I keep seeing claims that AI has 10Xed the productivity of software engineers.
-
@maxleibman We should also see 10x better software or 10x more features.
️We *can* have 10x lines of code!
-
@maxleibman I've never known engineering output to scale linearly with profit? Outside tiny startups.
At the small side, I expect we'll see a rise in very niche companies that previously may not have been feasible. Those with larger markets will ship faster and probably grow faster than in the past.
The big companies, are already big. eg I can't see AWS S3 suddenly going 10x, they already have the features and now it's mostly adoption and efficiency. Which takes the form of MS/Goog layoffs.
Funny enough, in the middle-tier, I've heard some are actually experiencing some margin erosion because features are suddenly a weak moat...
@kento Maybe not purely linearly, but they have to move together or it's not happening. You can't deliver 10X the value on the same salary and not affect the profitability of the company.
Unless, of course, either developer productivity is an extremely small contributor to profitability—and it's smaller than most people would assume, especially as you say at the big firms, but it would have to be implausibly small in this case—or AI is in aggregate creating negative impacts everywhere else in the org. Or some combination of both.
(Or, of course, the third possibility: "10X more productive" is true, but "productive" has been ill-defined. Somebody is producing 10X the lines of code but but only about the same economic value they were producing before. But that means 10X is a lie.)
And as you say, this is probably more possible in a small shops, but I'm not hearing about any of them taking off like a rocket, either.
-
@kento Maybe not purely linearly, but they have to move together or it's not happening. You can't deliver 10X the value on the same salary and not affect the profitability of the company.
Unless, of course, either developer productivity is an extremely small contributor to profitability—and it's smaller than most people would assume, especially as you say at the big firms, but it would have to be implausibly small in this case—or AI is in aggregate creating negative impacts everywhere else in the org. Or some combination of both.
(Or, of course, the third possibility: "10X more productive" is true, but "productive" has been ill-defined. Somebody is producing 10X the lines of code but but only about the same economic value they were producing before. But that means 10X is a lie.)
And as you say, this is probably more possible in a small shops, but I'm not hearing about any of them taking off like a rocket, either.
@kento I will also confess I may be calling this out too early. The AI coding mania really took off less than a year ago. But the stories need to start showing up at some point, with real numbers from real players.
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman They have. You can achieve 10x productivity with AI, but at about half the quality and 2x the cost (tokens).
-
@kento Maybe not purely linearly, but they have to move together or it's not happening. You can't deliver 10X the value on the same salary and not affect the profitability of the company.
Unless, of course, either developer productivity is an extremely small contributor to profitability—and it's smaller than most people would assume, especially as you say at the big firms, but it would have to be implausibly small in this case—or AI is in aggregate creating negative impacts everywhere else in the org. Or some combination of both.
(Or, of course, the third possibility: "10X more productive" is true, but "productive" has been ill-defined. Somebody is producing 10X the lines of code but but only about the same economic value they were producing before. But that means 10X is a lie.)
And as you say, this is probably more possible in a small shops, but I'm not hearing about any of them taking off like a rocket, either.
@maxleibman all this plays out in a competitive ecosystem.
From my perspective, GenAI is just a commodity productivity technology. Whether it's 10x or 0.5%... doesn't really matter.
If two companies are competing and one adopts the productivity technology... their competitors will follow along, getting the same advantage, returning to competing on price. It's table stakes that get passed on to customers as price/feature benefits, not an advantage, nor a profit moat.
(I don't know what the productivity factor is, or bringing an opinion on whether any of this is "good"... it's just an interesting system dynamic to think about)
-
I eagerly await someone arguing that software developer productivity isn't the only factor impacting the profits of a software company, because to do so will be to argue either or both that development is a very minimal part of the value of a software company, or that AI is in aggregate negative for all other productive activity in such companies.
@maxleibman I would actually argue that most of the positive impact of AI in my company has been outside of software development.
It's mostly small things like looking for a lost message, trying to figure out what a decision has been, summarising a thousand messages in Slack, and things like that. Sure, I still need to go check it, but at least I know where to look.
Definitely not worth the lunacy surrounding AI at the moment, but my life would be a lot harder if not for it, that's for sure.
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman Congrats, your cough is now ten times more productive!
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman Maybe not 10x, especially not for everything, but I would say I save at least 20% of my time. Sometimes it means that something gets done which otherwise would not have though. Sometimes what I am doing is turning an UB error path into a recoverable error. Ticks all the priority boxes, but mostly irrelevant for the bottom line, at best it will prevent a security incident in the future. And with AI we now find a lot more of those bugs with security implications that no user reported.
-
@kento I will also confess I may be calling this out too early. The AI coding mania really took off less than a year ago. But the stories need to start showing up at some point, with real numbers from real players.
@maxleibman @kento no think you're right. I think it's about 9-10 months since "omg coding agents are actually good now" so it's fair to start asking where all the success stories are. Especially since AI is supposed to make coding so much faster... If you can ship a hand-written app in a year then, given the claims being made, I'd expect a huge influx of successful new apps and features would already exist.
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman Two days ago, I posted that I can now achieve this 10x speed. Might contain sarcasm.
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
I'm an ex-Alpha geek computer pioneer with my brain turning to mush.
I have a sign by my egress door "Put on pants"The gif below is some screens stitched together from my Ai harness engine. It took me 2 years to put together a PHP content manager (like WP, but way shittier)...
...this harness would take me something like 4 years to put together. I made it in 2 months (of interactive, non contiguous progress) this is V4.
Oh its also in Python, a language I have 0% familiarity with.You DEFINITELY get the x10
You ask, where is the proof?
The proof is in the recruitment ads, all the geeks are bitching how few jobs there are. -
@maxleibman I've never known engineering output to scale linearly with profit? Outside tiny startups.
At the small side, I expect we'll see a rise in very niche companies that previously may not have been feasible. Those with larger markets will ship faster and probably grow faster than in the past.
The big companies, are already big. eg I can't see AWS S3 suddenly going 10x, they already have the features and now it's mostly adoption and efficiency. Which takes the form of MS/Goog layoffs.
Funny enough, in the middle-tier, I've heard some are actually experiencing some margin erosion because features are suddenly a weak moat...
GPT2.0 was "too dangerous to release" over 7 years ago. How many more years before anyone can actually give more than 1.5 questionable examples of those niche small startups that would otherwise not exist?
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman My (limited but trustworthy AFAIC) sources say it does increase but nowhere near that. In fact there seems to be a divide precisely between a single programmer's own perception (say "I'm doubling") and how that translates to org results e. g. what the financial dept accounts for ("It's a +20%, tops")
As a new, ongoing process, it may take time for the dust to settle and let us have a clear view of the whole thing.
What they agree on is that there's no going back on the field -
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman FWICS, it's mostly replaced "dig around on Google and StackExchange then cut-and-paste" as a primary source, which I suppose could be faster, but not really 10x.
I'd also expect it to acquire more hidden tech debt through lack of understanding (possibly quite a bit more than cut/paste code), so that 10x would reduce in time.
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman
Hey, 10x more LOC and the Excel folks are happy.A boss in a company I used to work at said: "Programmers are, in fact, just PowerPoint-to-code converters". This was some 15 years ago.
Managers. They never change.
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman As a software developer who uses it, it doesn't make you 10x more productive, it just trivializes the code writing part.
However, with few exceptions, writing the code wasn't the hard part, it was building something clean, secure, accessible and user friendly that was the hard part and still is.
It turns out interfacing with the various other departments in a company still takes the same amount of time as it always did.
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman I had to use 4GL tools in the late '80s and I was developing software 10 times faster. In the late noughties I worked with PC Soft's crappy, sexist, French Win- and WebDev and I was developing 10 times faster. That means that now, with AI, I can develop at least 1000x faster than I did at the start of my career, right? RighT?? *sob* RIGHT!!?!?!?!??!
-
I keep seeing claims that AI has 10Xed the productivity of software engineers. I’ve seen them here, this week, in fact.
Here’s my question: where’s the software company that has 10Xed their profits using AI?
If these kinds of gains were in reach, why hasn’t Microsoft or Google done it?
@maxleibman if it's like anything I've seen, companies choose a POC project that's well suited to an AI test - maybe looking at a non-critical legacy codebase. Set the deliverables in a well structured way, assign senior engineers and put a bunch of specialist support around the project (which doesn't get counted in the productivity equation).
Then voilà - max gains! Not repeatable of course but execs get an AI announcement and an AI company gets to claim 10x productivity -
J jwcph@helvede.net shared this topic
