A.I. researchers are negotiating $250M pay packages

Post by Eisenstein » Sun, Aug 03, 2025, 12:23 PM UTC

ath3nd wrote: post #450

bbor wrote: post #373

Earlier quoted context omitted.

I'm not a conspiracy person, but it's hard not to believe that some cruel god sent us crypto just a few years before we accidentally achieved AGI just to mess with us. So many people are confident that AGI is impossible and LLMs are a passing fad based to a large degree on the idea that SV isn't trustworthy -- I'd probably be there too, if I wasn't in the field myself. It's a hard pattern not to recognize, even if n~…

Dude, LLMs are a very sophisticated autocomplete with a few tricks down their sleeve. From that to AGI or Zuck's "superintelligence" is just light years... LLMs are a dead end for "intelligence". We are talking about the promises of the same crop of people who brought you full Autopilot on Tesla (still waiting since 2019 when it was supposed to happen), or the Boring Tunnel, the Metaverse, and their latest products a…

Post by Mistletoe » Sun, Aug 03, 2025, 12:29 PM UTC

jjmarr wrote: post #229

apparent wrote: post #208

This is the result of the winner-take-all (most) economy. If the very best LLM is 1.5x as good as good as the next-best, then pretty much everyone in the world will want to use the best one. That means billions of dollars of profit hang in the balance, so companies want to make sure they get the very best people (even if they have to pay hundreds of millions to get it). It's the same reason that sports stars, musicia…

The actual OpenRouter data says otherwise.[1] Right now, Google leads with only 28.4% marketshare. Anthropic (24.7%), Deepseek (15.4%), and Qwen (10.8%) are the runners-up. If this were winner-take-all market with low switching costs, we'd be seeing instant majority market domination whenever a new SOTA model comes out every few weeks. But this isn't happening in practice, even though it's much easier to switch model…

Post by hoppp » Sun, Aug 03, 2025, 12:43 PM UTC

tiagod wrote: post #462

ath3nd wrote: post #450

Earlier quoted context omitted.

Dude, LLMs are a very sophisticated autocomplete with a few tricks down their sleeve. From that to AGI or Zuck's "superintelligence" is just light years... LLMs are a dead end for "intelligence". We are talking about the promises of the same crop of people who brought you full Autopilot on Tesla (still waiting since 2019 when it was supposed to happen), or the Boring Tunnel, the Metaverse, and their latest products a…

Please prove that human brains don't work as a very sophisticated autocomplete.

Post by hoppp » Sun, Aug 03, 2025, 12:49 PM UTC

wyre wrote: post #434

HWR_14 wrote: post #339

Earlier quoted context omitted.

There is no guarantee that Meta or OpenAI would capture that increase in productivity as opposed to open models or otherwise driving the profits on the AI itself to zero.

As LLM's get smarter and more capable on smaller and smaller computers Anthropic, Google, Grok, OpenAI, and Meta (is there an acronym, like FAANG, for the giant AI companies? MOGGA? GOMAG?) will have to get creative with profit-drivers, when consumers will have a very capable LLM built-in to their computing device(s), and it can easily be worth the cost to a business to invest in the computing power to provide on-dev…

Post by Voultapher » Sun, Aug 03, 2025, 12:54 PM UTC

tonyhart7 wrote: post #443

Voultapher wrote: post #437

Slightly fixed title: "One A.I. researcher is negotiating a $250M pay package in made up money that Meta can produce by typing zeros into a spreadsheet." Paying with stock is a neat option for companies that are projected to grow - the very reason why all of big tech desperately wants to be perceived as growing - since it doesn't cost them anything, since they can dilute existing stock holdings at will by claiming th…

Yeah but its not made up money tho, meta(fb) must buy out stocks at certain point while its true that many of its would vested for period of time. it still cost them in the future what more embarrassing is that they do this to poach AI talent because they massively behind on AI races, like they literally still to the extend of king of social media (fb,instragram,whatsapp etc) they should do better given how much data…