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

Ilya Sutskever to leave OpenAI

twitter.com

Post reply on HN
Print view View original

711–720 of 855 posts

Re: Ilya Sutskever to leave OpenAI

#711

Post by sebastiennight »

tsimionescu wrote: post #648
Zambyte wrote: post #621

Earlier quoted context omitted.

I'm genuinely curious: what do you expect an "AGI" system to be able to do that we can't do with today's technology?

The simplest answer, without adding any extraordinary capabilities to the AGI that veer into magical intelligence, is to have AI assistants that can seemlessly interact with technology the way a human assistant would. So, if you want to meet with someone, instead of opening you calendar app and looking for an opening, you'd ask your AGI assistant to talk to their AGI assistant and set up a 1h meeting soon. Or, instea…

The examples you're providing seem to have been thoroughly solved already.

I'm at the European AI Conference for our startup tomorrow, and they use a platform that just booked me 3 meetings automatically with other people there based on our availability... It's not rocket science.

And you don't even need those narrow tools. You could easily ask GPT-4o (or lesser versions) something along the lines of :

> "you're going to interact with another AI assistant to book meetings for me: [here would be the details about the meeting]. Come up with a protocol that you'll send to the other assistant so it can understand what the meetings are about, communicate you their availability, etc. I want you to come up with the entire protocol, send it, and communicate with the other assistant end-to-end. I won't be available to provide any more context ; I just want the meeting to be booked. Go."

Re: Ilya Sutskever to leave OpenAI

#712

Post by somebodythere »

jpalawaga wrote: post #703
boringg wrote: post #691

Earlier quoted context omitted.

I've thought about this one for a long time having lived in both SV and Canada. It is a complicated one but there are a handful of critical road blocks in Canada that make it more challenging. (1) Access to size of market even if online being US vs 'foreign' has advantages in political arena/regulatory benefits (2) Significant tax advantages for US investors vs limited tax advantages for Canada (Angel+VC) (3) Risk Ap…

if the idea that high taxes disincentivizes people from building stuff, california would be a wasteland. but that's not what we see. people build because they still have a chance a making a lot of money. also, like canada can build successfully tech companies. yes, I realize there should be more canadian tech darlings, but I don't think it has to do with high taxes so much as it has to do with Canadians being comfort…

California benefits from a dominant market position. If you have the choice to found where the investors and the talent are, why would you pick Canada when the tax is the same?

Re: Ilya Sutskever to leave OpenAI

#713

Post by boringg »

jpalawaga wrote: post #703
boringg wrote: post #691

Earlier quoted context omitted.

I've thought about this one for a long time having lived in both SV and Canada. It is a complicated one but there are a handful of critical road blocks in Canada that make it more challenging. (1) Access to size of market even if online being US vs 'foreign' has advantages in political arena/regulatory benefits (2) Significant tax advantages for US investors vs limited tax advantages for Canada (Angel+VC) (3) Risk Ap…

if the idea that high taxes disincentivizes people from building stuff, california would be a wasteland. but that's not what we see. people build because they still have a chance a making a lot of money. also, like canada can build successfully tech companies. yes, I realize there should be more canadian tech darlings, but I don't think it has to do with high taxes so much as it has to do with Canadians being comfort…

You misunderstand my statement on tax.

Taxs on investors are quite high in Ontario/Canada compared to California. Not only does this minimize the outcome for the investor - it decreases the risk for making larger bets on big outcomes. In terms of exits -- you have a smaller playing field and fewer buyers being based in Canada vs USA. All the things add up to make a smaller opportunity for investors and builders and you work harder to pay more taxes to the government.

In terms for your ambition argument -- that could be an inherent problem in Canadian culture that no one wants to change the status quo - it is definitely a different culture than SV. The largest city is captured by financial industry for the most part which doesn't bode well for innovation.

Re: Ilya Sutskever to leave OpenAI

#714

Post by pixl97 »

sangnoir wrote: post #317
pixl97 wrote: post #121

Earlier quoted context omitted.

If the barrier to entry is lowered I don't see Nvidia losing a dime... Jevons paradox tells us that further massive amounts of training will occur in multi-modality and grounding models to reality. That would then kick off the race to true reasoning and AGI/ASI, and if achieved, all bets are off after that point. If no training breakthrough occurs, then your second option is much more likely.

> Jevons paradox tells us that further massive amounts of training will occur in multi-modality and grounding models to reality I was thinking along the lines of how Sun Microsystems and big-iron were dethroned by good-enough commodity x86 servers + Linux. The two coexisted... for a short while.

All while x86 morphed to x64 and gained back many of the features of big iron.

Much of the problem with PC software at the time was one of scaling. Getting software to use multiple processors effectively took decades, but the cost difference and ability to actually get the hardware cheap allowed it to win the market... then grow back into systems with hundreds of cores and terabytes of memory because the needs for big iron didn't go away.

With AI/training/LLMs/NNs scale, at least appears, intrinsic. We see this in the animal world. We see insects with a very basic brain capable of interacting with the world and surviving. Then we see mammals with larger brains capable of far more as the scale/complexity/interconnectivity of their brains increases. At least from what we can tell, throwing more power at any given problem will give us a better solution.

Re: Ilya Sutskever to leave OpenAI

#715

Post by Symmetry »

datameta wrote: post #706
Symmetry wrote: post #671

Earlier quoted context omitted.

A working memory that can preserve information indefinitely outside a particular context window and which can engage in multi-step reasoning that doesn't show up in its outputs. GPT4o's context window is 128k tokens which is somewhere on the order of 128kB. Your brain's context window, all the subliminal activations from the nerves in your gut and the parts of your visual field you aren't necessarily paying attention…

I'm curious what resources led you to calculate a 2MB context window, I'd like to learn more.

Looking up an estimate of the brain's input bandwidth at 10 million bits per second and multiplying by the second or two a subliminal stimuli can continue to affect a person's behavior. This is a very crude estimate and probably an order of magnitude off, but I don't think many orders of magnitude off.

Re: Ilya Sutskever to leave OpenAI

#716

Post by llm_trw »

titanomachy wrote: post #460
llm_trw wrote: post #409

Earlier quoted context omitted.

> 2-3x the money. That's if you're stuck in tech support. When you start doing actual ground breaking work it starts at x10 and goes up significantly.

If you’re a top-tier AI researcher like Ilya, for sure. I was thinking about your run-of-the mill FAANGish senior engineer.

E6 in facebook is something like USD 750k total comp.

In Canada you'd be lucky to get USD 180k total comp.

It really is x5 to x10 for seniour programmers.

Re: Ilya Sutskever to leave OpenAI

#717

Post by llm_trw »

tbossanova wrote: post #436
llm_trw wrote: post #408

Earlier quoted context omitted.

10B is going to old mates mates. No new startups are getting it. It's also like pulling teeth trying to explain to people that if we don't offer compensation commensurate with what they get in the US people will just leave for the US. There is some form of brain damage where even people who know how to code assume that because you can get a crud developer for $80k a year you should get an AI researcher for $150,000.…

Compensation isn’t just money. Staying near family and friends is hard to measure. I’ve never moved country for $ alone.

Great. But anyone good gets offered so much they move eventually.

It might not be a year after uni, might not be 10, but eventually they will move because the pay in the US is just so much better than anywhere else.

Re: Ilya Sutskever to leave OpenAI

#718

Post by jagrsw »

TimPC wrote: post #702
jagrsw wrote: post #627

Earlier quoted context omitted.

Some first ideas coming to mind: Engineering Level: Solve CO2 Levels End sickness/death Enhance cognition by integrating with willing minds. Safe and efficient interplanetary travel. Harness vastly higher levels of energy (solar, nuclear) for global benefit. Science: Uncover deeper insights into the laws of nature. Explore fundamental mysteries like the simulation hypothesis, Riemann hypothesis, multiverse theory, an…

The problem with CO2 levels is that no one likes the solution not that we don't have one. I highly doubt adding AGI to the mix is going to magically make things better. If anything we'll just burn more CO2 providing all the compute resources it needs. People want their suburban lifestyle with their red meat and their pick-up truck or SUV. They drive fuel inefficient vehicles long-distances to urban work environments…

Assuming embedded AI in every piece of robotics - sometimes directly, sometimes connected to a central server (this is doable even today) - it'll revolutionize industries: human-less mining, processing, manufacturing, services, and transportation. These factories would eventually produce and install enough solar power or build sufficient nuclear plants and energy infrastructure, making energy clean and free.

With abundant electric cars (at this future point in time) and clean electricity powering heating, transportation, and manufacturing, some AIs could be repurposed for CO2 capture.

It sounds deceptively easy, but from an engineering standpoint, it likely holds up. With free energy and AGI handling labor and thinking, we can achieve what a civilization could do and more (cause no individual incentives come into play).

However, human factors could be a problem: protests (luddites), wireheading, misuse of AI, and AI-induced catastrophes (alignment).

Re: Ilya Sutskever to leave OpenAI

#719

Post by ookdatnog »

nabla9 wrote: post #447
zoogeny wrote: post #362

Interesting, both Karpathy and Sutskever are gone from OpenAI now. Looks like it is now the Sam Altman and Greg Brockman show. I have to admit, of the four, Karpathy and Sutskever were the two I was most impressed with. I hope he goes on to do something great.

Top 6 science guys are long gone. Open AI is run by marketing, business, software and productization people. When the next wave of new deep learning innovations sweeps the world, Microsoft eats whats left of them. They make lots of money, but don't have future unless they replace what they lost.

> They make lots of money

Will they though? Last I heard OpenAI isn't profitable, and I don't know if it's safe to assume they every will be.

People keep saying that LLMs are an existential threat to search, but I'm not so sure. I did a quick search (didn't verify in any way if this is a feasible number) to find that Google on average makes about 30 cents in revenue per query. They make a good profit on that because processing the query costs them almost nothing.

But if processing a query takes multiple seconds on a high-end GPU, is that still a profitable model? How can they increase revenue per query? A subscription model can do that, but I'd argue that a paywalled service immediately means they're not a threat to traditional ad-supported search engines.

Re: Ilya Sutskever to leave OpenAI

#720

Post by pcurve »

sk11001 wrote: post #525
pcurve wrote: post #421

Earlier quoted context omitted.

Resignations lead to more resignations....unless mgmt. can get on top of it and remedy it quickly, which rarely happens. I've seen it happen way too many times working 25 years in tech.

This might not be bad from the perspective of the remaining employees, it might be that the annoying people are leaving the room.

In my experience, good ones leave first, followed by those who enjoyed working with them and or ones who are not longer able to get work done.
Post reply on HN