Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI adoption and Solow's productivity paradox

fortune.com

Post reply on HN
Print view View original

731–740 of 783 posts

Re: AI adoption and Solow's productivity paradox

#731

Post by simonw »

overgard wrote: post #712
simonw wrote: post #223

Earlier quoted context omitted.

It matters because as long as they are selling inference for less than it costs to serve they have a potential path to profitability. Training costs are fixed at whatever billions of dollars per year. If inference is profitable they might conceivably make a profit if they can build a model that's good enough to sign up vast numbers of paying customers. If they lose even more money on each new customer they don't have…

I'm curious just because you're well known in this space -- have you read Ed Zitron's work on the bubble, and if so what did you think of it? I'm somewhat in agreement with him that the financials of this just can't be reconciled, at least for OpenAI and Anthropic. But I also know that's not my field. I find his arguments a lot more convincing than the people just saying "ahh it'll work itself out" though.

My problem with Ed is that he's established a very firm position that LLMs are mostly useless and the business is a big scam, which makes it difficult to evaluate his reporting.

He often gathers good information but his analysis of that information appears to be heavily influenced by the conclusions he's already trying to reach.

I do pay attention to him but I'd like to see similar conclusions from other analysts against the same data before I treat them as robust.

I don't personally have the knowledge or experience of company finance to be able to confidently evaluate his findings myself!

Re: AI adoption and Solow's productivity paradox

#732

Post by bunderbunder »

Wojtkie wrote: post #724
SoftTalker wrote: post #121

Earlier quoted context omitted.

> Do they also make you write your own performance review and set your own objectives? Not to get off on a tangent but this has got to be a "tell" for how much a company is managed by formula and how much it's actually got thinking people running things. Every time I've had to write my own review I fill out the form with some corporatese bullshit, my supervisor approves it and adds some more bullshit, it disappears i…

This is pretty much how every performance review I've had in the past 5 years has gone. Even up to my VP it's considered pretty useless. I'm not sure where it actually gets used by HR, but I'm sure it's selectively applied.

The process’s entire purpose is to exist and be followed, so that when they need to they can point to it and say, “We followed the process.”

Re: AI adoption and Solow's productivity paradox

#733

Post by simonw »

amrocha wrote: post #730
simonw wrote: post #235

Earlier quoted context omitted.

Here are a few of mine from the past month - for all of them 90%+ of the code written by Claude Code: - https://github.com/simonw/sqlite-history-json - https://github.com/simonw/sqlite-ast - https://github.com/simonw/showboat - 292 stars - https://github.com/simonw/datasette-showboat - https://github.com/simonw/rodney - 290 stars and 4 contributors who aren't me or Claude - https://github.com/simonw/chartroom Noting…

Props for your work on these but they’re toys mate. These are things you built for yourself that other people happened to find useful. That’s great! I’m not shitting on that, but it doesn’t really convince me that AI coding really is this amazing productivity booster in all cases. It’s good for small greenfield projects, I’ll admit that.

Six useful small greenfield projects in two weeks is pretty good, especially when they weren't my primary focus for those two weeks.

I wouldn't call these toys either. If you want toys take a look at most of https://tools.simonwillison.net/ - these six are all real projects on GitHub with tests and documentation and release notes.

Re: AI adoption and Solow's productivity paradox

#734

Post by gzread »

aurareturn wrote: post #693
toldnotmywrath wrote: post #667

Earlier quoted context omitted.

Problem with much of this talk is receipts are always nowhere to be found.

But I don't see any receipts from the opposite side either.

You don't see any good software made by people who know how computers work?

Re: AI adoption and Solow's productivity paradox

#735

Post by VerifiedReports »

VerifiedReports wrote: post #263

"What had promised to be a boom to workplace productivity.." No. BOON. A BOON to workplace productivity. And then the writer doubles down on the error by proving it was not a typo, ending the sentence with "...was for several years a bust."

Yes, downvote facts.

Re: AI adoption and Solow's productivity paradox

#736

Post by reg_dunlop »

PowerElectronix wrote: post #461
gsf_emergency_6 wrote: post #142

Earlier quoted context omitted.

Stanford prof rebutts David's idea[0] that it's difficult to extract productivity from the data https://www.nber.org/system/files/working_papers/w25148/w251... I don't agree that real GDP measures what he thinks it measures, but he opines > Data released this week offers a striking corrective to the narrative that AI has yet to have an impact on the US economy as a whole. While initial reports suggested a year of ste…

GDP growth measurements have a big bias due to tariffs on, tariffs off, tariffs on again policies wrecking imports and exports numbers. Consumer spending is up, too, so I too fail to see that gdp growth while jobs are not as up as expected is due to AI making us more productive and not just people spending more after months of increased savings due to tariffs.

I'm curious how "consumer spending" is quantified here.

Is consumer spending measured in total dollars spent? If so, isn't that curious wrinkle in an economy of rising prices, and decreasing purchasing power?

If true, I believe less quantity could be purchased at a higher cost per person, making it appear that consumer spending is up.

Presumably these numbers are benchmarked/peg to some sort of constant and/or standardization

Re: AI adoption and Solow's productivity paradox

#737

Post by warkdarrior »

b40d-48b2-979e wrote: post #696
NoLinkToMe wrote: post #528

Earlier quoted context omitted.

Actually job numbers are depressed (hiring recession) and GDP numbers are still way up, both precisely due to the AI investment. More output with fewer people. Wild take to cite a recession when last quarter growth was 4.4%.

"The economy" is not GDP.

It is for the wealthy. And nobody cares what the economy is like for everyone else.

Re: AI adoption and Solow's productivity paradox

#738

Post by bloomca »

matwood wrote: post #291
aurareturn wrote: post #232

Earlier quoted context omitted.

The future of work is fewer human team members and way more AI assistants. I think companies will need fewer engineers but there will be more companies. Now: 100 companies who employ 1,000 engineers each What we are transitioning to: 1000 companies who employ 10 engineers each What will happen in the future: 10,000 companies who employ 1 engineer each Same number of engineers. We are about to enter an era of explosiv…

> smaller companies And large companies. The first half of my career was spent writing internal software for large companies. I believe it's still the case that the majority of software written is for internal software. AI will be a boon for these use cases as it will make it easier for every company big and small to have custom software for its exact use case(s).

I think a lot of companies are going to get burnt on these things. Sure it is easy to one-shot something which looks close, but then you are responsible for releasing/maintaining/improving.

Not to mention that you'd need to integrate it with lots of other vibe-coded products. It can be great for some use cases for sure, though, but identifying them can be tricky, as big orgs are pretty terrible at formulating what they need clearly.

Re: AI adoption and Solow's productivity paradox

#739

Post by chii »

xienze wrote: post #284
aurareturn wrote: post #273

Earlier quoted context omitted.

This implies that writing code by hand will remain the best way to create software. The seniors today who have got to senior status by writing code manually will be different than seniors of tomorrow, who got to senior status using AI tools. Maybe people will become more of generalists rather than specialists.

> The seniors today who have got to senior status by writing code manually will be different than seniors of tomorrow, who got to senior status using AI tools. That’s putting it mildly. I think it’s going to be interesting to see what happens when an entire generation of software developers who’ve only ever known “just ask the LLM to do it” are unleashed on the world. I think these people will have close to no unders…

> I think it’s going to be interesting to see what happens when an entire generation of software developers who’ve only ever known “just ask the LLM to do it” are unleashed on the world.

we only have to look today at how different software quality is compared to the "old days" - when compilers were not as good, and people wrote in assembly by hand.

Old software were fast and optimized. Hand written assembly used minimal resources. Today, people write bloated electron webapps packaged into a bundle.

And yet, look who is surviving in the competitive land of software darwinian natural selection?

Re: AI adoption and Solow's productivity paradox

#740

Post by itake »

akkanzn wrote: post #566
aurareturn wrote: post #249

Earlier quoted context omitted.

In my case, over the last 3 years, every dev who left was not replaced. We are doing more than ever. Our team shrunk by 50% but we are serving 200% more customers. Every time a dev left, we thought we're screwed. We just leveraged AI more and more. We are also serving our customers better too with higher retention rates. When we onboard a customer with custom demands, we used to have meetings about the ROI. Now we ju…

Could you provide some details on your company, code base, etc? These are wild claims and don’t match the reality I’m seeing everywhere else. How big is your team? How many customers? What’s your product? Can we see the code? How do you track defects? Etc. Part of the reason I’m struggling with this is because we’d be seeing OpenAI, Anthropic, etc. plastering these case studies everywhere if they existed. Instead, I’…

Not OP, but I am seeing this in my current company.

Companies are charged per token, which means heavy AI users deliver more and stress budgets. They recently announced significant payroll costs over the past ~3 years.

Those savings I think will partially be reclaimed by AI companies, enabling the high performers more ai model usage.

Post reply on HN