Live data from Hacker News

Ilya Sutskever to leave OpenAI

twitter.com

811–820 of 855 posts

Re: Ilya Sutskever to leave OpenAI

#811
post #734
post #562

Earlier quoted context omitted.

Isnt those just branches of the same company?

Now I'm curious how the NBA is structured. I always thought the "ownership" was who paid the players and signed the contracts (in essence, an NBA team is a company), and the NBA simply enforced the rules of the contracts (the "templating", if you will), but I'm sure it's much more complex than that. I made the analogy because NBA players often move from team to team and there are no non-competes keeping them from pla…

Maybe the NBA is a franchise and each team a franchisee?

Re: Ilya Sutskever to leave OpenAI

#812

Earlier quoted context omitted.

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…

This is definitely an interesting way to look at it. My initial reaction is to consider that I can enhance the capabilities of a system without increasing its inteligence. For example, if I give a monkey a hammer, it can do more than it could do when it didn't have the hammer, but it is not more intelligent (though it could probably learn things by interacting with the world with the hammer). That leads me to think:…

I'm sure some aspects of this can be achieved by manually programming GPT-4 links to other specific services. And obviously, some interaction tools would have to be written manually even for an AGI.

The difference though is the amount of work. Today if you wanted GPT-4 to work as I describe, you would have to write an integration for Gmail, another one for Office365, another one for Proton etc. You would probably have to create a management interface to give access to your auth tokens for each of these to OpenAI so they can activate these interactions. The person you want to sync with would have to do the same.

In contrast, an AGI that only has average human intelligence, or even below, would just need access to, say, Firefox APIs, and should easily be able to achieve all of this. And it would work regardless if the other side is a different AGI using a different provider, or even if they are just a regular human assistant.

Re: Ilya Sutskever to leave OpenAI

#813
post #633

Earlier quoted context omitted.

Not saying it’s going to be the same, but I’m sure computing progress looked pretty unimpressive from, say, 1975 to 1990 for the uninitiated. By the 90s they were still mainly used as fancy typewriters by “normal” people (my parents, school, etc) although the ridiculous potential was clear from day one. It just took a looong time to go from pong to ping and then to living online. I’m still convinced even this stage i…

It was only 11 years from pong to ping.

You and your family and friends were online in 1983? That’s quite remarkable.

Re: Ilya Sutskever to leave OpenAI

#814

Earlier quoted context omitted.

> That makes zero sense, governments invest in education to improve their own country The idea is precisely that not having SV types around _improves_ the country, i.e. makes it closer to the preferences of Canadians. And yes, having a foreign tech worker doing 9-to-5 in a large legacy company for thoroughly average salaries is very different from having a SV-style startup culture. There is very little process in Can…

> Compare: if a large minority of Icelanders wanted to work for the Baby (which Iceland doesn't have), theb stopping the brain drain (convincing them to work in the Merchant Fleet) is the best outcome, but funneling them out (training them in merchant navigation and watching them join the Danish Navy) would still be preferable to them engaging in their desired behavior anyway (form their own pirate gang preying on th…

I see why you'd read it that way, but it's meant as a metaphor, not an analogy, to help elucidate that a government may take steps to try and bring about their preferred outcome of retaining people, while not bringing about what is, from their perspective, even more undesirable outcome.

It's not perfect, but neither is anything else I couls.come up with. Take the following:

- Persia would prefer many trained accountants so that PersianAccountants, the shah's preferred supplier of accounting technology, can hire cheaply from a large talent pool.

- Ambitious, trained accountants leave for the U.S. to work on DisruptiveAccounting.io, earning big bucks and disrupting the U.S. accounting sector.

- If Persia changes the kingdom's procedures and incentives, the same accountants would stay and found DisruptiveAccounting East, instead of working for PersianAccountants. This would be strictly worse for the shah than letting them leave.

The problem is, if I were to use this metaphor, people would get hung up on the difference between democracy and monarchy (preferences of one autocrat vs. that of the majority) and most Americans just straught up do not understand why anybody, much less the majority, would prefer not disrupting the accounting sector.

I.e. if they don't understand what I'm saying about Canada not liking the third option, the metaphor falls flat: they also won't see why the shah doesn't like the third option.

So I had to look for a metaphor where the obvious alternative is undesirable to most Americans. Hence piracy. The problem is that there's another reading now (the software engineers will become criminals).

Do you have a metaphor that would avoid both issues? I'd love to hear it!

Re: Ilya Sutskever to leave OpenAI

#815
post #550

Earlier quoted context omitted.

There goes the so called superalignment: Ilya Jan Leike William Saunders Leopold Aschenbrenner All gone

So Satya Nadella paid $13 billion to have....Sam Altman :-))

Nah, they paid for the brand.

Re: Ilya Sutskever to leave OpenAI

#816
post #626

Earlier quoted context omitted.

Checking MSFT price, seems like the market thinks they made the right move and the shareholders are for sure seeing a return.

The market thinks Tesla is worth more than all other automakers combined, that GameStop is a reasonable investment, and laying off engineers is great!

But the market is always right.

Re: Ilya Sutskever to leave OpenAI

#817

Earlier quoted context omitted.

“Its biggest issue is in the productization.” That’s not true at all. The biggest issue is that it doesn’t work. You can’t actually trust ai systems and that’s not a product issue.

> You can’t actually trust ai systems For a lot of (very profitable) use cases, hallucinations and 80/20 are actually more than good enough. Especially when they are replacing solutions that are even worse.

What use cases? This kind of thing is stated all the time, never any examples.

Re: Ilya Sutskever to leave OpenAI

#818

Earlier quoted context omitted.

“Its biggest issue is in the productization.” That’s not true at all. The biggest issue is that it doesn’t work. You can’t actually trust ai systems and that’s not a product issue.

> You can’t actually trust ai systems For a lot of (very profitable) use cases, hallucinations and 80/20 are actually more than good enough. Especially when they are replacing solutions that are even worse.

What are examples of these (very profitable) use cases?

Producing spam has some margin on it, but is it really very profitable? And else?

Re: Ilya Sutskever to leave OpenAI

#819

Earlier quoted context omitted.

“Its biggest issue is in the productization.” That’s not true at all. The biggest issue is that it doesn’t work. You can’t actually trust ai systems and that’s not a product issue.

It works fine for some things. You just need a clearly defined task where LLM + human reviewer is on average faster (ie cheaper) than a human doing the same task themselves without that assistance.

Given the fact that you need to review, research, and usually correct every detail of AI output, how can that be faster than just doing it right yourself in the first place? Do you have some examples of such tasks?

Re: Ilya Sutskever to leave OpenAI

#820

Why do people treat these technologists doing career moves, as if this was lineup changes in a major league sports teams? Are these "first name" (ugh) "influencers" smart? Sure. Smart is not that rare. These people are technologists like most of you, they aren't notably smarter, they just got lucky in their career direction and specialization. They aren't business geniuses. They're just people filling roles. Do chang…

No, they are people defining companies which is a significantly less fungible placement and a self-defined role.
Post reply on HN