Live data from Hacker News

The turbulent AI era is here

gatesnotes.com

331–340 of 343 posts

Re: The turbulent AI era is here

#331

Earlier quoted context omitted.

not an unfair comparison but i think it's the wrong model. you were always able to buy time by hiring people to do things for you. right now it's agents. i think it's an equalizer in that, $200 agent is more accessible than a $10000 human

> you were always able to buy time by hiring people to do things for you. right now it's agents. i think it's an equalizer in that, $200 agent is more accessible than a $10000 human The other half of this is that, in the past, skill meant that a year of weekends could be turned into a viable business that had skill as a moat. Now, skill means almost nothing: the fleeting product of a drooling idiot is almost indistin…

I don't think that's fair. You can 3D print doodads and spin up an AI business to sell them, but it's not the same as selling a quality tool that's been forged in the failures of previous versions. Skill is still a moat.

Re: The turbulent AI era is here

#332
post #301

Earlier quoted context omitted.

yeah, fair on that point, that makes doing Sloppy SAAS really more accessible and that's what we're seeing On the other hand, what I see happening is: 0. corporation use AI plan ultra-X-high-pro-premium-enterprise 1. corporations firing medium-high paid developers with solid experience and experstise because low-medium paid developers are performing well with ultra-X-high-pro-premium-enterprise plan 2. AI models evol…

I’m a CEO not a dev and this is just patently false. I would never trust someone who lacks the fundamentals to create value even with the most advanced tools. I’d prefer to pay for the senior and fire the less senior - the implicit premium is to not have to worry about the lack of fundamentals for when problems need to be solved that are nuanced.

There are no nuanced problems

Re: The turbulent AI era is here

#333
post #329

Earlier quoted context omitted.

Excel sped up reviews >100x without Bill Gates preaching daily that devs were cooked.

Excel has limited uses. AI threatens replacement of all cognitive work. Don’t you see the difference?

> AI threatens replacement of all cognitive work

AI does nothing. CEOs want to use it to replace all cognitive work, it doesn't mean it is inevitable. That is why smart people are fighting against the narrative push.

Re: The turbulent AI era is here

#334
post #270

The notion of AI having an equalizing effect seems absurd to me. Buying tokens is about as pay-to-win as it gets. More so than anything we’ve seen in a long time. Before everyone had the same 24hr in a day which despite inequalities grounded things a bit. Now some people can run 100+ near human ability agents while others are still wondering how to get enough calories for the day. Someone in Sudan can’t afford a frea…

not an unfair comparison but i think it's the wrong model. you were always able to buy time by hiring people to do things for you. right now it's agents. i think it's an equalizer in that, $200 agent is more accessible than a $10000 human

Another idiot who doesn’t understand capitalism

Nobody pays for time. They are paying for a human form of capital.

Re: The turbulent AI era is here

#335
post #301

Earlier quoted context omitted.

I’m a CEO not a dev and this is just patently false. I would never trust someone who lacks the fundamentals to create value even with the most advanced tools. I’d prefer to pay for the senior and fire the less senior - the implicit premium is to not have to worry about the lack of fundamentals for when problems need to be solved that are nuanced.

There are no nuanced problems

You must be one of the LLM addicts.

Re: The turbulent AI era is here

#336
post #310

Earlier quoted context omitted.

> Before everyone had the same 24hr in a day which despite inequalities grounded things a bit. Now some people can run 100+ near human ability agents while others are still wondering how to get enough calories for the day. Someone in Sudan can’t afford a freakin $200 plan I don't think this is a useful model. People struggling to survive have other priorities. A better model for thinking about this is how cloud opene…

I can ship a new feature almost every day on the $20 Codex plan. The limiting factor has always been the drive to build something. General Artificial Intelligence is here and most will continue going about their normal routine

Yep exactly.

For people like scam Altman it’s a difficult paradox: we have a ‘intelligent’ technology box and the world largely… looks the same.

He blames it on inertia lmao - as if firms are not desperately trying to increase profits and getting rid of people without having to reinvest.

Re: The turbulent AI era is here

#337
post #251

I am not optimistic. What we learned from history is that despite all warnings and good intentions, governments, corps and citizens will make every possible mistake before we finally arrive at another precariously stable situation. Experts are famously bad at predicting the outcomes of new technologies/policies even in times of relatively low upheaval. And plans have a way of producing many significant unintended con…

WFH was a great example of this. Everyone seemed to be on board, this was the way of the future, etc. Yet it seems nobody in charge thought about downtown economies or commercial real estate, and it turns out having a bunch of empty office buildings is really bad for banks and pension plans that invest in them. Somehow the experts or economists didn't see this coming, and since the system in question is too big to fa…

Or maybe.. the people lies in you expecting too much of such people like economists? Who have a hammered in fixed model of the world.

Re: The turbulent AI era is here

#338
post #268

Earlier quoted context omitted.

> But it's impossible to measure how does a company or a person benefit from AI. I think he was proposing taxing the companies offering AI. > Re: AI killing jobs where you sit behind a desk The robots aren't far behind.

Companies offering AI will capture just a tiny slice of the value created by AI. So the tax would generate negligible revenue. I think humanoids capable of replacing a nurse for example are 20+ years away. But I could be wrong and hope I am.

Humanoids replacing a nurse Hahahha.

Wow you are in deep. I feel for people like you - you’ve lost all sight of reality. It’s sad.

Re: The turbulent AI era is here

#339
post #299

Earlier quoted context omitted.

That paper has had a pretty turbulent reception and looks pretty conclusively wrong at this point. It used an incorrect theoretical framing that assumed that data was being replaced rather than accumulated as a result of more training (see https://arxiv.org/abs/2404.01413 which explores this). This is incorrect because this simply isn't how real-world datasets are created via synthetic data generation (which generall…

We've already seen it start and they had to actively work against it: https://openai.com/index/where-the-goblins-came-from/

That article explains that the quirk was introduced during supervised fine-tuning.

> We unknowingly gave particularly high rewards for metaphors with creatures

It was the human feedback that caused the bias, not a change in training data.

Re: The turbulent AI era is here

#340
post #302

Earlier quoted context omitted.

This is still a nonsense argument. Do more of.. what? Much of what happens in software is not akin to a factory producing goods. You need projects - projects don’t fall out the sky and value creating ones are hard to come by!

The funny thing about that is, I would argue, the opposite. Much of what happens in software is like a factory producing goods. I've worked on many projects at many companies; 80% of the code was written some variation of what I'd written elsewhere in different languages. We had to slog through the boring, "Been there, done that" code to get to the 20% that was interesting. Also consider that there's a lot going on i…

> 80% of the code was written some variation of what I'd written elsewhere in different languages. We had to slog through the boring, "Been there, done that" code to get to the 20% that was interesting.

When you ignore the 80% mechanical work the last 20% is 100% of a project.

Post reply on HN