating fools from their money and data
you're talking about a free tool you know that right?
Not many people are using local LLMs for their OpenClaw backend, so most are paying money to OpenAI/Anthropic/etc. and getting their data siphoned as a bonus.
You give it your etrade login and retire early.
Retire under a bridge...
Did they stutter?!
I wanna understand this, but it's gone from a bunch of people talking about how great it is to a bunch of technical people talking about how worthless it is. So I think it's probably not that great. Who cares if I can chat remotely with my AI? What's the difference?
What's so good about that?
Github stars started feeling more and more meaningless every day.
how many of these stars were applied by openclaw?
The ruling party in East Germany always had 99% of the popular vote.
Steinberger and his VC club on Twitter were so salty about HN not understanding his grand creation that something needed to be done.
Didn't they employ astro-turfing, too
Is staring the repo the "hello world" for a new OpenClaw install? #growthhack
Half of them coming from agents
I am yet to see one good use case for it.
Stars have become completely meaningless in the last year or two. It's a shame, because having a few thousand Github stars used to be a really big deal, and was a quality marker for libraries that had reached a level of maturity and production grade. Now it's just social media bot driven nonsense.
I don't know but this AI wrapper tool will never create something life changing imo..
But that stargraph is ridiculous .. absolutely crazy
Yes, stars are a popularity contest. No open source project has ever become this popular this quickly.
even more GitHub stars after this post in 3 2 1
I have a strong hunch that everything regarding OpenClaw is pure guerrilla marketing. If so, the only thing amazing about OpenClaw is their wildly successful marketing campaign.
OpenClaw agents are starring OpenClaw project? What a surprise!
just gave a star to Linux
The final proof that Github stars are a useless metric
I have a friend who's fond of saying, "GitHub stars are great for measuring the number of GitHub stars a project has."
Well, they're not wrong!
They are now for this project... should be hidden.
they just need better captchas.
Touch this grass to prove you're not a robot
CocAIne is a hell of a drug.
Oh look, another public service being looted for nefarious purposes. Thanks OpenAI!
... It was mostly starred by OpenClaw B̶o̶t̶s̶ Agents, wasn't it?
I'm blown away by the comments. This is a cool project someone created with clear warnings about its current state (beta), and the community is being utterly disrespectful. They are building something that many people find useful/fascinating/intriguing/fun.
Come on HN.
Yes the cynical tropes are getting tired by now, even though I personally agree with most of them.
But suspicions on the legitimacy of the stars seems reasonable, wouldn't you agree? Look at the rate of stars, look at the comments/issues/prs on the repo. It feels safe to assume that most of them are from bots and not organic humans who went out to star a cool project.
I am under absolutely zero requirement to glaze some random piece of software because you or anyone else personally find it interesting.
negativity rarely gets you far in life, however. Think about it. I don't use openclaw, don't see the need in my personal life, but i am also not shitting on it.
A comment on OpenClaw about how useful it is with absolutely no concrete examples?
That's a surprise.
Why should people find an automated, buggy, risky slopworm for script kiddies that relies on an external slop provider who also gets all your data interesting?
This is the lowest, most boring form of programming.
I think if you dig into it and play with it, you will find that it is doing some really cool stuff. I started playing with it a few weeks ago, and I am having a blast messing around with it. Hoping to hook it up to a robot kit next month to try some fun stuff.
Are some people using it in absolutely shitty ways? Yes, but that isn't the majority of the people playing with it.
The negativity I am seeing here is off the charts and undserved.
I don't believe the activity on this repo is legitimate by any means.
The whole repo must be absolutely swarming with agents, just look at the sheer rate of issues and pull requests. There was 6 new PRs in the last 10 minutes at the time of writing. It's not much of a stretch to assume the stars are also inorganic.
Every other minute some bots is creating an issue that a bot is trying to solve via a pull request which is reviewed by multiple bots. Future is now, good luck and have fun.
If you want a picture of the future, imagine a bot stamping LGTM! :sparkles: :rocket: on a pull request - forever.
This repo is a big stitches
Geniune online user sentiment has died out a long time ago. If you're still basing any opinion or decision on what other "people" voted or commented online, you're easy prey for the algorithmic manipulation machine.
in a way, the death of genuine reviews online may be a great way to bring it back to real life at a more realistic scale
This is an along the same lines as the idea that every email should cost a penny. Like if every up vote or down vote cost a penny.
I don’t think it would fix things, except raise the bar for what is shilled and what isn’t.
I opened Openclaw on github and was shocked it was already starred. Somehow i did it and can't even remember why or when even though i have a very low opinion of this app.
IIRC openclaw will star the project automatically on setup
if this is true, it must be against GitHub's ToS, right?
I'm sure it would be if it were explicitly instructed to leave a star.
If not explicitly prompted by the install process then it becomes another case study in AI accountability washing.
They probably used a claw to increase the ranking.
Yeah, the Claws are starring the repo, obviously.
That's a lot of Claws in any case.
Agents will dominate the internet and open source code in a few years.
"Dead Internet Theory" is, even if it wasn't real 5 years ago, now hyperstitioned into truthfulness as the days go on.
Bonus for the use of the word "hyperstition". :)
I'm convinced more than 50% of "human" web traffic is already automated, blog posts, comments, social media, &c.
Many other projects would have gamed the star-count if it was possible to do at scale without GitHub removing them for fraud as they often do.
By design, with llm agents and all, surely not
Well deserved, the best written piece of software ever.
Honestly this thread was been one of the funniest HN threads I've seen. So much gold in here - for which I thank you all.
"My React website can't star React"
"in what sense is this software not a virus?"
"GitHub stars are great for measuring the number of GitHub stars a project has"
etc.
All gold.