I have an open source project and started receiving a lot of security vulnerability reports in the last few months. A lot of them are extremely corner cases, but there were some legit ones. They're all fixed now. Closed source software won't receive any reports, but it will be exploited with AI. So I definitely agree with the message of this article.
As someone who works on closed source software and has done for a couple of decades, most companies won't even know about that and of those who do only a fraction give enough of a shit about it to do anything until they are caught with their pants down.
Having worked in quite a few agency/consultancy situations, it is far more productive to smash your head against a wall till bleeding, than to get a client to pay for security. The regular answer: "This is table stakes, we pay you for this." Combined with: "Why has velocity gone down, we don't pay you for that security or documentation crap."
There are unexploited security holes in enterprise software you can drive a boring machine through. There is a well paid "security" (aka employee surveillance) company using python2.7 (no, not patched) on each and every machine their software runs on. At some of the biggest companies in this world. They just don't care for updating this, because, why should they. There is no incentive. None.
Yea, its fundamentally an issue of asymmetric economics.
Running AI scanners internally costs money, dev time, and management buy in to actually fix the mountain of tech debt the scanners uncover. As you said there is no incentive for that
But for bad actors the cost of pointing an LLM at an exposed endpoint or reverse engineered binary has dropped to near zero. The attackers tooling just got exponentially cheaper and faster, while the enterprise defenders budget remained at zero.
In theory though, there is now a new way for community to support open source, but running vulnerability scans in white-hat mode, reporting and patching. That way they burn tokens for a project they love. Even if they couldn't actually contribute code before.
There should be a way to donate your unused tokens on every cycle to open source like rounding up at the chekout!
That sounds like a great idea. I'd love to be able to contribute the remainder of my monthly AI subscriptions for something like this, especially since some of them bill and refresh their quotas by calendar month.
Hang on, why is it costly for in-house to run AI scanners but near zero for threat actors to do the same?
I've seen multiple proprietary places now including a routine AI scan of their code because it's so cheap and they may as well use-up unused tokens at the end of the week.
I mean, it's literally zero because they already paid for CC for every developer. You can't get cheaper than that.
As I mentioned above, we actually do run these AI scanners on our code, but the problem is it's simply not enough. These AI scanners, including STRIX, don't find everything. Each scanning tool actually finds different results from the other, and so it's impossible to determine a benchmark of what's secure and what's not.
> As I mentioned above, we actually do run these AI scanners on our code, but the problem is it's simply not enough. These AI scanners, including STRIX, don't find everything. Each scanning tool actually finds different results from the other, and so it's impossible to determine a benchmark of what's secure and what's not.
Yeah, but with closed source it's cheaper for the defender than for the attacker - the defender can scan their sources and their PRs as well as the compiled output. The attacker can only scan the compiled output, and they have to perform repeated scans.
I think it makes it all the more apparent that writing EAL4 code with as little design competence as possible was taking advantage of some strange scarcity economics.. It's now even easier to make something with endless technical debt and security vs backwards compatibility liability but is anyone going to keep paying for things that aren't correct and to the point if some market participants structure their agent usage toward verifiable quality and don't actually have more cost any more?
> What makes you so sure that closed-source companies won't run those same AI scanners on their own code?
How many companies take the time to use penetration testing tools, that have been available for many years, to verify their software (or pay a penetration testing company to do a more thorough job than they have the experience to do internally)?
Some, certainly. Many, possibly. Most, I would wager not.
The economic motivation just simply isn't there.
I'm sure we could cherry pick a few examples of companies where things like quality and security really are part of the culture, not just feel-good lip services. The reality is that companies are in business to make money and corner cutting is the easiest way to pad the margins.
Because they're a company. Even if the bar to entry can fit a normal sized american, doesn't mean they will do it, or do it in a systematic way; We know very well that nothing about AI is naturally systematic, so why would you assume it'll happen in a systematic way.
The point being that there are always going to be more eyes, and more knowledge of available tools (i.e. including "D, E and F"), and more experience using them, with open source than with a single in-house dev team.
If true then logically it will be sufficient to run this "master model" once before any code release for the level playing field to be restored. After all, even open-source software is private until it is released.
> If true then logically it will be sufficient to run this "master model" once before any code release for the level playing field to be restored.
I'm struggling to see how it is a level playing field:
1. Closed-source: defender runs llms to check the sources for vulns, runs llms on each PR, runs llm on deployment of the compiled output. Attacker runs llm only on compiled output.
2. Open-source: both attacker and defender runs llms on source, on PRs and on compiled output.
> Closed source software won't receive any reports
Not from the automated repo scanners, but bug bounty programs can generate a lot of reports in my experience. AI tools are becoming a problem there, too, because amateurs are drawn to the bounties and will submit anything the AI hallucinates.
Closed source companies can (and should!) also run their own security audits rather than passively waiting for volunteers to spend their tokens on it.
Those bug bounty programs now have to compete against the market for 0-days. I suppose they always did, but it seems the economics have changed in the favour of the bad actors - at least from my uninformed standpoint.
That still exists in the OSS world too, having your code out there is no panacea. I think we'll see a real swarm of security issues across the board, but I would expect the OSS world to fare better (perhaps after a painful period).
Of course everyone should do their own due diligence, but my point is mostly that open source will have many more eyes and more effort put into it, both by owners, but also community.
That's absolutely our plan. We have bug bounty programs, we have internal AI scanners, we have manual penetration testing, and a number of other things that enable us to push really hard to find this stuff internally rather than relying on either the good people in the open source community or hackers to find our vulnerabilities.
+1, at this point all companies need to be continuously testing their whole stack. The dumb scanners are now a thing of the past, the second your site goes live it will get slammed by the latest AI hackers
> Not from the automated repo scanners, but bug bounty programs can generate a lot of reports in my experience. AI tools are becoming a problem there, too, because amateurs are drawn to the bounties and will submit anything the AI hallucinates
You don't even need a bug bounty program. In my experience there's an army of individuals running low-quality security tools spamming every endpoint they can think (webmaster@ support@ contact@ gdpr@ etc.) with silly non-vulnerabilities asking for $100. They suck now but they will get more sophisticated over time.
I don't follow. It seems obvious that there's more to gain for attackers using AI agents to exploit open source repositories, than there is for good samaritan defenders. In this new closed-source world (for Cal.com), there's nothing stopping them from running their own internal security agent audits, all whilst at least blocking the easiest method of finding zero-days - that is, being open source.
This really just seems like Strix marketing. Which is totally fair, but let's be reasonable here, any open-source business stands to lose way more by continuing to be open-source vs. relying on the benevolence of people scanning their code for them.
> It seems obvious that there's more to gain for attackers using AI agents to exploit open source repositories, than there is for good samaritan defenders.
Actually the opposite is obvious - the comment you replied too talked about an abundance of good Samaritan reports - it's strange to speculate on some nebulous "gain" when responding to facts about more then enough reports concerning open source code.
> In this new closed-source world (for Cal.com), there's nothing stopping them from running their own internal security agent audits
That's one good Samaritan for a closed source app vs many for an open source one. Open source wins again.
> any open-source business stands to lose way more
That doesn't make any sense - why would it lose more when it has many more good Samaritans working for it for free?
You seem to forget that the number of vulnerabilities in a certain app is finite, an open source app will reach a secure status much faster than a closed source one, in addition to also gaining from shorter time to market.
In fact, open source will soon be much better and more capable due to new and developing technological and organizational advancements which are next to impossible to happen under a closed source regime.
The main drawback is that you will need to be able to patch quick in the next 3-5 years. We are already seeing this in a few solutions getting attention from various AI-driven security topics and our previous stance of letting fixes "ripen" on the shelf for a while - a minor version or two - is most likely turning problematic. Especially if attackers start exploiting faster and botnets start picking up vulnerabilities faster.
But at that point, "fighting fire with fire" is still a good point. Assuming tokens are available, we could just dump the entire code base, changesets and all, our dependent configuration on the code base, company-internal domain knowledge and previous upgrade failures into a folder and tell the AI to figure out upgrade risks. Bonus points if you have decent integration tests or test setups to all of that through.
It won't be perfect, but combine that with a good tiered rollout and increasing velocity of rollouts are entirely possible.
It's kinda funny to me -- a lot of the agentic hype seems to be rewarding good practices - cooperation, documentation, unit testing, integration testing, local test setups hugely.
Let's say finding a security issue takes 10M tokens. If one company has to pay for it, they most likely won't bother. It's purely a cost/benefit thing for them.
But if you have an open source project, you might get a 1000 people looking at it, each only has to spend 10k tokens to find the same flaws.
A new user is much more likely to scan the codebase and report vulnerabilities so they can be fixed than illegally exploit them since most people aren't criminals
Some users might be tech sensitive and have the capacity to check the codebase
If a company want to use your platform, it can run an audit with its own staff
These are people really concerned about the code, not "good samaritans"
I’ve recently set up nightly automated pentest for my open-source project. I’m considering starting to publish these reports as proof of security posture.
If the cost of security audit becomes marginal, it would seem reasonable to expect projects to publish results of such audits frequently.
There’s probably a quite hefty backlog of medium- and low-severity issues in existing projects for maintainers to suffer through first though.
We actually run AI scanners on our code internally, so we get the benefit of security through obscurity while also layering on AI vulnerability scanning, manual human penetration testing, and a huge array of other defence mechanisms.
"Security through obscurity" is a term popularized entirely by the long-standing consensus among security researchers and any expert not being paid to say otherwise that this is a bad idea that doesn't work
> Closed source software won't receive any reports, but it will be exploited with AI.
This is what worries me about companies sleeping on using AI to at a bare minimum run code audits and evaluate their security routinely. I suspect as models get better we're going to see companies being hacked at a level never seen before.
Right now we've seen a few different maintainers for open source packages get hacked, who knows how many companies have someone infiltrating their internal systems with the help of AI because nobody wants to do the due dilligence of having a company do security audits on their systems.
Assembly is still source code so really it comes down to if the copy protection is obscuring the executable code to the point where the LLM is not able to retrieve it on its own. And if it can't someone motivated could give it the extra help it needs to start tracing how outside inputs gets handled by the application.
absolutely agree with you if we're talking about clean room reverse engineering; but in the context of finding vulnerabilities it's a completely different story
I mean-- to an LLM is there really any difference between the actual source and disassembled source? Informative names and comments probably help them too, but it's not clear that they're necessary.
Which models have you had good luck with when working with ghidra?
I analyze crash dumps for a Windows application. I haven't had much luck using Claude, OpenAI, or Google models when working with WinDbg. None of the models are very good at assembly and don't seem to be able to remember the details of different calling conventions or even how some of the registers are typically used. They are all pretty good at helping me navigate WinDbg though.
> Closed source software won't receive any reports, but it will be exploited with AI
How so? AI won't have access to the source code. In some cases AI may have access to deployed binaries (if your business deploys binaries) but I am not aware that it has the same capabilities against compiled code than source code.
But in a SAAS world, all AI has access to is your API. It might be still be up to no good but surely you will be several orders of magnitude less exposed than with access to source code.
Claude is already shockingly good at reverse engineering. Try it – it's really a step change. It has infinite patience which was always the limited resource in decompiling/deobfuscating most software.
It's SaaS though. You don't have access to the binary to decompile. There's only so much you can reverse-engineer through public URLs and APIs, especially if the SaaS uses any form of automatic detection of bot traffic.
Thanks you. This is what the parent post was trying to say. Don't know why it is down-voted. AI or not, if the API end points are well secured, for example use uuid-v7, then their is little that the ai can gain from just these points.
The opposite is true. Open source barely matters to attackers, especially ones that can be automated. It mostly enables more people (or agents, or people with agents) to notice and fix your vulnerabilities. Secrecy and other asymmetries in the information landscape disproportionately benefit attackers, and the oft-repeated corporate claim that proprietary software is more secure is summarily discounted by most cybersecurity professionals, whether in industry or academic research. This is also seldom the motivation for making products proprietary, but it's more PR-friendly to claim that closing your source code is for security reasons than it is to say that it's for competitive advantage or control over your customers
Yes exactly! I'm so glad I took this route with my startup. We can't bury our heads in the sand and think the vulnerabilities don't exist just because we don't know about them.
Exactly. I respect their decision to go closed source if that's what they need to do to make it a viable business, but just be honest about it. Don't make up some excuse around security and open source.
it's not necessarily about people self hosting it, it's about people preferring to pay for hosted stuff that is open source (e.g. I pay for Plausible).
Now it's a lot easier to rewrite open source stuff to get around licensing requirements and have an LLM watch the repo and copy all improvements and fixes, so the bar for a competitor to come along and get 10 years of work for free it a lot lower.
We've run an extremely profitable business for five years, raised a seed and a Series A, and grown at 300% a year sustainably while being open source.
Going closed source actually hurts our business more than it benefits it. But it ultimately protects customer data, and that's what we care about the most.
I think if it ultimately protects customer data in a significant way, I would be for it.
Are you able to share any more detail on how you determined this is the best route? It would be a significant implication for many other pieces of open source software also if so.
(And I say this is someone who just recommended cal.com to someone a few days ago specifically citing the fact that it was open source, that led to increased trust in it.)
I think if you are committed to switching back to open source as soon as the threat landscape changes, and you have some metric for what that looks like, that would be valuable to share now.
I would like to see the analysis that you're referencing around open source being 5-10x less secure.
I literally have a Claude Code skill called "/delib" that takes takes in any nodejs project/library and converts it to a dependency-less project only using the standard library.
It started as a what-if joke, but it's turned out to be amazing. So yeah, npmjs.com is just reference site for me now, and node_modules stays tiny.
And the output is honestly superior. I end up with smaller projects, clean code, and a huge suite of property-based tests from the refactor process. And it's fully automatic.
It varies from project to project, but applications benefit a lot more than libraries. When I de-lib a normal express app it might add a few hundred lines of code and a few thousand new tests, but if I de-lib an library then depends on how ancient it is. The older the library is, the higher the chances that most of what it needs is built-in to the standard library.
Thank you. I have been thinking about the same approach. However my worry is the open source libraries often gets more eyeballs and CPU cycles and ends up much more refined over a period of time.
separating codebase and leaving 'cal.diy' for hobbyists is pretty much the classic open-core path. the community phase is over and they need to protect their enterprise revenue.
blaming AI scanners is just really convenient PR cover for a normal license change.
Yes. Before AI the source was a demonstration of your substance. Users would be encouraged to reach out to maintainers to pay for upgrades or custom tweaks or training. Or indirectly pay for advertising while reading docs. After AI those revenue streams have collapsed. Now you have to withdraw enough of the work to make it hard for an individual to recreate with an LLM. The open source needs to be restricted to a rich interaction layer. Cloudflare just announced they are using that model with their services which were already closed source but now they are exposing them through new APIs. So they can capitalize on existing services that were not ripe enough for SaaS before AI, that had to be handled by their in-house professionals services folks. With this move they are using AI to expand/automate their white glove professional services business to smaller customers.
This is part of it for sure. It is also true that many open source business depended on it not being worth the trouble to figure out the hosting setup, ops etc, and the code. Typical open source businesses also make a practice of running a few features back on the public repo.
Now I can take an open source repo and just add the missing features, fix the bugs, deploy in a few hours. The value of integration and bug-fixing when the code is available is now a single capable dev for a few hours, instead of an internal team. The calculus is completely different.
Yes, it feels like they've been looking for an excuse to go closed-source, and this one is plausible enough to make it sound like they're only doing it because they "have to".
1) Pulls you in with a catchy title, that at first glance seems like a dunk on Cal.com (whatever that is).
2) Takes the "we understand your pain" approach to empathize w/ Cal.com, so you feel like you're on the good vibes side.
3) Provides a genuine response to the actual problem Cal.com is dealing with. Something you can't dismiss out of hand.
4) But in the end of the day, the response aligns perfectly with the product they're promoting (a click away to the homepage!)
This mix of genuine ideas and marketing is quite potent. Not saying this is all bad or anything, just found it a bit funny. The mixed-up-ness is the point!
That's exactly how this read to me too. Ultimately, the whole article is written by a company that does AI vulnerability scanning, and it's to try and get you to sign up for their service.
As it mentions in their article, Strix actually scans the Cal.com codebase and reports vulnerabilities to us. But the reality is, they actually miss so many vulnerabilities that other platforms do find. There's no one platform that seems to be able to reliably find all vulnerabilities, and so simply adopting AI scanners just isn't enough.
Is it good marketing though? I mean personally I do not use AI, and I don't think this opinion of mine will change. I can't look into the future, but right now I don't use nor do I depend on AI. I guess it may work for some people, but even then I am unsure whether that is really good marketing. Riding on a hype train (which AI right now still is) is indeed easier, so that has to be considered.
>Security through obscurity is a losing bet against automation
Security through obscurity is only problematic if that is the only, or a primary, layer of defense. As an incremental layer of deterrence or delay, it is an absolutely valid tactic, with its primary function being imposing higher costs on the attacker.
As such if, as people are postulating post-Mythos, security comes down to which side spends more tokens, it is an even more valid strategy to impose asymmetric costs on the attacker.
"With enough AI-balls (heheh) all bugs are shallow."
From a security perspective, the basic calculus of open versus closed comes down to which you expect to be case for your project: the attention donated by the community outweighs the attention (lowered by openness) invested by attackers, or, the attention from your internal processes outweighs the attention costs (increased by obscurity) on attackers. The only change is that the attention from AI is multifold more effective than from humans, otherwise the calculus is the same.
I wonder whether cal actually has concerns about security (in which case, they're wrong, this argument was false when people made it decades ago), or whether they just took a convenient excuse to do something they wanted to do anyway because Open Source SaaS businesses are hard.
Isn’t the real danger now not the ability to find security vulnerabilities, but rather, the ability of anyone to ask an LLM agent to rewrite your open source project in another language and thus work around whatever license your project has?
This is happening quite a lot actually. People just feed an existing project into their agent harness and have it regenerate more or less the same with a few tweaks and then they publish it.
I'm not sure how this works in the legal sense. A human could ostensibly study an existing project and then rewrite it from scratch. The original work's license shouldn't apply as long as code wasn't copy & pasted, right?
What happens when an automated tool does the same? It's basically just a complicated copy & paste job.
A lot of open source projects already have licenses that allow forking and selling the fork, it hasn't been a problem most of the time... there's a lot more to operating open source as a business beyond just shipping the code
> A lot of open source projects already have licenses that allow forking and selling the fork
If we go by the OSI's definition, a project that doesn't allow this is not "open source". So all open source projects -- not just "a lot" -- allow this.
Reading between the lines, it seems like they were working with cal.com and used red team bots to find vulnerabilities in cal.com’s code. And they probably found bugs a lot faster than cal.com could fix them. So the CEO balked at the estimated cost of fixing and took his ball home.
This article is effectively an announcement that cal.com is riddled with vulnerabilities, which should be easy to find in an archive of their code.
Alternatively those scanning tools have the same issue all other security scanners have in that they have too many false positives. And when tuning them to have only few false positives, they miss the true positives.
Then the real work is in investigating each false positive. Can still be useful compared to manual review, but requires real resources.
Meanwhile the flood of false positives causes reputation loss if not addressed. Reputation loss that closed source software does not get. Hence perhaps going closed source.
I decided to not open source my latest project but it has nothing to do with security concerns. My code is perfectly secure and bug-free.
My concern is mostly financial. Most people would be in a better position to monetize my software than I am... Using AI to obfuscate the origin while appropriating all the key innovations. I wouldn't get any credit.
Also, I'm not really interested in humans anymore. I have human fatigue.
have you like used AI? it's barely usable for anything other than grunt work. You actually have to tell the truth for the cynical but correct trope to work. If it's about now might be time to re-calibrate the hype, if it's about the future, well then there the worst case scenario is far worse than the worst case scenario cynics and LLM company CEOs talk about, human extinction is really the medium bad case if they actually manage to create AGI.
I mean, bold statement but statistically speaking it's almost certainly incorrect. I will say that, irrespective of whether source is open or closed, I would be deeply skeptical of a project that made this assertion.
I mean that's fair enough but I don't think one person keeping their code closed source really changes all that much. And it depends on what your software does, no-ones out there really replacing (while actually saving any money at least) anything complicated with vibecoding, and GPL violations already happened before.
I don’t believe for a second that the real reason is security by obscurity. They probably believe they can make more money not being open source and this sounds like a better excuse than ”we wanna make more money”.
Probably, and I don’t care and kinda wish they boldly said so too. It’s their product to do with what they want, they built it.
One of the ugliest parts of open source is people believing they’re entitled to you working for free forever. And instead of being thankful you gave years of your labor for free, people get angry at you for not continuing to do so forever. And try to shame you as if you’re somehow greedy if that changes.
Do you work exclusively pro-bono on open source projects? Or do you work a job where you only go in if you get paid?
I have a large open source project and noticed the number of LLM generate PR is making it unmanageable. Every two weeks, I go in, kill all of them and when someone complains or asks why, I realize it was a real person and then I merge it.
Yes, I "fixed" it by disabling pull requests on the repository. I'm still happy to pull from other people's branches (and do say so in CONTRIBUTING.md)
> Security testing has to become an automated, integral part of the CI/CD pipeline. When a developer opens a pull request, an AI agent should immediately attempt to exploit it. When infrastructure changes, an AI should autonomously validate the new attack surface. You do not beat automated attackers by turning off the lights; you beat them by running better automation on the inside.
This feels like the core of the article, but it doesn’t prove the need for open source.
I'll admit that I agree with a lot of the post but that I can't fully wrap myself around the cybersecurity situation today, is it basically:
-if code is open source or closed source, AI bots can still look for exploits
-so we need to use AI to develop a checklist program regardless to check for currently known and unknown exploits given our current state of AI tools
-we have to just keep running AI tools looking for more security issues as AI models become more powerful, which empowers AI bots attacking but also then AI bots to defensively find exploits and mitigate them
-so it's an ongoing effort to work on
I understand the logic of closing the source to prevent AI bot scans of the code but also fundamentally people won't trust your closed source code because it could contain harmful code, thus forcing it to be open source
Edit: Another thing that comes to mind is people are often dunking here on "vibe coding" however can't we just develop "standards / tools" to "harden" vibe coded software and also help guide well for decisions related to architecture of the program, and so on?
> Closing your source code does not stop an AI from probing your API or finding an authorization bypass in your webhooks.
I see this trope a lot in security discussions. “Obscurity isn’t security” or “since you can’t protect against X you may as well do Y”.
This is a harmful trope, which discourages perfectly good protections. Sure, closing source is not a perfect protection, but it is a defense against a large band of attacks.
Think of the entire field of potential vulnerability probes attackers have. Closing the source closes many of them off, likely most of them.
A pen-tester model with implementation will be loads more effective than one with only a black box. And that will give cal.com time to run the pen testing model on the source and address the vulns , hopefully before they are exploited.
I tested this myself, first using black box model attacks, secondly using the source code. The model with the source found and exploited the vulns instantly . The model without failed.
The lesson is: obscurity is not security ALONE, but it is a component of security.
First we blamed AI for layoffs, next we are blaming AI for the AI bait and switch.
It's entirely possible this CEO sincerely believes this, but that means you as a potential customer should stay away: now you know that the CEO of this company has no idea how technology works even at an executive level and/or that he doesn't consult his experts before making decisions.
That's literally not it, a CEO can know how technology work and not apply it for its management, many people do things they "dislike" or don't believe in everyday.
Well, that's what I mean, this guy is using this issue as an scapegoat to close source the software and increase revenues as a result.
The pipeline goes like this:
Use open source license to gain traction and credibility > establish a customer base > pull the rug on open source to get everyone who depends on your product but isn't yet paying to pay.
But... playing devil's advocate, if AI makes it very easy to find exploits without the source code, wouldn't it be doubly effective finding them with the source code as well? And why is the dichotomy posed by this blog post "open source with AI reviews by everyone" vs "closed source but only the bad guys use AI"? What if the scenario was: closed source and the authors/security team use every AI tool at their disposal to find bugs? What do the community's eyeballs add to this equation, assuming (big if) AI review of exploits is such a force multiplier?
Before any knee-jerk reactions: big fan of open source, I'm not arguing this will kill it, I don't have the faintest idea what Cal.com is and I think a world without FOSS would be a tragedy, I run linux and most of my software on my personal PC (other than games) is FOSS.
Pretty overreaching claim about another company's internal decisions and open source in general. There is a lot of incentive to stop open source these days.
One of which I am experiencing right now is somebody just copying my repo, not crediting me, didn't even try to change the README. It's pretty discouraging.
The other is security reasons, the premise that volunteers will report vulnerabilities really matter if you are big enough for small portion of people to dedicate themselves, for the most part people take open source tool use it and then forget about it, they only want stuff fixed.
Lastly, open source development kinda sucks so far. I'v been working on a few different tools and the amount of trolling and just bad faith actors I had to deal with is exhausting. On top of that there is a constant stream of people just demanding stuff to be fixed quickly.
The idea of tying source code to sustenance will soon be history. We will all remember the days when adding some few thousand smart lines of code meant you could gain notoriety and through cheap viral copy expand those traits to wealth and worth. But software has always just been zeros and ones, the value only happens when interpreted.
The future is sharing, you may not believe because your income is tied to being clever. Long term we are all more clever because of the sharing, and your contribution sometimes does not add to your personal success. Asking a company or its individuals to forego their success will not make them add more to our future. But they will add to our future nonetheless, because they all feel like we all do, that adding is what we are all meant to do.
Open Source was always open to "many eyes" in theory exposing itself to zero-day vulnerabilities. But the "many eyes" go for the good and the bad actors.
As far as I am concerned... Way to go Cal.com, and a good reminder to never use your services.
At the same time, I heavily support open-source and contribute a lot, but I can't necessarily agree that security-through-obfuscation doesn't play a major role in slowing down attacks. Cloudflare have based its whole security being closed-source (for example on its anti-bot mechanism) to be hard to reverse engineer, and they remain leaders as of today with few serious security breaches.
Some things just can't be truly secure as well, ddos protection is mostly a guessing/preventive game, exposing your firewall config/scripts will make you more vulnerable than NOT.
If your codebase isn't exposed, attackers are constrained by the network and other external restrictions which greatly reduce the number of possible trials, even with a swarm of residential proxies, it's not the same at all from inspecting a codebase in depth with thousand of agents and all models.
> In the past, exploiting an application required a highly skilled hacker with years of experience and a significant investment of time to find and exploit vulnerabilities. The reality is that humans don’t have the time, attention, or patience to find everything.
I read this as:
"We figured no one was looking so we just shipped unsafe garbage for years. We never once did an internal audit, never once paid a hacker to try to exploit our product, never thought we'd get caught with our substandard products."
If a guy in his basement with $200 dollars can ruin your company then you were trading on vapor the entire time. I'm sorry you had to find out this way.
I'm hopeful the article is right about its prediction, although I'm under the impression the attacker/defender dynamic is asymmetric and the defender on the loosing end. I hope someone can proof me wrong though...
Making the assumption that the same amount of money needed to attack a critical vulnerability is also required to find and fix it.
Lets say we have a project with 100 modules, and it costs us $100 000 to check these modules for vulnerabilities. What is stopping an attacker from spending the same amount of money to scan, lets say 10 modules but this time with 10x the number of tokens per module than the defender had when hardening the software?
Closing your source doesn't close your attack surface,it just closes the
community that would have helped you defend it. Security through obscurity
is a kind of tradeoff, not a strategy.. i mean that's what I feel.
It's a good question - is blackbox hacking as effective as whitebox hacking, for AI agents? I've gotta assume someone at Anthropic is putting together an eval as we speak.
I don't really know, but I have a story which might prompt some conversation about it.
At $WORK we had a system which, if you traced its logic, could not possibly experience the bug we were seeing in production. This was a userspace control module for an FPGA driver connected to some machinery you really don't want to fuck around with, and the bug had wasted something like three staff+ engineer-years by the time I got there.
Recognizing that the bug was impossible in the userspace code if the system worked as intended end-to-end, the engineers started diving into verilog and driver code, trying to find the issue. People were suspecting miscompilations and all kinds of fun things.
Eventually, for unrelated reasons, I decided to clean up the userspace code (deleting and refactoring things unlocks additional deletion and refactoring opportunities, and all said and done I deleted 80% of the project so that I had a better foundation for some features I had to add).
For one of those improvements, my observation was just that if I had to write the driver code to support the concurrency we were abusing I'd be swearing up a storm and trying to find any way I could to solve a simpler problem instead.
Long story short, I still don't know what the driver bug was, but the actual authors must've felt the same way, since when I opted for userspace code with simpler concurrency demands the bug disappeared.
Tying it back to AI and hacking, the white box approach here literally didn't work, and the black box approach easily illuminated that something was probably fucky. Given that AI can de-minify and otherwise spot patterns from fairly limited data, I wouldn't be shocked if black-box hacking were (at least sometimes) more token-efficient than white-box.
This seems to be extremely common. Been a very long time since I looked at Linux kernel stuff, but there were numerous drivers that disabled hardware acceleration or offloading features simply because they became unreliable if they were given heavy loads or deep queues.
I’m afraid the balance of open source wasn’t broken today. It happened quite a long time ago. It’s just not something people usually talk about. Companies have been using open source code for years to build paid products without giving anything back. Take PHP as an example: a language widely used across the internet, yet with a very limited budget.
Strix was so close to being the hero we deserve. I think these blue torches like strix should offer their services for free to open source ships out at sea. There are 3 wins here, GLOBAL GOOD WILL, testimonial and reviews, and market loyalty reward.
Can any of the AI systems read binary yet? Perhaps generate source code from object file? Is so, that would make access to source redundant for that type of analysis.
AI assisted decompiling has been a thing for a while now, from what I know most people are using assisted tooling for it.
With that said it at least seems possible to be able to be able to read binary itself, but most of the magic there is in execution, so you'd have to have an LLM behave kind of like a processor I think.
I mean "yes but" lots from 2015 and before, on a scholarly paper search engine. (I do not have access to most, but there are some public ones)
I mean as a convention when dealing with cryptography, so far the only organization that has succeeded in doing closed-source cryptography securely, has been the USA's "NSA", and mostly their algorithms are public.
I mostly work in the closed source world, however my observation from all the code bases I've seen is "mostly open source are more secure", except when very thorough following of formal security specifications are followed, and then security is as good as the specifications. (YMMV there, of course).
feels like the real shift is not open vs closed, but reaction time AI attackers don’t need perfect access anymore, just enough surface and time. So the question becomes:
can you detect+respond faster than they can iterate in that sense, open source might even help -more eyes reduces time to fix, not just time to find
every line of code is a liability. open, closed, doesn't matter. companies will have to treating it that way--which means actual engineering--or they will get burnt, and hard.
which is why am I happy to launch the first AI-first meeting scheduling app, yes it costs 5000 USD/month but we have almost completely eliminated any use of code in our application, offering enterprise quality security by having Claude (TM) Opus (TM) 4.8 (TM) directly reply to all HTTPS requests - end to end AI driven agentic software is finally here.
Of course this neglects why mostly free things that were posted on the internet generally won. Take Microsoft for example. All their money makers are licensed, yet at the same time you can download almost every single one for free and install it.
The people that go behind paywalls don't realize how much they'll have to spend on marketing to catch up to those that are open.
And that's only frames the current state, where models are very expensive to train. Once model training is close to the point where a group of individuals can afford it, it's pretty much game over for our current paradigm. The software police will be running around trying to play whack-a-mole on open weight models with people all over the world.
Why would I create content that I don't get paid for and I don't even get credit for? Everyone who creates free content right now is simply doing the work of AI companies to make them more useful for free.
Search engines will cease to exist, so no one will search your content and then click on your link. AI will simply regurgitate your content and take the money for tokens or subscription and not acknowledge you at all.
>There isn't a rule of economics that says better technology makes more, better jobs for horses. It sounds shockingly dumb to even say that
--Humans need not apply.
It's kind of funny that you think you're going to be making money writing software. If you lock up your software who exactly are you selling it to anyway? It's like you're thinking 25% through the situation then going "I can stay where I am and I don't have to change anything", and then crying later when it doesn't work.
What are you going to do, advertise in BYTE magazine (dead). On Instagram? With a sandwich board on a Seattle street corner? What does the software market even look like in the AI age.
And much like how Google and Amazon eat your lunch now whenever they way, successful AI companies will buy up some software ideas and feed them to their models (which will be stolen later by other models). Anyone that sees your software will mock up a useful clone of it pretty quick the first time they see it. And foreign AI companies will just right out steal it.
You're right you won't create content that you don't get paid for, you just won't be creating anything while competing with the other unemployed masses for strawberry picking jobs.
buying into the marketing I see? Might as well making stealing easier just cause they will probably manage to steal everything in the end is an interesting thought.
Like yes if the AI thing works out the theft will be retroactively okay because the people that stole will now be the ruling class, but assuming AI models will write good code one day and do all work, when that hasn't happened yet is just repeating corporate propaganda while presuming being convinced you are some sort of cynical truth tell that's telling it how it is.
I don't think this will happen. If most content goes behind a paywall, releasing content for free will again become a valuable source of attention. It used to be so before the web got filled with so much free content that it lost any value.
> Today, Cal.com announced they are transitioning their core codebase away from open source. The reasoning provided by their CEO, Bailey Pumfleet, is that AI has automated vulnerability discovery at scale, making code scanning and exploitation "near zero-cost". In this new world, they argue, "transparency becomes exposure."
Laughable and hilarious. Extremely short sighted. I can show code generated by Claude Opus 4.6 at the highest compute intensity that lacks even basic checks in input validation that was clearly provided in the spec.
There's no point in arguing with crypto and AI bros. They are the same tribe. AI crowd however might learn their lessons sooner because the universe isn't forgiving or flexible.
Note: I use AI code generators all the time but I take them as very very dumb transpilers no matter how expensive their input/output pricing it and I learned that hard way.
Which works if you assume that AI can find 100% of your bugs.
It can't. So this is a complete waste of your time and will hide actual bugs behind a layer of confidence _and_ obscurity.
You're going to actually have to sit down and figure out how to provide real security in your product while earning profits. This is called "work." I understand Silicon Valley would like to earn money and not work. I am eager for these people to get their comeuppance.
Open Source as such will never "die", but we only need to look at what happened in, say, the last 5 or 10 years. Private entities with a commercial interest, have been flexing their muscles. Microsoft - also known as Microslop these days - with Github is probably the most famous example still, but you can see other examples. One that annoys me personally is Shopify's recent influence - rubygems.org is basically just shopifygems.org now. See: https://blog.rubygems.org/2026/04/15/rubygems-org-has-a-publ...
"Contributors from both the RubyGems client team and Shopify are already working with us on making native gems a better experience for the Ruby community. "
There is a lot more I could add to this (see my complaint about how rubygems.org hijacks gems past the 100.000 download barrier; this was why I retired from using rubygems.org, and then the year afterwards ruby core purged numerous developers. The handwriting is soooooo clear that shopify flexed their muscles here).
I think we need to make open source development more accessible to everyone, not just corporations throwing their money to gain influence and leverage. I don't have a great idea to make this model work; economic incentives kind of have to be there too, I get that part, and I am not sure which models could work. But right now we really have a big problem. We can also see this with age sniffing (age verification - see the article that pointed at Meta at orchestrating influence and lobbyism) and many more changes. Something has to change. Hopefully some people cleverer than me can come up with models that are actually sustainable, even if it may not necessarily be a "fund an open source developer for a year". There could be a more wide-spread "achieve xyz" or some other lower finance effort - but again, I don't have a good suggestion here. Hopfully something improves here though, because I am getting really tired of private interests constantly sabotaging and ruining the whole ecosystem while claiming they do "improve" an ecosystem. We have the old "War is peace. Freedom is slavery. Ignorance is strength." going again. Opposite day, every day.
There is another product I use that has a freemium model. They hope to monetize a paid tier for users who use the product a lot.
In order to build trust, they open source their product. I forked it, removed the blocks from the freemium feature in 15 minutes using Claude Code. Never published the code to anyone else, just used it myself
Unfortunately, I think it isn’t going to be tenable for systems to be fully open sourced going forward.
Open source is one of the main reasons I recommended cal.com to everyone — I just did so yesterday again in fact!
I'm disappointed to hear this especially since I don't think the rationale makes sense, from what I understand of the security landscape, and it also makes me a little more skeptical of cal.com in general.
I agree, it is shortsighted (next quarter syndrome). First of all the AI does not need source to find vulnerabilities and further it breaks the unwritten contract to exchange source for eyeballs which creates better source. I guess the CEO wants less security and stopped evolution of it's code.
I can't believe we still have people out there buying this baby-brain idea of "If muh code is open than people will find vulns!!" This has been disproven for 20+ years catch up.
AI generated bullshit PRs are clearly the bigger issue in the OSS space.
> We are an open-source project ourselves, recently crossing 24k stars, and our framework processes over 15 billion LLM tokens daily to find software vulnerabilities.
I'm sorry, but talking Github stars comes off as borderline pathetic.
Open source is dead, AI-pundits are applying the wrong lessons. No one has to accept AI or play the game all these AI companies don’t work if everyone stops publishing. Let the AI generated content industry have the publish space, they're very adamant about taking it over and watering it down with slop.
I wrote some very nice expressive text for our deployment guide. My project manager took the guide and had Gemini break it down into plain boring bullet points. AI and the pundits can gf themselves in their journey to kill human expression.
Here is what I wrote in the guide:
"Post Deploy Responsibility
If you made it this far, say “Wow I really did it and it was so easy!”
Did you say it? Good. Now you are entirely responsible for any issues or bugs that may arise from the newly deployed code. Don’t go anywhere until the deploy has finished (usually takes a few minutes). While an issue or bug may not leave you directly at fault, you are responsible for coordinating any rollbacks or remediations that may be needed until the next deploy."
Here is what the product manager slopped it into:
"- Post deploy responsibility
- You are responsible for performing QA upon deployment
- You are responsible for any issues or bugs that may arise from newly deployed code
- You are responsible for coordinating any rollbacks or remediations that may be needed until the next deploy"
My paragraph wasn't long, hard to understand, or poorly written. I wouldn't have objected to a rewording or some changes but the project manager chose to just copy paste it into Gemini and copy and paste it back. So my take is that they didn't understand what I wrote. Which is a few sentences long and frankly sad if a paragraph is too intense for you to read. When my project manager did this during the meeting I said, "RIP human expression" and their response was a very hasty "no that's not what's happening". This is what all the pundits want to do to everyone and society. Don't believe them that "it's just a tool", that is just a tactic to get you to rollover so they can shove more AI in your face.
there seems to be a certain type of person that thinks if an LLM can generate content faster than means it's better, said belief appears to also be linked with thinking that if text is long it should be broken down so it's easier to understand, it would seem a bit contradictory considering LLM generated code is usually more difficult to correctly review than code a human wrote, but at the core of it might be a poor understanding of systems, those bullet-points are going to get skimmed over or read but not understood, not everything can be easy.
And your paragraph had a much bigger impact on the reader. Your paragraph reads like an experienced senior developer teaching you to not screw things up, while the AI generated bullet points sound like generic ToS that everyone ignores.
Thank you, my intent was to write something memorable but also calm any exterior worries about deploying code and the feeling of “being in trouble because prod is broked”.
Enshittification has come for VC backed open-source. As someone on Twittter said, open source has deemed commercial open source obsolete especially when users can point Calude Code to calcom on GitHub and ask it to make them scheduling features directly into their product. That’s what spooked Cal.
Don’t get me wrong but if virtually all modern software infrastructure lives on top of open source and they’re mostly fine then I’d imagine that you can make a scheduling webapp secure independent to if it’s OSS or not.
It’s OK if there’s another reason for this transition, just be transparent about it and don’t treat your users as children.
They don’t owe you a complete list of reasons why they’re close sourcing their software. They are not a publicly traded company and no one (customers) actually cares if the product is open source or not.
No-one cares if a product is closed-source, a lot of people do care if a product is open source, and really unless a product is truly unique (almost no product is) that is often the main selling point.
Open source software is just a method for corporations to get free evangelization of their product. Companies are now saying “with the assault of AI it is no longer valuable to keep the code open.” They don’t need the handful of Twitter addicted software devs that will evangelize for them anymore because they lose too much by publishing the code. For small companies, the product is most likely the extent of their brand and offerings. Not worth the risk of going out of business, especially when competitors probably aren’t open source.
I've always used and advocated for Cal.com because it's open source. I understand you need to make money and this is no longer the GTM, but don't lie about it.
I have an open source project and started receiving a lot of security vulnerability reports in the last few months. A lot of them are extremely corner cases, but there were some legit ones. They're all fixed now. Closed source software won't receive any reports, but it will be exploited with AI. So I definitely agree with the message of this article.
> Closed source software won't receive any reports, but it will be exploited with AI.
What makes you so sure that closed-source companies won't run those same AI scanners on their own code?
It's closed to the public, it's not closed to them!
As someone who works on closed source software and has done for a couple of decades, most companies won't even know about that and of those who do only a fraction give enough of a shit about it to do anything until they are caught with their pants down.
Seconded.
Having worked in quite a few agency/consultancy situations, it is far more productive to smash your head against a wall till bleeding, than to get a client to pay for security. The regular answer: "This is table stakes, we pay you for this." Combined with: "Why has velocity gone down, we don't pay you for that security or documentation crap."
There are unexploited security holes in enterprise software you can drive a boring machine through. There is a well paid "security" (aka employee surveillance) company using python2.7 (no, not patched) on each and every machine their software runs on. At some of the biggest companies in this world. They just don't care for updating this, because, why should they. There is no incentive. None.
Yea, its fundamentally an issue of asymmetric economics.
Running AI scanners internally costs money, dev time, and management buy in to actually fix the mountain of tech debt the scanners uncover. As you said there is no incentive for that
But for bad actors the cost of pointing an LLM at an exposed endpoint or reverse engineered binary has dropped to near zero. The attackers tooling just got exponentially cheaper and faster, while the enterprise defenders budget remained at zero.
In theory though, there is now a new way for community to support open source, but running vulnerability scans in white-hat mode, reporting and patching. That way they burn tokens for a project they love. Even if they couldn't actually contribute code before.
There should be a way to donate your unused tokens on every cycle to open source like rounding up at the chekout!
That sounds like a great idea. I'd love to be able to contribute the remainder of my monthly AI subscriptions for something like this, especially since some of them bill and refresh their quotas by calendar month.
Hang on, why is it costly for in-house to run AI scanners but near zero for threat actors to do the same?
I've seen multiple proprietary places now including a routine AI scan of their code because it's so cheap and they may as well use-up unused tokens at the end of the week.
I mean, it's literally zero because they already paid for CC for every developer. You can't get cheaper than that.
If a company specifically doesn't have a dedicated security team (or even a person), this will never get done.
Most software companies sadly don't hire a dedicated (software) security expert.
Yup, closed source software is a huge pile of shit with good marketing teams. Always was.
As I mentioned above, we actually do run these AI scanners on our code, but the problem is it's simply not enough. These AI scanners, including STRIX, don't find everything. Each scanning tool actually finds different results from the other, and so it's impossible to determine a benchmark of what's secure and what's not.
> As I mentioned above, we actually do run these AI scanners on our code, but the problem is it's simply not enough. These AI scanners, including STRIX, don't find everything. Each scanning tool actually finds different results from the other, and so it's impossible to determine a benchmark of what's secure and what's not.
Yeah, but with closed source it's cheaper for the defender than for the attacker - the defender can scan their sources and their PRs as well as the compiled output. The attacker can only scan the compiled output, and they have to perform repeated scans.
I think it makes it all the more apparent that writing EAL4 code with as little design competence as possible was taking advantage of some strange scarcity economics.. It's now even easier to make something with endless technical debt and security vs backwards compatibility liability but is anyone going to keep paying for things that aren't correct and to the point if some market participants structure their agent usage toward verifiable quality and don't actually have more cost any more?
> What makes you so sure that closed-source companies won't run those same AI scanners on their own code?
How many companies take the time to use penetration testing tools, that have been available for many years, to verify their software (or pay a penetration testing company to do a more thorough job than they have the experience to do internally)?
Some, certainly. Many, possibly. Most, I would wager not.
The economic motivation just simply isn't there. I'm sure we could cherry pick a few examples of companies where things like quality and security really are part of the culture, not just feel-good lip services. The reality is that companies are in business to make money and corner cutting is the easiest way to pad the margins.
More eyes, more chances that someone will actually use the tools. Also, the tools and how you use them are not all the same.
With enough copies of GPT printing out the same bulleted list, all bugs are
1. shallow
2. hollow
3. flat
...
Because they're a company. Even if the bar to entry can fit a normal sized american, doesn't mean they will do it, or do it in a systematic way; We know very well that nothing about AI is naturally systematic, so why would you assume it'll happen in a systematic way.
Came here to say the same. Same tools + private. In security two different defense-mechanisms are always better than one.
Same tools A, B and C, but minus tools D, E and F, and with a smaller chance that any tools at all will even be used.
Not claiming that it's a slam dunk for open source, but the inverse does not seem correct either.
> Same tools A, B and C, but minus tools D, E and F,
Why "minus D, E and F"? After all, once you have the harness set up, there's no additional work to add in new models, right?
The point being that there are always going to be more eyes, and more knowledge of available tools (i.e. including "D, E and F"), and more experience using them, with open source than with a single in-house dev team.
There's no more "eyes" though, it's all models, and they are all converging pretty damn fast.
If true then logically it will be sufficient to run this "master model" once before any code release for the level playing field to be restored. After all, even open-source software is private until it is released.
> If true then logically it will be sufficient to run this "master model" once before any code release for the level playing field to be restored.
I'm struggling to see how it is a level playing field:
1. Closed-source: defender runs llms to check the sources for vulns, runs llms on each PR, runs llm on deployment of the compiled output. Attacker runs llm only on compiled output.
2. Open-source: both attacker and defender runs llms on source, on PRs and on compiled output.
Fair enough
[dead]
> Closed source software won't receive any reports
Not from the automated repo scanners, but bug bounty programs can generate a lot of reports in my experience. AI tools are becoming a problem there, too, because amateurs are drawn to the bounties and will submit anything the AI hallucinates.
Closed source companies can (and should!) also run their own security audits rather than passively waiting for volunteers to spend their tokens on it.
Those bug bounty programs now have to compete against the market for 0-days. I suppose they always did, but it seems the economics have changed in the favour of the bad actors - at least from my uninformed standpoint.
That still exists in the OSS world too, having your code out there is no panacea. I think we'll see a real swarm of security issues across the board, but I would expect the OSS world to fare better (perhaps after a painful period).
Of course everyone should do their own due diligence, but my point is mostly that open source will have many more eyes and more effort put into it, both by owners, but also community.
But also tools that might not be nice and report security vulnerabilities, but exploit them.
There is no guarantee that open means that they will be discovered.
That's absolutely our plan. We have bug bounty programs, we have internal AI scanners, we have manual penetration testing, and a number of other things that enable us to push really hard to find this stuff internally rather than relying on either the good people in the open source community or hackers to find our vulnerabilities.
+1, at this point all companies need to be continuously testing their whole stack. The dumb scanners are now a thing of the past, the second your site goes live it will get slammed by the latest AI hackers
> Not from the automated repo scanners, but bug bounty programs can generate a lot of reports in my experience. AI tools are becoming a problem there, too, because amateurs are drawn to the bounties and will submit anything the AI hallucinates
So just like a pre-AI or worse?
Worse. [0]
[0] https://hackerone.com/reports/3595764
You don't even need a bug bounty program. In my experience there's an army of individuals running low-quality security tools spamming every endpoint they can think (webmaster@ support@ contact@ gdpr@ etc.) with silly non-vulnerabilities asking for $100. They suck now but they will get more sophisticated over time.
I don't follow. It seems obvious that there's more to gain for attackers using AI agents to exploit open source repositories, than there is for good samaritan defenders. In this new closed-source world (for Cal.com), there's nothing stopping them from running their own internal security agent audits, all whilst at least blocking the easiest method of finding zero-days - that is, being open source.
This really just seems like Strix marketing. Which is totally fair, but let's be reasonable here, any open-source business stands to lose way more by continuing to be open-source vs. relying on the benevolence of people scanning their code for them.
> It seems obvious that there's more to gain for attackers using AI agents to exploit open source repositories, than there is for good samaritan defenders.
Actually the opposite is obvious - the comment you replied too talked about an abundance of good Samaritan reports - it's strange to speculate on some nebulous "gain" when responding to facts about more then enough reports concerning open source code.
> In this new closed-source world (for Cal.com), there's nothing stopping them from running their own internal security agent audits
That's one good Samaritan for a closed source app vs many for an open source one. Open source wins again.
> any open-source business stands to lose way more
That doesn't make any sense - why would it lose more when it has many more good Samaritans working for it for free?
You seem to forget that the number of vulnerabilities in a certain app is finite, an open source app will reach a secure status much faster than a closed source one, in addition to also gaining from shorter time to market.
In fact, open source will soon be much better and more capable due to new and developing technological and organizational advancements which are next to impossible to happen under a closed source regime.
The main drawback is that you will need to be able to patch quick in the next 3-5 years. We are already seeing this in a few solutions getting attention from various AI-driven security topics and our previous stance of letting fixes "ripen" on the shelf for a while - a minor version or two - is most likely turning problematic. Especially if attackers start exploiting faster and botnets start picking up vulnerabilities faster.
But at that point, "fighting fire with fire" is still a good point. Assuming tokens are available, we could just dump the entire code base, changesets and all, our dependent configuration on the code base, company-internal domain knowledge and previous upgrade failures into a folder and tell the AI to figure out upgrade risks. Bonus points if you have decent integration tests or test setups to all of that through.
It won't be perfect, but combine that with a good tiered rollout and increasing velocity of rollouts are entirely possible.
It's kinda funny to me -- a lot of the agentic hype seems to be rewarding good practices - cooperation, documentation, unit testing, integration testing, local test setups hugely.
It's a token game.
Let's say finding a security issue takes 10M tokens. If one company has to pay for it, they most likely won't bother. It's purely a cost/benefit thing for them.
But if you have an open source project, you might get a 1000 people looking at it, each only has to spend 10k tokens to find the same flaws.
A new user is much more likely to scan the codebase and report vulnerabilities so they can be fixed than illegally exploit them since most people aren't criminals
Exactly. Who even hacks stuff? Most people will report the issue to earn xp and level up than actually exploit it.
Some users might be tech sensitive and have the capacity to check the codebase If a company want to use your platform, it can run an audit with its own staff These are people really concerned about the code, not "good samaritans"
Isn’t that security by obscurity?
I’ve recently set up nightly automated pentest for my open-source project. I’m considering starting to publish these reports as proof of security posture.
If the cost of security audit becomes marginal, it would seem reasonable to expect projects to publish results of such audits frequently.
There’s probably a quite hefty backlog of medium- and low-severity issues in existing projects for maintainers to suffer through first though.
What do you use for the pentests? any oss libraries?
This is a sandbox escape pentest so the only tooling needed is Claude Code and a simple prompt that asks it to follow a workflow: https://github.com/airutorg/airut/blob/main/workflows/sandbo...
We actually run AI scanners on our code internally, so we get the benefit of security through obscurity while also layering on AI vulnerability scanning, manual human penetration testing, and a huge array of other defence mechanisms.
"Security through obscurity" is a term popularized entirely by the long-standing consensus among security researchers and any expert not being paid to say otherwise that this is a bad idea that doesn't work
> Closed source software won't receive any reports, but it will be exploited with AI.
This is what worries me about companies sleeping on using AI to at a bare minimum run code audits and evaluate their security routinely. I suspect as models get better we're going to see companies being hacked at a level never seen before.
Right now we've seen a few different maintainers for open source packages get hacked, who knows how many companies have someone infiltrating their internal systems with the help of AI because nobody wants to do the due dilligence of having a company do security audits on their systems.
Assembly is still source code so really it comes down to if the copy protection is obscuring the executable code to the point where the LLM is not able to retrieve it on its own. And if it can't someone motivated could give it the extra help it needs to start tracing how outside inputs gets handled by the application.
given what the clankers can do unassisted and what more they can do when you give them ghidra, no software is 'closed source' anymore
Guess that kind of depends on your definition of "source", I personally wouldn't really agree with you here.
absolutely agree with you if we're talking about clean room reverse engineering; but in the context of finding vulnerabilities it's a completely different story
I mean-- to an LLM is there really any difference between the actual source and disassembled source? Informative names and comments probably help them too, but it's not clear that they're necessary.
Which models have you had good luck with when working with ghidra?
I analyze crash dumps for a Windows application. I haven't had much luck using Claude, OpenAI, or Google models when working with WinDbg. None of the models are very good at assembly and don't seem to be able to remember the details of different calling conventions or even how some of the registers are typically used. They are all pretty good at helping me navigate WinDbg though.
i agree with his too,
but with cal.com i dont think this is about security lol
open source will always be an advantage just you need to decide wether it aligns with you business needs
> Closed source software won't receive any reports, but it will be exploited with AI
How so? AI won't have access to the source code. In some cases AI may have access to deployed binaries (if your business deploys binaries) but I am not aware that it has the same capabilities against compiled code than source code.
But in a SAAS world, all AI has access to is your API. It might be still be up to no good but surely you will be several orders of magnitude less exposed than with access to source code.
Claude is already shockingly good at reverse engineering. Try it – it's really a step change. It has infinite patience which was always the limited resource in decompiling/deobfuscating most software.
It's SaaS though. You don't have access to the binary to decompile. There's only so much you can reverse-engineer through public URLs and APIs, especially if the SaaS uses any form of automatic detection of bot traffic.
Thanks you. This is what the parent post was trying to say. Don't know why it is down-voted. AI or not, if the API end points are well secured, for example use uuid-v7, then their is little that the ai can gain from just these points.
The opposite is true. Open source barely matters to attackers, especially ones that can be automated. It mostly enables more people (or agents, or people with agents) to notice and fix your vulnerabilities. Secrecy and other asymmetries in the information landscape disproportionately benefit attackers, and the oft-repeated corporate claim that proprietary software is more secure is summarily discounted by most cybersecurity professionals, whether in industry or academic research. This is also seldom the motivation for making products proprietary, but it's more PR-friendly to claim that closing your source code is for security reasons than it is to say that it's for competitive advantage or control over your customers
Yes exactly! I'm so glad I took this route with my startup. We can't bury our heads in the sand and think the vulnerabilities don't exist just because we don't know about them.
This might be the most painfully obvious advertisement I’ve ever seen on a forum.
I didn't mean it as such, but I can see why it would seem so. I've edited the link out now. Thanks for the feedback.
> The reasoning provided by their CEO, Bailey Pumfleet, is that AI has automated vulnerability discovery at scale,
That sounds like an excuse. The real reason is probably that it's hard to make a viable business out of developing open source.
Exactly. I respect their decision to go closed source if that's what they need to do to make it a viable business, but just be honest about it. Don't make up some excuse around security and open source.
I don't know if I fully agree with this -- how many people were actually self-hosting cal infra? I def could be wrong though
The issue isn’t would-be customers going to the trouble of self hosting to save a measly $30/month.
The issue is competitors popping up to clone your offering with your own codebase.
it's not necessarily about people self hosting it, it's about people preferring to pay for hosted stuff that is open source (e.g. I pay for Plausible).
Now it's a lot easier to rewrite open source stuff to get around licensing requirements and have an LLM watch the repo and copy all improvements and fixes, so the bar for a competitor to come along and get 10 years of work for free it a lot lower.
[flagged]
We've run an extremely profitable business for five years, raised a seed and a Series A, and grown at 300% a year sustainably while being open source.
Going closed source actually hurts our business more than it benefits it. But it ultimately protects customer data, and that's what we care about the most.
I think if it ultimately protects customer data in a significant way, I would be for it.
Are you able to share any more detail on how you determined this is the best route? It would be a significant implication for many other pieces of open source software also if so.
(And I say this is someone who just recommended cal.com to someone a few days ago specifically citing the fact that it was open source, that led to increased trust in it.)
I did find the video valuable, for reference for others: https://www.youtube.com/watch?v=JYEPLpgCRck
I think if you are committed to switching back to open source as soon as the threat landscape changes, and you have some metric for what that looks like, that would be valuable to share now.
I would like to see the analysis that you're referencing around open source being 5-10x less secure.
By this logic, Linux should switch to closed-source.
All your servers are Linux, so imagine how insecure you are - must switch to windows ASAP.
AI makes a great scapegoat. Need to lay off people? "AI." Need to switch to closed source? "AI."
It’s also now ridiculously easy to simply cherry pick from open source without actually “using” it.
“I need to do foo in my app. Libraries bar and baz do these bits well. Pick the best from each and let’s implement them here”
I’d not be surprised if npmjs.com and its ilk turn into more a reference site than a package manager backend soon.
I literally have a Claude Code skill called "/delib" that takes takes in any nodejs project/library and converts it to a dependency-less project only using the standard library.
It started as a what-if joke, but it's turned out to be amazing. So yeah, npmjs.com is just reference site for me now, and node_modules stays tiny.
And the output is honestly superior. I end up with smaller projects, clean code, and a huge suite of property-based tests from the refactor process. And it's fully automatic.
It's that easy yes, and someday, we will literally be able to prompt "Redo the Linux kernel entirely in Zig" and it will practically make a 1:1 copy.
Interesting - I am interested to know how’s it impacting the codebase size interms of lines of code.
It varies from project to project, but applications benefit a lot more than libraries. When I de-lib a normal express app it might add a few hundred lines of code and a few thousand new tests, but if I de-lib an library then depends on how ancient it is. The older the library is, the higher the chances that most of what it needs is built-in to the standard library.
Thank you. I have been thinking about the same approach. However my worry is the open source libraries often gets more eyeballs and CPU cycles and ends up much more refined over a period of time.
Ironically, given the recent supply chain attacks, that may be also more secure.
separating codebase and leaving 'cal.diy' for hobbyists is pretty much the classic open-core path. the community phase is over and they need to protect their enterprise revenue.
blaming AI scanners is just really convenient PR cover for a normal license change.
Yes. Before AI the source was a demonstration of your substance. Users would be encouraged to reach out to maintainers to pay for upgrades or custom tweaks or training. Or indirectly pay for advertising while reading docs. After AI those revenue streams have collapsed. Now you have to withdraw enough of the work to make it hard for an individual to recreate with an LLM. The open source needs to be restricted to a rich interaction layer. Cloudflare just announced they are using that model with their services which were already closed source but now they are exposing them through new APIs. So they can capitalize on existing services that were not ripe enough for SaaS before AI, that had to be handled by their in-house professionals services folks. With this move they are using AI to expand/automate their white glove professional services business to smaller customers.
I'd think it's also much easier to spin up a (in some area) slightly better clone and eat into their revenue.
This is part of it for sure. It is also true that many open source business depended on it not being worth the trouble to figure out the hosting setup, ops etc, and the code. Typical open source businesses also make a practice of running a few features back on the public repo.
Now I can take an open source repo and just add the missing features, fix the bugs, deploy in a few hours. The value of integration and bug-fixing when the code is available is now a single capable dev for a few hours, instead of an internal team. The calculus is completely different.
Yes, it feels like they've been looking for an excuse to go closed-source, and this one is plausible enough to make it sound like they're only doing it because they "have to".
What an uncharitable take
I mean, it's hard to make a viable business regardless of if the tech is OSS or not, but it's often seen as more challenging this way.
Brilliant piece of content marketing:
1) Pulls you in with a catchy title, that at first glance seems like a dunk on Cal.com (whatever that is).
2) Takes the "we understand your pain" approach to empathize w/ Cal.com, so you feel like you're on the good vibes side.
3) Provides a genuine response to the actual problem Cal.com is dealing with. Something you can't dismiss out of hand.
4) But in the end of the day, the response aligns perfectly with the product they're promoting (a click away to the homepage!)
This mix of genuine ideas and marketing is quite potent. Not saying this is all bad or anything, just found it a bit funny. The mixed-up-ness is the point!
That's exactly how this read to me too. Ultimately, the whole article is written by a company that does AI vulnerability scanning, and it's to try and get you to sign up for their service.
As it mentions in their article, Strix actually scans the Cal.com codebase and reports vulnerabilities to us. But the reality is, they actually miss so many vulnerabilities that other platforms do find. There's no one platform that seems to be able to reliably find all vulnerabilities, and so simply adopting AI scanners just isn't enough.
I'm sad to see this article being so upvoted while being kind of empty.
The real content could fit in a comment.
Is it good marketing though? I mean personally I do not use AI, and I don't think this opinion of mine will change. I can't look into the future, but right now I don't use nor do I depend on AI. I guess it may work for some people, but even then I am unsure whether that is really good marketing. Riding on a hype train (which AI right now still is) is indeed easier, so that has to be considered.
They are in HN front page, therefore it’s good marketing.
>Security through obscurity is a losing bet against automation
Security through obscurity is only problematic if that is the only, or a primary, layer of defense. As an incremental layer of deterrence or delay, it is an absolutely valid tactic, with its primary function being imposing higher costs on the attacker.
As such if, as people are postulating post-Mythos, security comes down to which side spends more tokens, it is an even more valid strategy to impose asymmetric costs on the attacker.
"With enough AI-balls (heheh) all bugs are shallow."
From a security perspective, the basic calculus of open versus closed comes down to which you expect to be case for your project: the attention donated by the community outweighs the attention (lowered by openness) invested by attackers, or, the attention from your internal processes outweighs the attention costs (increased by obscurity) on attackers. The only change is that the attention from AI is multifold more effective than from humans, otherwise the calculus is the same.
I wonder whether cal actually has concerns about security (in which case, they're wrong, this argument was false when people made it decades ago), or whether they just took a convenient excuse to do something they wanted to do anyway because Open Source SaaS businesses are hard.
Isn’t the real danger now not the ability to find security vulnerabilities, but rather, the ability of anyone to ask an LLM agent to rewrite your open source project in another language and thus work around whatever license your project has?
This is happening quite a lot actually. People just feed an existing project into their agent harness and have it regenerate more or less the same with a few tweaks and then they publish it.
I'm not sure how this works in the legal sense. A human could ostensibly study an existing project and then rewrite it from scratch. The original work's license shouldn't apply as long as code wasn't copy & pasted, right?
What happens when an automated tool does the same? It's basically just a complicated copy & paste job.
> A human could ostensibly study an existing project and then rewrite it from scratch.
And likely there would be enough similarities that the rewrite would be considered a derived work under copyright law.
> The original work's license shouldn't apply as long as code wasn't copy & pasted, right?
You don't need to do a literal copy & paste for it to be copyright infringement.
> What happens when an automated tool does the same? It's basically just a complicated copy & paste job.
Sounds like copyright infringement to me.
A lot of open source projects already have licenses that allow forking and selling the fork, it hasn't been a problem most of the time... there's a lot more to operating open source as a business beyond just shipping the code
> A lot of open source projects already have licenses that allow forking and selling the fork
If we go by the OSI's definition, a project that doesn't allow this is not "open source". So all open source projects -- not just "a lot" -- allow this.
You can do the same for closed source projects.
There are real limitations of course.
Reading between the lines, it seems like they were working with cal.com and used red team bots to find vulnerabilities in cal.com’s code. And they probably found bugs a lot faster than cal.com could fix them. So the CEO balked at the estimated cost of fixing and took his ball home.
This article is effectively an announcement that cal.com is riddled with vulnerabilities, which should be easy to find in an archive of their code.
Alternatively those scanning tools have the same issue all other security scanners have in that they have too many false positives. And when tuning them to have only few false positives, they miss the true positives.
Then the real work is in investigating each false positive. Can still be useful compared to manual review, but requires real resources.
Meanwhile the flood of false positives causes reputation loss if not addressed. Reputation loss that closed source software does not get. Hence perhaps going closed source.
I decided to not open source my latest project but it has nothing to do with security concerns. My code is perfectly secure and bug-free.
My concern is mostly financial. Most people would be in a better position to monetize my software than I am... Using AI to obfuscate the origin while appropriating all the key innovations. I wouldn't get any credit.
Also, I'm not really interested in humans anymore. I have human fatigue.
>My concern is mostly financial.
Then AI will eat your lunch anyway if the financial part has anything at all to do with the code.
AI can decompile code very well.
have you like used AI? it's barely usable for anything other than grunt work. You actually have to tell the truth for the cynical but correct trope to work. If it's about now might be time to re-calibrate the hype, if it's about the future, well then there the worst case scenario is far worse than the worst case scenario cynics and LLM company CEOs talk about, human extinction is really the medium bad case if they actually manage to create AGI.
> My code is perfectly secure and bug-free.
I mean, bold statement but statistically speaking it's almost certainly incorrect. I will say that, irrespective of whether source is open or closed, I would be deeply skeptical of a project that made this assertion.
I assumed they were trying to be humorous . Although I find that type of humour obnoxious enough that it would put me off the project.
I gave it a good minute of reading and re-reading because I thought it SURELY was meant tongue in cheek, but I couldn’t make it work.
Maybe I was being too generous - jongjong seems cynical and old enough but can't read similar "humour" from others: https://news.ycombinator.com/item?id=47426320
I previously failed to summarise HN guidelines on sarcasm: https://news.ycombinator.com/item?id=38585465
Humans are fine, the problem is your worth.
I mean that's fair enough but I don't think one person keeping their code closed source really changes all that much. And it depends on what your software does, no-ones out there really replacing (while actually saving any money at least) anything complicated with vibecoding, and GPL violations already happened before.
I don’t believe for a second that the real reason is security by obscurity. They probably believe they can make more money not being open source and this sounds like a better excuse than ”we wanna make more money”.
Probably, and I don’t care and kinda wish they boldly said so too. It’s their product to do with what they want, they built it.
One of the ugliest parts of open source is people believing they’re entitled to you working for free forever. And instead of being thankful you gave years of your labor for free, people get angry at you for not continuing to do so forever. And try to shame you as if you’re somehow greedy if that changes.
Do you work exclusively pro-bono on open source projects? Or do you work a job where you only go in if you get paid?
I have a large open source project and noticed the number of LLM generate PR is making it unmanageable. Every two weeks, I go in, kill all of them and when someone complains or asks why, I realize it was a real person and then I merge it.
is anyone else seeing this / fixed this problem ?
Yes, I "fixed" it by disabling pull requests on the repository. I'm still happy to pull from other people's branches (and do say so in CONTRIBUTING.md)
> kill all of them and when someone complains or asks why, I realize it was a real person and then I merge it.
I mean an AI skill is perfectly capable of doing this exact same thing.
Related:
Cal.com is going closed source
https://news.ycombinator.com/item?id=47780456
> Security testing has to become an automated, integral part of the CI/CD pipeline. When a developer opens a pull request, an AI agent should immediately attempt to exploit it. When infrastructure changes, an AI should autonomously validate the new attack surface. You do not beat automated attackers by turning off the lights; you beat them by running better automation on the inside.
This feels like the core of the article, but it doesn’t prove the need for open source.
I'll admit that I agree with a lot of the post but that I can't fully wrap myself around the cybersecurity situation today, is it basically:
-if code is open source or closed source, AI bots can still look for exploits
-so we need to use AI to develop a checklist program regardless to check for currently known and unknown exploits given our current state of AI tools
-we have to just keep running AI tools looking for more security issues as AI models become more powerful, which empowers AI bots attacking but also then AI bots to defensively find exploits and mitigate them
-so it's an ongoing effort to work on
I understand the logic of closing the source to prevent AI bot scans of the code but also fundamentally people won't trust your closed source code because it could contain harmful code, thus forcing it to be open source
Edit: Another thing that comes to mind is people are often dunking here on "vibe coding" however can't we just develop "standards / tools" to "harden" vibe coded software and also help guide well for decisions related to architecture of the program, and so on?
> Closing your source code does not stop an AI from probing your API or finding an authorization bypass in your webhooks.
I see this trope a lot in security discussions. “Obscurity isn’t security” or “since you can’t protect against X you may as well do Y”.
This is a harmful trope, which discourages perfectly good protections. Sure, closing source is not a perfect protection, but it is a defense against a large band of attacks.
Think of the entire field of potential vulnerability probes attackers have. Closing the source closes many of them off, likely most of them.
A pen-tester model with implementation will be loads more effective than one with only a black box. And that will give cal.com time to run the pen testing model on the source and address the vulns , hopefully before they are exploited.
I tested this myself, first using black box model attacks, secondly using the source code. The model with the source found and exploited the vulns instantly . The model without failed.
The lesson is: obscurity is not security ALONE, but it is a component of security.
Related ongoing threads:
Cal.com is going closed source - https://news.ycombinator.com/item?id=47780456
Cybersecurity looks like proof of work now - https://news.ycombinator.com/item?id=47769089
First we blamed AI for layoffs, next we are blaming AI for the AI bait and switch.
It's entirely possible this CEO sincerely believes this, but that means you as a potential customer should stay away: now you know that the CEO of this company has no idea how technology works even at an executive level and/or that he doesn't consult his experts before making decisions.
That's literally not it, a CEO can know how technology work and not apply it for its management, many people do things they "dislike" or don't believe in everyday.
Well, that's what I mean, this guy is using this issue as an scapegoat to close source the software and increase revenues as a result.
The pipeline goes like this:
Use open source license to gain traction and credibility > establish a customer base > pull the rug on open source to get everyone who depends on your product but isn't yet paying to pay.
I'm pro FOSS, militantly so. FSF-style.
But... playing devil's advocate, if AI makes it very easy to find exploits without the source code, wouldn't it be doubly effective finding them with the source code as well? And why is the dichotomy posed by this blog post "open source with AI reviews by everyone" vs "closed source but only the bad guys use AI"? What if the scenario was: closed source and the authors/security team use every AI tool at their disposal to find bugs? What do the community's eyeballs add to this equation, assuming (big if) AI review of exploits is such a force multiplier?
Before any knee-jerk reactions: big fan of open source, I'm not arguing this will kill it, I don't have the faintest idea what Cal.com is and I think a world without FOSS would be a tragedy, I run linux and most of my software on my personal PC (other than games) is FOSS.
Pretty overreaching claim about another company's internal decisions and open source in general. There is a lot of incentive to stop open source these days.
One of which I am experiencing right now is somebody just copying my repo, not crediting me, didn't even try to change the README. It's pretty discouraging.
The other is security reasons, the premise that volunteers will report vulnerabilities really matter if you are big enough for small portion of people to dedicate themselves, for the most part people take open source tool use it and then forget about it, they only want stuff fixed.
Lastly, open source development kinda sucks so far. I'v been working on a few different tools and the amount of trolling and just bad faith actors I had to deal with is exhausting. On top of that there is a constant stream of people just demanding stuff to be fixed quickly.
The idea of tying source code to sustenance will soon be history. We will all remember the days when adding some few thousand smart lines of code meant you could gain notoriety and through cheap viral copy expand those traits to wealth and worth. But software has always just been zeros and ones, the value only happens when interpreted.
The future is sharing, you may not believe because your income is tied to being clever. Long term we are all more clever because of the sharing, and your contribution sometimes does not add to your personal success. Asking a company or its individuals to forego their success will not make them add more to our future. But they will add to our future nonetheless, because they all feel like we all do, that adding is what we are all meant to do.
So Cal.com favors security through obscurity.
Open Source was always open to "many eyes" in theory exposing itself to zero-day vulnerabilities. But the "many eyes" go for the good and the bad actors.
As far as I am concerned... Way to go Cal.com, and a good reminder to never use your services.
At the same time, I heavily support open-source and contribute a lot, but I can't necessarily agree that security-through-obfuscation doesn't play a major role in slowing down attacks. Cloudflare have based its whole security being closed-source (for example on its anti-bot mechanism) to be hard to reverse engineer, and they remain leaders as of today with few serious security breaches.
Some things just can't be truly secure as well, ddos protection is mostly a guessing/preventive game, exposing your firewall config/scripts will make you more vulnerable than NOT.
If your codebase isn't exposed, attackers are constrained by the network and other external restrictions which greatly reduce the number of possible trials, even with a swarm of residential proxies, it's not the same at all from inspecting a codebase in depth with thousand of agents and all models.
> In the past, exploiting an application required a highly skilled hacker with years of experience and a significant investment of time to find and exploit vulnerabilities. The reality is that humans don’t have the time, attention, or patience to find everything.
I read this as:
"We figured no one was looking so we just shipped unsafe garbage for years. We never once did an internal audit, never once paid a hacker to try to exploit our product, never thought we'd get caught with our substandard products."
If a guy in his basement with $200 dollars can ruin your company then you were trading on vapor the entire time. I'm sorry you had to find out this way.
I'm hopeful the article is right about its prediction, although I'm under the impression the attacker/defender dynamic is asymmetric and the defender on the loosing end. I hope someone can proof me wrong though...
Making the assumption that the same amount of money needed to attack a critical vulnerability is also required to find and fix it.
Lets say we have a project with 100 modules, and it costs us $100 000 to check these modules for vulnerabilities. What is stopping an attacker from spending the same amount of money to scan, lets say 10 modules but this time with 10x the number of tokens per module than the defender had when hardening the software?
Closing your source doesn't close your attack surface,it just closes the community that would have helped you defend it. Security through obscurity is a kind of tradeoff, not a strategy.. i mean that's what I feel.
It's a good question - is blackbox hacking as effective as whitebox hacking, for AI agents? I've gotta assume someone at Anthropic is putting together an eval as we speak.
I don't really know, but I have a story which might prompt some conversation about it.
At $WORK we had a system which, if you traced its logic, could not possibly experience the bug we were seeing in production. This was a userspace control module for an FPGA driver connected to some machinery you really don't want to fuck around with, and the bug had wasted something like three staff+ engineer-years by the time I got there.
Recognizing that the bug was impossible in the userspace code if the system worked as intended end-to-end, the engineers started diving into verilog and driver code, trying to find the issue. People were suspecting miscompilations and all kinds of fun things.
Eventually, for unrelated reasons, I decided to clean up the userspace code (deleting and refactoring things unlocks additional deletion and refactoring opportunities, and all said and done I deleted 80% of the project so that I had a better foundation for some features I had to add).
For one of those improvements, my observation was just that if I had to write the driver code to support the concurrency we were abusing I'd be swearing up a storm and trying to find any way I could to solve a simpler problem instead.
Long story short, I still don't know what the driver bug was, but the actual authors must've felt the same way, since when I opted for userspace code with simpler concurrency demands the bug disappeared.
Tying it back to AI and hacking, the white box approach here literally didn't work, and the black box approach easily illuminated that something was probably fucky. Given that AI can de-minify and otherwise spot patterns from fairly limited data, I wouldn't be shocked if black-box hacking were (at least sometimes) more token-efficient than white-box.
>simpler concurrency demands
This seems to be extremely common. Been a very long time since I looked at Linux kernel stuff, but there were numerous drivers that disabled hardware acceleration or offloading features simply because they became unreliable if they were given heavy loads or deep queues.
feels like people are arguing the wrong axis tbh
- it’s not open vs closed anymore, it’s more like bug finding going a few devs poking around to basically infinite parallel scanners
- so now you don’t get a couple of thoughtful reports, you get a many edge cases and half-real junk. fixing capacity didn’t change though
- closing the repo doesn’t really save you, it just switches from white-box to black-box… and that’s getting pretty damn good anyway
real problem is: vuln discovery scaled, patching didn’t. now everything is a backlog game
How long before LLM perform perfect disassembly exploitation...
https://x.com/steipete/status/2044423791405924562 very soon it seems...
I’m afraid the balance of open source wasn’t broken today. It happened quite a long time ago. It’s just not something people usually talk about. Companies have been using open source code for years to build paid products without giving anything back. Take PHP as an example: a language widely used across the internet, yet with a very limited budget.
Great PR piece by Strix, but I find mixed messages.
Cal.com folks are getting a red team for free, wouldn't that further convince them their closed source software is strong enough?
Isn't Strix's business companies paying for scans regardless of whether the software scanned is open source or closed?
Strix was so close to being the hero we deserve. I think these blue torches like strix should offer their services for free to open source ships out at sea. There are 3 wins here, GLOBAL GOOD WILL, testimonial and reviews, and market loyalty reward.
Can any of the AI systems read binary yet? Perhaps generate source code from object file? Is so, that would make access to source redundant for that type of analysis.
AI assisted decompiling has been a thing for a while now, from what I know most people are using assisted tooling for it.
With that said it at least seems possible to be able to be able to read binary itself, but most of the magic there is in execution, so you'd have to have an LLM behave kind of like a processor I think.
Yes, the current meta for ctfs, which includes challenges for exploiting binaries, is to just throw an LLM at it.
Is there any recent research on whether open or closed-source projects are more secure? I am genuinely curious if anyone has studied the question.
I mean "yes but" lots from 2015 and before, on a scholarly paper search engine. (I do not have access to most, but there are some public ones)
I mean as a convention when dealing with cryptography, so far the only organization that has succeeded in doing closed-source cryptography securely, has been the USA's "NSA", and mostly their algorithms are public.
I mostly work in the closed source world, however my observation from all the code bases I've seen is "mostly open source are more secure", except when very thorough following of formal security specifications are followed, and then security is as good as the specifications. (YMMV there, of course).
Seems like flimsy reasoning from the Cal.com CEO. How should we think about Strix vs. foundational model releases like Mythos?
feels like the real shift is not open vs closed, but reaction time AI attackers don’t need perfect access anymore, just enough surface and time. So the question becomes: can you detect+respond faster than they can iterate in that sense, open source might even help -more eyes reduces time to fix, not just time to find
We need to be more bullish as a community.
So many people discussing things like UBI, however we selfishly create our own little projects all the time.
We need to center our shared efforts, send open source is a step on that.
Nowadays, every closed source solution should be seen as 'you are the product' type of deal.
every line of code is a liability. open, closed, doesn't matter. companies will have to treating it that way--which means actual engineering--or they will get burnt, and hard.
which is why am I happy to launch the first AI-first meeting scheduling app, yes it costs 5000 USD/month but we have almost completely eliminated any use of code in our application, offering enterprise quality security by having Claude (TM) Opus (TM) 4.8 (TM) directly reply to all HTTPS requests - end to end AI driven agentic software is finally here.
a lot of the vulnerabilities in web-apps are people trying to be too smart for their own good.
use battle-tested frameworks such as Rails, Django then you won't make rookie security mistakes.
Except that Django got so many criticals we can't even list them on a thread here, but yeah, using known and ancient frameworks is generally smart.
It's just an excuse. Classic open source rug pull here.
All content is going to go behind paywalls.
There is zero incentive or reason for content creators to let AI slurp their content for free and distribute it and get all the money from it.
Everything new will be licensed and if AI companies want access to it, they will need to pay for it, just like we will.
Will it help? AI authors will just then buy those subscriptions and in the big picture it won't cost that much.
Of course this neglects why mostly free things that were posted on the internet generally won. Take Microsoft for example. All their money makers are licensed, yet at the same time you can download almost every single one for free and install it.
The people that go behind paywalls don't realize how much they'll have to spend on marketing to catch up to those that are open.
And that's only frames the current state, where models are very expensive to train. Once model training is close to the point where a group of individuals can afford it, it's pretty much game over for our current paradigm. The software police will be running around trying to play whack-a-mole on open weight models with people all over the world.
Why would I create content that I don't get paid for and I don't even get credit for? Everyone who creates free content right now is simply doing the work of AI companies to make them more useful for free.
Search engines will cease to exist, so no one will search your content and then click on your link. AI will simply regurgitate your content and take the money for tokens or subscription and not acknowledge you at all.
>There isn't a rule of economics that says better technology makes more, better jobs for horses. It sounds shockingly dumb to even say that
--Humans need not apply.
It's kind of funny that you think you're going to be making money writing software. If you lock up your software who exactly are you selling it to anyway? It's like you're thinking 25% through the situation then going "I can stay where I am and I don't have to change anything", and then crying later when it doesn't work.
What are you going to do, advertise in BYTE magazine (dead). On Instagram? With a sandwich board on a Seattle street corner? What does the software market even look like in the AI age.
And much like how Google and Amazon eat your lunch now whenever they way, successful AI companies will buy up some software ideas and feed them to their models (which will be stolen later by other models). Anyone that sees your software will mock up a useful clone of it pretty quick the first time they see it. And foreign AI companies will just right out steal it.
You're right you won't create content that you don't get paid for, you just won't be creating anything while competing with the other unemployed masses for strawberry picking jobs.
buying into the marketing I see? Might as well making stealing easier just cause they will probably manage to steal everything in the end is an interesting thought.
Like yes if the AI thing works out the theft will be retroactively okay because the people that stole will now be the ruling class, but assuming AI models will write good code one day and do all work, when that hasn't happened yet is just repeating corporate propaganda while presuming being convinced you are some sort of cynical truth tell that's telling it how it is.
I don't think this will happen. If most content goes behind a paywall, releasing content for free will again become a valuable source of attention. It used to be so before the web got filled with so much free content that it lost any value.
I disagree. AI will slurp their content so quickly that no one will notice.
> Today, Cal.com announced they are transitioning their core codebase away from open source. The reasoning provided by their CEO, Bailey Pumfleet, is that AI has automated vulnerability discovery at scale, making code scanning and exploitation "near zero-cost". In this new world, they argue, "transparency becomes exposure."
Laughable and hilarious. Extremely short sighted. I can show code generated by Claude Opus 4.6 at the highest compute intensity that lacks even basic checks in input validation that was clearly provided in the spec.
There's no point in arguing with crypto and AI bros. They are the same tribe. AI crowd however might learn their lessons sooner because the universe isn't forgiving or flexible.
Note: I use AI code generators all the time but I take them as very very dumb transpilers no matter how expensive their input/output pricing it and I learned that hard way.
PS: Edit to fix typos.
Security by obscurity is flawed.
> The real solution: fight fire with fire
Which works if you assume that AI can find 100% of your bugs.
It can't. So this is a complete waste of your time and will hide actual bugs behind a layer of confidence _and_ obscurity.
You're going to actually have to sit down and figure out how to provide real security in your product while earning profits. This is called "work." I understand Silicon Valley would like to earn money and not work. I am eager for these people to get their comeuppance.
"Open Source Isn't Dead."
Well ...
Open Source as such will never "die", but we only need to look at what happened in, say, the last 5 or 10 years. Private entities with a commercial interest, have been flexing their muscles. Microsoft - also known as Microslop these days - with Github is probably the most famous example still, but you can see other examples. One that annoys me personally is Shopify's recent influence - rubygems.org is basically just shopifygems.org now. See: https://blog.rubygems.org/2026/04/15/rubygems-org-has-a-publ...
"Contributors from both the RubyGems client team and Shopify are already working with us on making native gems a better experience for the Ruby community. "
There is a lot more I could add to this (see my complaint about how rubygems.org hijacks gems past the 100.000 download barrier; this was why I retired from using rubygems.org, and then the year afterwards ruby core purged numerous developers. The handwriting is soooooo clear that shopify flexed their muscles here).
I think we need to make open source development more accessible to everyone, not just corporations throwing their money to gain influence and leverage. I don't have a great idea to make this model work; economic incentives kind of have to be there too, I get that part, and I am not sure which models could work. But right now we really have a big problem. We can also see this with age sniffing (age verification - see the article that pointed at Meta at orchestrating influence and lobbyism) and many more changes. Something has to change. Hopefully some people cleverer than me can come up with models that are actually sustainable, even if it may not necessarily be a "fund an open source developer for a year". There could be a more wide-spread "achieve xyz" or some other lower finance effort - but again, I don't have a good suggestion here. Hopfully something improves here though, because I am getting really tired of private interests constantly sabotaging and ruining the whole ecosystem while claiming they do "improve" an ecosystem. We have the old "War is peace. Freedom is slavery. Ignorance is strength." going again. Opposite day, every day.
There are no answers, only compromises.
Corporations are about money.
Individuals need to eat.
Governments love to concentrate power.
There is another product I use that has a freemium model. They hope to monetize a paid tier for users who use the product a lot.
In order to build trust, they open source their product. I forked it, removed the blocks from the freemium feature in 15 minutes using Claude Code. Never published the code to anyone else, just used it myself
Unfortunately, I think it isn’t going to be tenable for systems to be fully open sourced going forward.
Open source is one of the main reasons I recommended cal.com to everyone — I just did so yesterday again in fact!
I'm disappointed to hear this especially since I don't think the rationale makes sense, from what I understand of the security landscape, and it also makes me a little more skeptical of cal.com in general.
This is just an excuse to close source their project while blaming AI. Spineless bullshit excuse instead of owning your choices.
Shame
I agree, it is shortsighted (next quarter syndrome). First of all the AI does not need source to find vulnerabilities and further it breaks the unwritten contract to exchange source for eyeballs which creates better source. I guess the CEO wants less security and stopped evolution of it's code.
It's like the layoffs. Let's blame this thing we wanted to do for a while on AI.
I can't believe we still have people out there buying this baby-brain idea of "If muh code is open than people will find vulns!!" This has been disproven for 20+ years catch up.
AI generated bullshit PRs are clearly the bigger issue in the OSS space.
> We are an open-source project ourselves, recently crossing 24k stars, and our framework processes over 15 billion LLM tokens daily to find software vulnerabilities.
I'm sorry, but talking Github stars comes off as borderline pathetic.
Open source is dead, AI-pundits are applying the wrong lessons. No one has to accept AI or play the game all these AI companies don’t work if everyone stops publishing. Let the AI generated content industry have the publish space, they're very adamant about taking it over and watering it down with slop.
I wrote some very nice expressive text for our deployment guide. My project manager took the guide and had Gemini break it down into plain boring bullet points. AI and the pundits can gf themselves in their journey to kill human expression.
Here is what I wrote in the guide:
"Post Deploy Responsibility
If you made it this far, say “Wow I really did it and it was so easy!”
Did you say it? Good. Now you are entirely responsible for any issues or bugs that may arise from the newly deployed code. Don’t go anywhere until the deploy has finished (usually takes a few minutes). While an issue or bug may not leave you directly at fault, you are responsible for coordinating any rollbacks or remediations that may be needed until the next deploy."
Here is what the product manager slopped it into:
"- Post deploy responsibility
My paragraph wasn't long, hard to understand, or poorly written. I wouldn't have objected to a rewording or some changes but the project manager chose to just copy paste it into Gemini and copy and paste it back. So my take is that they didn't understand what I wrote. Which is a few sentences long and frankly sad if a paragraph is too intense for you to read. When my project manager did this during the meeting I said, "RIP human expression" and their response was a very hasty "no that's not what's happening". This is what all the pundits want to do to everyone and society. Don't believe them that "it's just a tool", that is just a tactic to get you to rollover so they can shove more AI in your face.there seems to be a certain type of person that thinks if an LLM can generate content faster than means it's better, said belief appears to also be linked with thinking that if text is long it should be broken down so it's easier to understand, it would seem a bit contradictory considering LLM generated code is usually more difficult to correctly review than code a human wrote, but at the core of it might be a poor understanding of systems, those bullet-points are going to get skimmed over or read but not understood, not everything can be easy.
And your paragraph had a much bigger impact on the reader. Your paragraph reads like an experienced senior developer teaching you to not screw things up, while the AI generated bullet points sound like generic ToS that everyone ignores.
Thank you, my intent was to write something memorable but also calm any exterior worries about deploying code and the feeling of “being in trouble because prod is broked”.
[dead]
[dead]
[dead]
[dead]
[dead]
[dead]
Enshittification has come for VC backed open-source. As someone on Twittter said, open source has deemed commercial open source obsolete especially when users can point Calude Code to calcom on GitHub and ask it to make them scheduling features directly into their product. That’s what spooked Cal.
I support and encourage the death of “““open source””” so that we may retvrn to Free Software.
cofounder here
going closed source does not mean we are not fighting fire with fire
we are using a handful of internal AI vulnerability scanners for months now
being open source simply reduces risk by 5x to 10x according to several security researchers we are working with https://cal.com/blog/continuous-ai-pentesting-vulnerability-...
Don’t get me wrong but if virtually all modern software infrastructure lives on top of open source and they’re mostly fine then I’d imagine that you can make a scheduling webapp secure independent to if it’s OSS or not.
It’s OK if there’s another reason for this transition, just be transparent about it and don’t treat your users as children.
They don’t owe you a complete list of reasons why they’re close sourcing their software. They are not a publicly traded company and no one (customers) actually cares if the product is open source or not.
No-one cares if a product is closed-source, a lot of people do care if a product is open source, and really unless a product is truly unique (almost no product is) that is often the main selling point.
Open source software is just a method for corporations to get free evangelization of their product. Companies are now saying “with the assault of AI it is no longer valuable to keep the code open.” They don’t need the handful of Twitter addicted software devs that will evangelize for them anymore because they lose too much by publishing the code. For small companies, the product is most likely the extent of their brand and offerings. Not worth the risk of going out of business, especially when competitors probably aren’t open source.
I've always used and advocated for Cal.com because it's open source. I understand you need to make money and this is no longer the GTM, but don't lie about it.