Again, you need to question the premise. Perhaps all the sales and hype you heard simply wasn't true?
In reality, many organizations have already implemented the AI-based improvements to their systems that they need. That work is done, people are enjoying it. The AI vendors want to take it farther. Some coders want to take it farther. Some leaders are pushing it due to FOMO. But "the masses" do not want more. Step outside of the tech silos, and you'll find that most people do not want more AI than we already have.
The signals are there but the usefulness/blast radius is being limited to "I'm not a software developer but I have this specific issue I need to write software to solve. I've done that and here is a Linkedin post explaining what it is and how I did it."
I think we're looking at the wrong demographic/professional sector and throwing up our hands. You have to look at people who don't have as much professional experience with it because everyone you didn't write software in the 2010's is writing it now.
If before LLMs it took 20 years to build something, with LLMs it might still take 2-5 years, and they've been around of only 5 years. So, you're asking this too early.
No, you're not! It's only a matter of attention. Today all eyes are on the LLM itself, but the big change is happening underground. I'm not a developer — I'm just a curious arborist who owns a company that prunes trees in Italy. In 4 months we have totally changed our company with the help of AI. AI, LLMs are just tools. It's the use that changes the output!
It's undeniable that it has improved the productivity in _some_ areas of development. But my point stand nonetheless, if development is improved, it seems to be difficult to surface that to end user.
The premise, originally, was that AI would empower workers to do more with less. Granted this is anecdotal, but most of the stuff I use today is much the same as it was 5 years ago. It seems the world is improving at the same rate as it did before, generally speaking.
Code monkeys who spent 90%+ of their day typing code are probably 10x faster today because of AI.
I'm not convinced most software developers ever spent the majority of their time coding though... A huge part of the job is taking messy user requirements and converting them into technical requirements, and perhaps you could let AI do that for a while but after some time it will blow up in your face if you don't have any opinion on the technical requirements. Software developers also spend a lot of their time thinking about how to architect solutions, considering different technologies and libraries to use, thinking about how to model data, etc...
If I were to guess before AI I suspect the average software developer spent 50% of their day typing code into an editor, so even if AI made this 10-20x faster, that still wouldn't be a 2x in output unless they're also faster at the other parts of the job too... And maybe they are a bit... So maybe the average developer is 2x, or even 3x more productive today with AI. But 10x is so absurd that unless you were a junior developer building Wordpress themes or something I have no idea how you could be working at anywhere remotely close to 10x velocity you were previously.
I mean spent 4 or my 8 work hours in meetings on a single day last week, that time certainly didn't go 10x faster because of AI... Is my career as a SWE some extreme outlier, or do other people also have meetings and do other things in their day that doesn't involving producing code?
> if AI is such a game changing platform
Again, you need to question the premise. Perhaps all the sales and hype you heard simply wasn't true?
In reality, many organizations have already implemented the AI-based improvements to their systems that they need. That work is done, people are enjoying it. The AI vendors want to take it farther. Some coders want to take it farther. Some leaders are pushing it due to FOMO. But "the masses" do not want more. Step outside of the tech silos, and you'll find that most people do not want more AI than we already have.
The signals are there but the usefulness/blast radius is being limited to "I'm not a software developer but I have this specific issue I need to write software to solve. I've done that and here is a Linkedin post explaining what it is and how I did it."
I think we're looking at the wrong demographic/professional sector and throwing up our hands. You have to look at people who don't have as much professional experience with it because everyone you didn't write software in the 2010's is writing it now.
If before LLMs it took 20 years to build something, with LLMs it might still take 2-5 years, and they've been around of only 5 years. So, you're asking this too early.
No, you're not! It's only a matter of attention. Today all eyes are on the LLM itself, but the big change is happening underground. I'm not a developer — I'm just a curious arborist who owns a company that prunes trees in Italy. In 4 months we have totally changed our company with the help of AI. AI, LLMs are just tools. It's the use that changes the output!
That sounds really interesting. Could you elaborate a bit more on how AI has helped you to fundamentally change your company?
Well, in software development is does increase the velocity 10x.
It's undeniable that it has improved the productivity in _some_ areas of development. But my point stand nonetheless, if development is improved, it seems to be difficult to surface that to end user.
The premise, originally, was that AI would empower workers to do more with less. Granted this is anecdotal, but most of the stuff I use today is much the same as it was 5 years ago. It seems the world is improving at the same rate as it did before, generally speaking.
Code monkeys who spent 90%+ of their day typing code are probably 10x faster today because of AI.
I'm not convinced most software developers ever spent the majority of their time coding though... A huge part of the job is taking messy user requirements and converting them into technical requirements, and perhaps you could let AI do that for a while but after some time it will blow up in your face if you don't have any opinion on the technical requirements. Software developers also spend a lot of their time thinking about how to architect solutions, considering different technologies and libraries to use, thinking about how to model data, etc...
If I were to guess before AI I suspect the average software developer spent 50% of their day typing code into an editor, so even if AI made this 10-20x faster, that still wouldn't be a 2x in output unless they're also faster at the other parts of the job too... And maybe they are a bit... So maybe the average developer is 2x, or even 3x more productive today with AI. But 10x is so absurd that unless you were a junior developer building Wordpress themes or something I have no idea how you could be working at anywhere remotely close to 10x velocity you were previously.
I mean spent 4 or my 8 work hours in meetings on a single day last week, that time certainly didn't go 10x faster because of AI... Is my career as a SWE some extreme outlier, or do other people also have meetings and do other things in their day that doesn't involving producing code?
- I'm seeing lots of internal apps to help our customer success teams.
- I'm seeing prototypes escape Figma and live as code for a faster/closer demo experience for product managers.
AI-native firms will be a game changer I think, the Black Swan event is approaching.