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

Mark Zuckerberg freezes AI hiring amid bubble fears

telegraph.co.uk

Post reply on HN
Print view View original

711–720 of 936 posts

Re: Mark Zuckerberg freezes AI hiring amid bubble fears

#711

Post by miki123211 »

janalsncm wrote: post #678
h4ck_th3_pl4n3t wrote: post #629

Earlier quoted context omitted.

By definition, transformers can never exceed average. That is the thing, and what companies pushing LLMs don't seem to realize yet.

Can you expand on this? For tasks with verifiable rewards you can improve with rejection sampling and search (i.e. test time compute). For things like creative writing it’s harder.

For creative writing, you can do the same, you just use human verifiers rather than automatic ones.

LLMs have encountered the entire spectrum of qualities in its training data, from extremely poor writing and sloppy code, to absolute masterpieces. Part of what Reinforcement Learning techniques do is reinforcing the "produce things that are like the masterpieces" behavior while suppressing the "produce low-quality slop" one.

Because there are humans in the loop, this is hard to scale. I suspect that the propensity of LLMs for certain kinds of writing (bullet points, bolded text, conclusion) is a direct result of this. If you have to judge 200 LLM outputs per day, you prize different qualities than when you ask for just 3. "Does this look correct at a glance" is then a much more important quality.

Re: Mark Zuckerberg freezes AI hiring amid bubble fears

#712

Post by miki123211 »

me551ah wrote: post #414
miki123211 wrote: post #276

Earlier quoted context omitted.

I have an overwhelming feeling that what we're trying to do here is "Netflix over DialUp." We're clearly seeing what AI will eventually be able to do, just like many VOD, smartphone and grocery delivery companies of the 90s did with the internet. The groundwork has been laid, and it's not too hard to see the shape of things to come. This tech, however, is still far too immature for a lot of use cases. There's enough…

Or maybe not. Scaling AI will require an exponential increase in compute and processing power, and even the current LLM models take up a lot of resources. We are already at the limit of how small we can scale chips and Moore’s law is already dead. So newer chips will not be exponentially better but will be more of incremental improvements, so unless the price of electricity comes down exponentially we might never see…

> We are already at the limit of how small we can scale chips

I strongly suspect this is not true for LLMs. Once progress stabilizes, doing things like embedding the weights of some model directly as part of the chip will suddenly become economical, and that's going to cut costs down dramatically.

Then there's distillation, which basically makes smaller models get better as bigger models get better. You don't necessarily need to run a big model al of the time to reap its benefits.

> so unless the price of electricity comes down exponentially

This is more likely than you think. AI is extremely bandwidth-efficient and not too latency-sensitive (unlike e.g. Netflix et al), so it's pretty trivial to offload AI work to places where electricity is abundant and power generation is lightly regulated.

> Most companies are already running AI models at a loss, scaling the models to be bigger(like GPT 4.5) only makes them more expensive to run.

"We're profitable on inference. If we didn't pay for training, we'd be a very profitable company." Sam Altman, OpenAI CEO[1].

[1] https://www.axios.com/2025/08/15/sam-altman-gpt5-launch-chat...

Re: Mark Zuckerberg freezes AI hiring amid bubble fears

#713

Post by ehnto »

throw310822 wrote: post #682
malfist wrote: post #607

Earlier quoted context omitted.

How many years of not seeing returns this quarter does it take before its all hype?

"technology too expensive to be offered at a profit (yet)" != hype

The MIT report that has everyone talking was about 95% of companies not seeing return on investment in using AI, and that is with the VC subsidised pricing. If it gets more expensive that math only gets worse.

I can't predict the future, but one possibility is that AI will not be a general purpose replacement for human effort like some hope, but rather a more expensive than expected tool for a subset of use cases. I think it will be an enduring technology, but how it actually plays out in the economy is not yet clear.

Re: Mark Zuckerberg freezes AI hiring amid bubble fears

#715

Post by hyperbovine »

thfuran wrote: post #472
me551ah wrote: post #414

Earlier quoted context omitted.

Or maybe not. Scaling AI will require an exponential increase in compute and processing power, and even the current LLM models take up a lot of resources. We are already at the limit of how small we can scale chips and Moore’s law is already dead. So newer chips will not be exponentially better but will be more of incremental improvements, so unless the price of electricity comes down exponentially we might never see…

We know for a fact that human level general intelligence can be achieved on a relatively modest power budget. A human brain runs on somewhere from about 20-100W, depending on how much of the rest of the body's metabolism you attribute to supporting it.

The fact that the human brain, heck all brains, are so much more efficient than “state of the art” nnets, in terms of architecture, power consumption, training cost, what have you … while also being way more versatile and robust … is what convinces me that this is not the path that leads to AGI.

Re: Mark Zuckerberg freezes AI hiring amid bubble fears

#716

Post by jmyeet »

danpalmer wrote: post #706

Zuckerberg either doesn't have the resolve for changing the business, or just keeps picking the wrong directions (depending on your biases). First Facebook tried to pivot into mobile, pushed really hard for a short time and then flopped. Then Facebook tried really hard to make the Metaverse a thing, and for a while, but eventually Meta stopped finding it interesting and significantly reduced investment. Then AI was t…

It's important to analyze decisions within the context at the time, not the modern context.

When Facebook went into gaming, it was about the time they went public and they were in search of revenue. At the time, FB games were huge. It was the era of Farmville. Some thought that FB and Zynga would be the new Intel and MIcrosoft. This was also long before mobile gaming was really big so gaming wasn't an unreasonable bet.

Waht really killed FB Gaming was not having a mobile platform. They tried. But they failed. We could live in a very different world if FB partnered with Google (who had Android) but both saw each other as an existential threat.

After this, Zuckerberg paid $1 billion for Instagram. This was a 100x decision, much like Google buying Youtube.

But in the last 5-10 years the company has seemed directionless. FB itself has fallen out of favor. Tiktok came out of nowhere and has really eaten FB's lunch.

The Metaverse was the biggest L. Tends of billions of dollars got thrown at this before any product market fit was found. VR has always been a solution looking for a problem. Companies have focused on how it can benefit them but consumers just don't want headsets strapped to their heads. It's never grown beyond a niche and never shown signs that it would.

This was so disastrous that the company lost like 60%+ of its value and seemingly it's been abandoned now.

Meta also dabbled with cryptocurrencies and NFTs. Also abandoned.

Social media really seems to have settled into a means of following public figures. Individuals generally seem to interact with each other via group texts.

Meta has a massive corpus of posts, comments, interactions, etc to train AI. But what does Meta do with AI? Can they build a moat? It's never been clear to me what the end goal is.

Re: Mark Zuckerberg freezes AI hiring amid bubble fears

#717

Post by foxylad »

danpalmer wrote: post #706

Zuckerberg either doesn't have the resolve for changing the business, or just keeps picking the wrong directions (depending on your biases). First Facebook tried to pivot into mobile, pushed really hard for a short time and then flopped. Then Facebook tried really hard to make the Metaverse a thing, and for a while, but eventually Meta stopped finding it interesting and significantly reduced investment. Then AI was t…

Cory Doctorow has a compelling theory that the megatech companies have to appear to be startups, or else their share price reverts to normal multiples. Hence the continuous string of increasingly over-hyped "game-changing technologies" they all (not just Meta) keep rolling out.

VR, blockchain and LLMs have their value, but it's a tiny fraction of the insane amounts of money being pumped into these bubbles. There will be tears before bedtime.

Re: Mark Zuckerberg freezes AI hiring amid bubble fears

#718

Post by xuki »

danpalmer wrote: post #706

Zuckerberg either doesn't have the resolve for changing the business, or just keeps picking the wrong directions (depending on your biases). First Facebook tried to pivot into mobile, pushed really hard for a short time and then flopped. Then Facebook tried really hard to make the Metaverse a thing, and for a while, but eventually Meta stopped finding it interesting and significantly reduced investment. Then AI was t…

For the past 15 years, mobile has been the main revenue source for Facebook. As big as Facebook is, they're at the mercy of the 2 competitors: Apple and Google. Apple has been very hostile to Facebook, because Facebook make a shitload of money off Apple's platform and they refused to pay a certain percentage to Apple - unlike Google who is paying 20B a year to access iOS users. Apple tried to cut Facebook off with ATT on iOS 14, but it didn't work.

Because of this, Zuckerberg has to be incredibly paranoid about controlling his company destiny, to stop relying on others' platforms to deliver ads. It would be catastrophic for Facebook to not be a main player for the next computing platform, and they're currently making a lot of money from their other businesses. Zuckerberg is ruthless and he is paranoid, he has total control of Facebook and he will use all the resources to control the next big thing. I think it comes down to this: Zuckerberg believes it's cheaper to be wrong than to miss out on the next platform, and Facebook can afford to be wrong (to a certain extend).

Re: Mark Zuckerberg freezes AI hiring amid bubble fears

#719

Post by mlsu »

seu wrote: post #422

These changes in direction (spending billions, freezing hiring) over just a few months show that these people are as clueless as to what's going to happen with AI, as everyone else. They just have the billions and therefore dictate where money goes, but that's it.

This is a structural problem with our economy, much larger than just Facebook. Due its large scale concentration, the allocation of capital in the economy as a whole has become far less efficient over the last 20 years.

Re: Mark Zuckerberg freezes AI hiring amid bubble fears

#720

Post by Rebelgecko »

rs186 wrote: post #387
cm2012 wrote: post #381

Earlier quoted context omitted.

I know a company that replaced their sales call center with an AI calling bot instead. The bot got better sales and higher feedback scores from customers.

And I happen to know a different company that regrets their decision to do something similar: https://tech.co/news/klarna-reverses-ai-overhaul Is my anecdotal evidence any better than yours?

I would argue yes, because you provided a source and a verifiable company name
Post reply on HN