frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Open in hackernews

A.I. Researchers Are Negotiating $250 Million Pay Packages

https://www.nytimes.com/2025/07/31/technology/ai-researchers-nba-stars.html
54•jrwan•3h ago

Comments

merelysounds•2h ago
https://archive.is/t9HRT
siva7•2h ago
Ok people, is this for real like are these detached IC roles or are these articles talking about executive rolles filled by a.i. researchers?
Avicebron•2h ago
As far as I know it's only one guy that got this offer, https://mattdeitke.com/ (aside from the others who had the mythical 100 million dollar poaching package).
flappyeagle•2h ago
Source: I know people who have both accepted and declined 100M+ packages

They are IC roles for the most part

mr90210•2h ago
Are you aware of the terms of such offers?

I suppose those $100M are spread across years and potentially contingent upon achieving certain milestones.

flappyeagle•39m ago
Even amongst the packages, there is a range. One example package was 100 guaranteed up to 250 based on milestones and incentives over five years
coderatlarge•1h ago
they are executive roles in the sense that you are required to profitably allocate a scarce perishable resource (gpu time) way more expensive than any regular engineer’s time.
flappyeagle•38m ago
Yeah, you could definitely look at it that way. They are IC roles in the sense that their job is to tell computer computers what to do but maybe that’s old-fashioned thinking at this point.
apwell23•1h ago
murati was offered 1B by meta apparently
smokel•2h ago
It's similar to how, near the end of a Monopoly game, a player might indiscriminately hand over a stash of $100 bills to acquire Mediterranean Avenue, even though the property is mortgaged.
mathgeek•2h ago
Which analogy(s) are you going for? The world is about to end so money is essentially worthless? The players with all the money are going to move on to something else soon? The game ceased to be fun for anyone so they all want to find other things to do?

I assume you are going for “there are no more useful resources to acquire so those with all the resources overpay just to feel like they own those last few they don’t yet own”.

qgin•2h ago
I think the idea is the end of the game is nearing (AGI) and specific dollar amounts mean less than the binary outcome of getting there first.
tough•2h ago
if we get AGI and a post-scarcity age what makes these people think they -reaching- AGI will make them kings.

seems like governments will have a thing to say about who's able to run that AGI or not.

GPU's run on datacenters which exist in countries

ElevenLathe•1h ago
Yes but countries are run by governments, which are composed of people, who can be bribed. If you believe that AI will make you the richest person in human history, you presumably can see that the problem of government can be solved with enough money.
mhb•1h ago
Presumably they think that, whatever chance they have of becoming kings if they get there first is more than the chance if someone else does. In we get AGI, we is doing all the work.
smokel•2h ago
I was going for irony, not analogy. Unfortunately, even though some incompetent fools think it is, life is not a game.
mark_l_watson•1h ago
I saw the ‘forgetting about money, moving on to other challenges’ thing happen about 30 years ago. A childhood friend sold his company for about 300 million (a billion in today’s dollar devaluation?). My friend and his wife continued to live in their same house. The only thing he did was to purchase eight houses for extended family members who didn’t own their own homes, he also got his daughter expensive horse back riding lessons and a horse, and he said he and his wife drank more expensive wine. He did continue to play “The Infinite Game” by staying in the tech industry - it seemed like he loved the game, the money was only to help other people in his life.
cornfieldlabs•26m ago
Is that friend Josh Kopelman?
saubeidl•1h ago
Capitalism is about to break. The revolution is coming.
fidotron•2h ago
Good for those involved being offered such packages, but it really does raise the question of what exactly those offering them are so afraid of.

For example, Meta seem to be spending so much so they don't later have to fight a war against an external Facebook-as-chatbot style competitor, but it's hard to see how such a thing could emerge from the current social media landscape.

MichaelZuo•2h ago
They just want the best, and they’re afraid of having second rates, B-players, etc., causing a bozo explosion. That seems like all the motivation that’s needed.

Why why would they need fears about a quasi-facebook chatbot?

lores•1h ago
Just like in football, buying all the best players pretty much guarantees failure as egos and personal styles clash and take precedence over team achievement. The only reasons one would do that are fear, vanity, and stupidity, and those have to be more important than getting value for the extraordinary amounts of money invested.
HarHarVeryFunny•1h ago
Yeah, pretty much agree.

The only case where this may have made sense - but more for an individual rather than a team - is Google's aqui-rehire of Noam Shazeer for $1B. He was the original creator of the transformer architecture, had made a number of architectural improvements while at Character.ai, and thus had a track record of being able to wring performance out of it, which at Google-scale may be worth that kind of money.

HarHarVeryFunny•1h ago
Coming from Meta, I have to wonder if the reason for this isn't more down to Zuck's ego and history. He seems to have somewhat lost interest in FaceBook, and was previously all-in on the Metaverse as the next big thing, which has failed to take off as a concept, and now wants to go all-in on "super-intelligence" (seems to lack ambition - why not "super-duper extra special intelligence"?) with his new vision being smart glasses as the universal AI interface. He can't seem to get past the notion that people want to wear tech on their head and live in augmented reality.

Anyhow, with the Metaverse as a flop, and apparently having self-assessed Meta's current LLM efforts as unsatisfactory, it seems Zuck may want to rescue his reputation by throwing money at it to try to make his next big gamble a winner. It seems a bit irrational given that other companies, and countries, have built SOTA LLMs without needing to throw NBA/NFL/rockstar money around.

InterviewFrog•1h ago
Here is the uncomfortable truth. Only a small group of people are capable of operating at an elite level. The talent pool is extremely small and the companies want the absolute best.

It is the same thing in sports as well. There will only ever be one Michael Jordan one Lionel Messi one Tiger Woods one Magnus Carlsen. And they are paid a lot because they are worth it.

>> Meta seem to be spending so much so they don't later have to fight a war against an external Facebook-as-chatbot style competitor

Meta moved on from facebook a while back.It has been years since I last logged into facebook and hardly anybody I know actually post anything there. Its a relic of the past.

firesteelrain•2h ago
When will the bubble pop?
mr90210•2h ago
I think we’d need a major war or a pandemic of sorts because we have become pretty good at maintaining such bubble inflated.

Whenever and however it comes, it’s going to be a bloodbath because we haven’t had a proper burst since 2008. I don’t count 2020.

nl•2h ago
Eventually people might consider that just maybe... it's not a bubble...
firesteelrain•1h ago
I just don’t think the industry is moatless. Where there is a moat in my opinion is airgap because few are pursuing this and not everyone wants their data in the Cloud.
DonsDiscountGas•1h ago
2000 was a bubble, and yet the internet continued to eat the world after it popped. I expect we'll see something similar
impossiblefork•2m ago
There is definitely a bubble though. Tesla has 28 times larger market cap than other well-run competitors, for example, and there's a bunch of other firms with similarly crazy numbers.

AI is great and it's the future, and a bunch of people will probably eventually turn it into very powerful systems able to solve industrially important maths and software development problems, but that doesn't meant they'll make huge money from that.

mattlondon•1h ago
When more than 1 company has "AGI", or whatever we're calling it, and people realise it is not just a license to print money.

Some people are rightly pointing out that for quite a lot of things right now we probably already have AGI to a certain extent. Your average AI is way better than the average schmuck on the street in basically anything you can think of - maths, programming, writing poetry, world languages, music theory. Sure there are outliers where AI is not as good as a skilled practitioner in foo, but I think the AGI bar is about being "about as good as the average human" and not showing complete supremacy in every niche. So far the world has been disrupted sure, but not ended.

ASI of course is the next thing, but that's different.

Enginerrrd•1h ago
I think the AI is only as good as the person wrangling it a lot of the time. I think it's easy for really competent people to get an inflated sense of how good the AI is in the same way that a junior engineer is often only as good as the senior leading them along and feeding them small chunks of work. When led with great foresight, careful calibration, and frequent feedback and mentorship, a mediocre junior engineer can be made to look pretty good too. But take away the competent senior and youre left pretty lacking.

I've gotten some great results out of LLM's, but thats often because the prompt was well crafted, and numerous iterations were performed based on my expertise.

You couldn't get that out of the LLM without that person most of the time.

impossiblefork•8m ago
Nah. The models are great, but the models can also write a story where characters who in the prompt are clearly specified as never having met are immediately addressing each other by name.

These models don't understand anything similar to reality and they can be confused by all sorts of things.

This can obviously be managed and people have achieved great things with them, including this IMO stuff, but the models are despite their capability very, very far from AGI. They've also got atrocious performance on things like IQ tests.

mhb•1h ago
That's a lot of confidence that this is a bubble rather than an existential race. Maybe you're making bank betting that view?
snowstormsun•1h ago
2027
mark_l_watson•1h ago
Perhaps when society balances benefits of AI against energy and environmental costs? I have worked through two ‘AI winters’ when funding dried up. This might happen again.

I think a possible scenario is that we see huge open source advances in training and inference efficiency that ends up making some of the mega-investments in AI infrastructure look silly.

What will probably ‘save’ the mega-spending is (unfortunately!) the application of AI to the Forever Wars for profit.

chvid•38m ago
When we are seeing down rounds on OpenAI. OpenAI is currently valued at 300B.
abtinf•2h ago
HN shows this as google.com, but the link is directly to nytimes.com. Bug?
stevage•2h ago
The link is a google.com redirect. You probably have an extension installed that auto-resolves such redirects.
_Algernon_•2h ago
It's a redirect.
tomhow•2h ago
We updated the link to its redirect URL.
tomhow•2h ago
At $250M, top AI salaries dwarf the Manhattan Project and the Space Race - https://news.ycombinator.com/item?id=44765193 - Aug 2025 (66 comments)
jokoon•2h ago
Meanwhile I'm not sure that training myself to do ai would increase my odds of getting a job
ivape•2h ago
At those prices they have to be hiring god gifted talent. I can’t imagine that being just a regular academic grinder with top grades. Arod made $250 million and it was considered huge news.
gosub100•2h ago
aww, they can't be sleazy CEO types who "make number go up" for 2-3 years and leave with a golden parachute?
ivape•1h ago
No, it’s a 24-year old:

https://nypost.com/2025/08/01/business/meta-pays-250m-to-lur...

dmurray•1h ago
What really, private industry pays top performing individuals more than the government ever did?
normie3000•1h ago
How can I get one of these jobs? I am currently an OK web dev.
lcnPylGDnU4H9OF•1h ago
Get a PhD in a related field like math or computer science.
seanbarry•1h ago
And have spent the last 15 years working on the cutting edge of AI research.
coderatlarge•1h ago
actually applied math or statistics.
BSOhealth•1h ago
These figures are for a very small number of potential people. This leaves out that frontier AI is being developed by an incredibly small number of extremely smart people who have migrated between big tech, frontier AI, and others.

Yes, the figures are nuts. But compare them to F1 or soccer salaries for top athletes. A single big name can drive billions in that context at least, and much more in the context of AI. $50M-$100M/year, particularly when some or most is stock, is rational.

AIPedant•1h ago
A very major difference is that top athletes bring in real tangible money via ticket / merch sales and sponsorships, whereas top AI researchers bring in pseudo-money via investor speculation. The AI money is far more likely to vanish.
brandall10•34m ago
It's best to look at this as expected value. A top AI research has the potential to bring in a lot more $$ than a top athlete, but of course there is a big risk factor on top of that.
jgalt212•3m ago
If you imagine hard enough, you can expect anything. e.g. Extraordinary Popular Delusions and the Madness of Crowds
magic_man•21m ago
Top athletes they have stats to measure. I guess for these researchers I guess there are papers? How do you know who did what with multiple authors? How do you figure out who is Jordan vs Steve Kerr?
unyttigfjelltol•1h ago
I take this as a contrarian signal that Meta has hit serious roadblocks improving their AI despite massive data advantages and are throwing a bunch of "Hail Mary" desperation passes to achieve meaningful further progress.
aleph_minus_one•1m ago
> I take this as a contrarian signal that Meta has hit serious roadblocks improving their AI despite massive data advantages

Just a thought:

Assuming that Meta's AI is actually good. Could it rather be that having access to a massive amount of data does not bring that much of a business value (including for training AIs)?

Evidence for my hypothesis: if you want to gain a deep knowledge about some complicated specific scientific topic, you typically don't want to read a lot of shallow texts tangentially related to this topic, but the few breakthrough papers and books of the smartest mind who moved the state of art in the respective area. Or some of the few survey monographs of also highly smart people who work in the respective area who have a vast overview about how these deep research breakthroughs fit into the grander scheme of things.

TheAceOfHearts•1h ago
Wonder what their contracts look like. Are these people gonna be grinding their ass off at the Meta offices working crazy hours? Does Zucc have a strong vision, leadership, and management skills to actually push and enable these people to achieve maximum success? And if so, what does that form of success look like? So far the vision that Zucc has outlined has been rather underwhelming, but maybe the vision which he shares with insiders is different from his public persona.

I can't help but think that the structure of this kinda hints at there being a bit of a scam-y element, where a bunch of smart people are trying to pump some rich people out of as much money as possible, with questionable chances at making it back. Imagine that the people on The List had all the keys needed to build AGI already if they put their knowledge together, what action do you think they would take?

walterbell•38m ago
> Imagine.. had all the keys needed

.. that had already leaked and would later plummet in value.

Compressing Icelandic name declension patterns into a 3.27 kB trie

https://alexharri.com/blog/icelandic-name-declension-trie
105•alexharri•3h ago•31 comments

The Art of Multiprocessor Programming 2nd Edition Book Club

https://eatonphil.com/2025-art-of-multiprocessor-programming.html
23•eatonphil•1h ago•4 comments

We may not like what we become if A.I. solves loneliness

https://www.newyorker.com/magazine/2025/07/21/ai-is-about-to-solve-loneliness-thats-a-problem
103•defo10•3h ago•139 comments

6 Weeks of Claude Code

https://blog.puzzmo.com/posts/2025/07/30/six-weeks-of-claude-code/
39•mpweiher•2h ago•10 comments

OpenAI's "Study Mode" and the risks of flattery

https://resobscura.substack.com/p/openais-new-study-mode-and-the-risks
59•benbreen•2d ago•34 comments

Caches: LRU vs. Random

https://danluu.com/2choices-eviction/
25•gslin•2d ago•2 comments

WebGPU enables local LLM in the browser. Demo site with AI chat

https://andreinwald.github.io/browser-llm/
5•andreinwald•34m ago•3 comments

Microsoft is open sourcing Windows 11's UI framework

https://www.neowin.net/news/microsoft-is-taking-steps-to-open-sourcing-windows-11-user-interface-framework/
68•bundie•6h ago•61 comments

Cerebras Code

https://www.cerebras.ai/blog/introducing-cerebras-code
395•d3vr•16h ago•151 comments

Palo Alto Networks agrees to buy CyberArk for $25B

https://techcrunch.com/2025/07/30/palo-alto-networks-agrees-to-buy-cyberark-for-25-billion/
45•vmatsiiako•2d ago•27 comments

ThinkPad designer David Hill spills secrets, designs that never made it

https://www.theregister.com/2025/08/02/thinkpad_david_hill_interview/
31•LorenDB•2h ago•1 comments

Aerodynamic drag in small cyclist formations: shielding the protected rider [pdf]

http://www.urbanphysics.net/2025_Formation_Paper_Preprint_v1.pdf
43•PaulHoule•3d ago•11 comments

Coffeematic PC – A coffee maker computer that pumps hot coffee to the CPU

https://www.dougmacdowell.com/coffeematic-pc.html
243•dougdude3339•16h ago•73 comments

New research finds that ivermectin could help control malaria transmission

https://www.ndm.ox.ac.uk/news/new-research-supports-ivermectin-as-an-effective-strategy-to-control-malaria-transmission
36•rguiscard•2h ago•34 comments

Hardening mode for the compiler

https://discourse.llvm.org/t/rfc-hardening-mode-for-the-compiler/87660
135•vitaut•12h ago•44 comments

Why leather is best motorcycle protection [video]

https://www.youtube.com/watch?v=xwuRUcAGIEU
126•lifeisstillgood•2d ago•99 comments

Cadence Guilty, Pays $140M for Exporting Semi Design Tools to PRC Military Uni

https://www.justice.gov/opa/pr/cadence-design-systems-agrees-plead-guilty-and-pay-over-140-million-unlawfully-exporting
18•737min•1h ago•3 comments

This Month in Ladybird

https://ladybird.org/newsletter/2025-07-31/
316•net01•8h ago•95 comments

At 17, Hannah Cairo solved a major math mystery

https://www.quantamagazine.org/at-17-hannah-cairo-solved-a-major-math-mystery-20250801/
394•baruchel•22h ago•160 comments

Palo Alto Networks closing on over $20B acquisition of CyberArk

https://www.calcalistech.com/ctechnews/article/hksugkiwxe
7•tomashertus•3d ago•0 comments

JavaScript retro sound effects generator

https://github.grumdrig.com/jsfxr/
112•selvan•4d ago•21 comments

Weather Model based on ADS-B

https://obrhubr.org/adsb-weather-model
222•surprisetalk•3d ago•35 comments

I couldn't submit a PR, so I got hired and fixed it myself

https://www.skeptrune.com/posts/doing-the-little-things/
295•skeptrune•21h ago•180 comments

Ask HN: Who is hiring? (August 2025)

206•whoishiring•23h ago•238 comments

A.I. Researchers Are Negotiating $250 Million Pay Packages

https://www.nytimes.com/2025/07/31/technology/ai-researchers-nba-stars.html
54•jrwan•3h ago•61 comments

Ethersync: Peer-to-peer collaborative editing of local text files

https://github.com/ethersync/ethersync
157•blinry•4d ago•31 comments

Yearly Organiser

https://neatnik.net/calendar/
77•anewhnaccount2•4d ago•21 comments

North Korea sent me abroad to be a secret IT worker

https://www.bbc.com/news/articles/c15wk77zxngo
47•tellarin•4h ago•21 comments

Does the Bitter Lesson Have Limits?

https://www.dbreunig.com/2025/08/01/does-the-bitter-lesson-have-limits.html
162•dbreunig•18h ago•79 comments

Ferroelectric Helps Break Transistor Limits

https://spectrum.ieee.org/negative-capacitance-schottky-limit
19•pseudolus•4d ago•1 comments