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

771–780 of 783 posts

Re: AI adoption and Solow's productivity paradox

#771

Post by Wilder7977 »

aurareturn wrote: post #761
Wilder7977 wrote: post #748

Earlier quoted context omitted.

YouTube is a platform, it's not a product. And in this case, created a new market. A market in which, by the way, still very few people (relative to those who try) are successful. In fact I wouldn't be surprised if the percentage would be much smaller than 10%. A quick search leads to different answer, but https://alanspicer.com/what-percentage-of-youtubers-make-mon... suggests that 0.25% of all YouTube channels make…

0.25% is 77,500 thousand channels by the way. Now count how many TV stations there were in the 90s and 2000s. This is just Youtube alone. What about streamers on Twitch? Youtube? Tiktokers? IG influencers? Plenty of people are making money creating video content online. The internet, ease and advancements in video editing tools, and cheap portable cameras all came together to allow millions of content creators instea…

77.500 channels which make any money. Now plenty of those make a handful of dollars per month. Also, 77k worldwide.

I am not going to deny that YouTube (and all social media) created new markets. But how is this not an argument that shows that when N people suddenly do some activity, only a tiny minority is successful and gains some market share?

If tomorrow a product that is made by 3 companies will see competitors by 10000 1-man operations, maybe you will have 30 different successful products, or 100. 9900 of those 10000 will still be out of luck. I

YouTube is not an example of a market that being exposed to a flood of players gets shares somewhat equally between those players or that allows a significant number of the to survive with it. Nor is twitch or any of the other platforms.

Re: AI adoption and Solow's productivity paradox

#772

Post by csa »

Wilder7977 wrote: post #749
csa wrote: post #661

Earlier quoted context omitted.

> And those companies will do what? Produce products in uber-saturated markets? > Or magically 9900 more products or markets will be created, all of them successful? Yes. Products will become more tailored/bespoke rather than a lot of the one size fits all approach that is pervasive now.

And if it's so cheap and bespoke, why buying it and not making it in house? What about access to people with know-how of that product? You use a product that only 4 other companies use, you can be sure you won't find any new hire that knows how to use it. To me it seems the reality works in the opposite way. Among the many products built, some will be successful and will swallow the whole market, like now with basica…

> And if it's so cheap and bespoke, why buying it and not making it in house?

0. Sure, some products will be made in house. That said, being able to spec a product well is a skill that is not as common as some folks seem to think. It also assumes that an org is large enough to have a good internal dev team, which is both rare and relatively expensive.

1. It sloughs responsibility, which many folks want to do.

2. It allows for creation to be done not by committee and/or with less impact from internal politics.

3. It facilitates JIT product/tool development while minimizing costs.

That’s off the top of my head.

The realities of business often point to internal development not being ideal.

Re: AI adoption and Solow's productivity paradox

#773

Post by Telaneo »

alephnerd wrote: post #192
tstrimple wrote: post #161

Earlier quoted context omitted.

> On hacker news, a very tech literate place I think this is the prior you should investigate. That may be what HN used to be. But it's been a long time since it has been an active reality. You can still see actual expert opinions on HN, but they are the minority more and more.

I think one longtime HN user (Karrot_Kream I think) pinpointed the change in HN discourse to sometime in mid 2022 to early 2023 when the rate of new users spiked to 40k per month and remained at that elevated rate. From personal experience, I've also noticed that some of the most toxic discourse and responses I've received on this platform are overwhelmingly from post-2022 users.

It's still September.

Re: AI adoption and Solow's productivity paradox

#774

Post by ssvt »

Maybe I’m slow (alright, I know I am) but it seems to me that HN has jumped the shark with it’s apparent shift to “all AI, all the time” making me lose interest.

Yes, I won’t let the door hit me in the ass on the way out…

Re: AI adoption and Solow's productivity paradox

#775

Post by docmars »

fastasucan wrote: post #630
docmars wrote: post #587

Earlier quoted context omitted.

True, but you'd be surprised how much you can tighten up a codebase by asking a heftier model to do a security review and suggest fixes.

At what point do people really know if it has been tightened up if they never look at the code?

That's the catch -- a team would need to care enough about quality, or don't at their own peril.

Re: AI adoption and Solow's productivity paradox

#776

Post by jurschreuder »

franktankbank wrote: post #694
jurschreuder wrote: post #124

Workers may see the LLM as a productivity boost because they can basically cheat a their homework. As a CEO I see it as a massive clog up of vast amounts of content that somebody will need to check. A DDoS of any text-based system. The other day I got a document of 155 pages in Whatsapp. Thanx. Same with pull requests. Who will check all this?

Who gave you the 155 page doc? How quickly were they fired?

A customer, he also did some research himself that I should look at using Claude

Re: AI adoption and Solow's productivity paradox

#777

Post by jurschreuder »

thisoneisreal wrote: post #206
jurschreuder wrote: post #124

Workers may see the LLM as a productivity boost because they can basically cheat a their homework. As a CEO I see it as a massive clog up of vast amounts of content that somebody will need to check. A DDoS of any text-based system. The other day I got a document of 155 pages in Whatsapp. Thanx. Same with pull requests. Who will check all this?

Just yesterday one of my junior devs got an 800-line code review from an AI agent. It wasn't all bad, but is this kid literally going to have to read an essay every time he submits code?

Yes it's like "double it and pass it to the next developer".

Even more troublesome is importing libraries. I have no idea which ones are AI generated and they are better and better at hiding their original authors.

Re: AI adoption and Solow's productivity paradox

#778

Post by jorvi »

Gracana wrote: post #606
jorvi wrote: post #580

Earlier quoted context omitted.

Hehe, yeah there's some terms that just are linguistically unintuitive. "Skill floor" is another one. People generally interpret that one as "must be at least this tall to ride", but it actually means "amount of effort that translates to result". Something that has a high skill floor (if you write "high floor of skill" it makes more sense) means that with very little input you can gain a lot of result. Whereas a low…

Are you sure about skill floor? I've only ever heard it used to describe the skill required to get into something, and skill ceiling describes the highest level of mastery. I've never heard your interpretation, and it doesn't make sense to me.

Yes, I am very sure. And it isn't that difficult to understand, it is skill input graphed against effectiveness output. A higher floor just means that with 1 skill, you are guaranteed at least X (say, 20) effectiveness output.

https://imgur.com/tOHltkx

The confusion comes from people using "skill floor" for "learning curve" instead of "effectiveness".

But this is a thing where definitions have shifted over time. Like jealousy. People use "jealousy" when they really mean "envy", but correcting someone on it will usually just get you scorn and ridicule, because like I mentioned, language is fluid.

Re: AI adoption and Solow's productivity paradox

#779

Post by Gracana »

jorvi wrote: post #778
Gracana wrote: post #606

Earlier quoted context omitted.

Are you sure about skill floor? I've only ever heard it used to describe the skill required to get into something, and skill ceiling describes the highest level of mastery. I've never heard your interpretation, and it doesn't make sense to me.

Yes, I am very sure. And it isn't that difficult to understand, it is skill input graphed against effectiveness output. A higher floor just means that with 1 skill, you are guaranteed at least X (say, 20) effectiveness output. https://imgur.com/tOHltkx The confusion comes from people using "skill floor" for "learning curve" instead of "effectiveness". But this is a thing where definitions have shifted over time. Like…

If the skill floor is high and therefore "effectiveness" is the same for a wide range of skill levels, isn't that the same as having a high barrier to entry? It seems that any activity or game where it takes a lot of skill before you can differentiate yourself from other players would be described that way.

Re: AI adoption and Solow's productivity paradox

#780

Post by aakresearch »

riskable wrote: post #579
mcny wrote: post #408

Earlier quoted context omitted.

I would like to see the day when the context size is in gigabytes or tens of billions of tokens, not RAG or whatever, actual context.

Context size helps some things but generally speaking, it just slows everything down. Instead of huge contexts, what we need is actual reasoning . I predict that in the next two to five years we're going to see a breakthrough in AI that doesn't involve LLMs but makes them 10x more effective at reasoning and completely eliminates the hallucination problem. We currently have "high thinking" models that double and tripl…

Upvoted, as it basically 99% matches my own thinking. Very well said. But I, personally, would not predict a breakthrough in this direction in the next 2-5 years, as there is no pathway from current LLM tech to "true reasoning". In my mental model LLM operates in "raster space" with "linguistic tokens" being "rasterization units". For "true reasoning" an AI entity has to operate fluently in "vector space", so to speak. LLM can somewhat simulate "reasoning" to a limited degree, and even that it only does with brute force - massive CPU/GPU/RAM resources, enormous amount of training data and giant working contexts. And still, that "simulation" is incomplete and unverifiable.

I would argue that the research needed to enable such "vector operation" is nowhere near the stage to come to fruition in the next decade. So, my prediction is, maybe, 20-50 years for this to happen, if not more.

Post reply on HN