Live data from Hacker News

Ilya Sutskever's SSI Inc raises $1B

reuters.com

201–210 of 803 posts

Re: Ilya Sutskever's SSI Inc raises $1B

#201

Earlier quoted context omitted.

Which basic counting tests do they still fail? Recent examples I've seen fall well within the range of innumeracy that people routinely display. I feel like a lot of people are stuck in the mindset of 10 years ago, when transformers weren't even invented yet and state-of-the-art models couldn't identify a bird, no matter how much capabilities advance.

> Recent examples I've seen fall well within the range of innumeracy that people routinely display. But the company name specifically says "superintelligence" The company isn't named "as smart as the average redditor, Inc"

Right. They don't think that state-of-the-art models are already superintelligent, they're aiming to build one that is.

Re: Ilya Sutskever's SSI Inc raises $1B

#202

Lots of comments either defending this ("it's taking a chance on being the first to build AGI with a proven team") or saying "it's a crazy valuation for a 3 month old startup". But both of these "sides" feel like they miss the mark to me. On one hand, I think it's great that investors are willing to throw big chunks of money at hard (or at least expensive) problems. I'm pretty sure all the investors putting money in…

I think the plan is to raise a lot of cash and then more and then maybe something comes up that brings us closer to AGI(i.e something better than LLM). The investors know that AGI is not really the goal but they can’t miss the next trillion dollar company.

Re: Ilya Sutskever's SSI Inc raises $1B

#203

Earlier quoted context omitted.

I wonder if "Super Intelligence" means anything .. just LLMs, or maybe they are pursuing new architectures and shooting for AGI ?

In 2022 Ilya Sutskever claimed there wasn't a distinction: > It may look—on the surface—that we are just learning statistical correlations in text. But it turns out that to ‘just learn’ the statistical correlations in text, to compress them really well, what the neural network learns is some representation of the process that produced the text. This text is actually a projection of the world. ( https://www.youtube.co…

> newer LLMs appear to be great at arithmetic, but they still fail basic counting tests

How does the performance of today's LLMs contradict Ilya's statement?

Re: Ilya Sutskever's SSI Inc raises $1B

#205
post #121

Earlier quoted context omitted.

The neural networks in human brains are very different from artificial neural networks though. In particular, they seem to learn in a very different way than backprop. But there is no reason the company can't come up with a different paradigm.

that is very weak evidence for the impossibility claim

It was refuting the weak evidence for possibility stated above.

Re: Ilya Sutskever's SSI Inc raises $1B

#207
post #171

“…a straight shot to safe superintelligence and in particular to spend a couple of years doing R&D on our product before bringing it to market," Gross said in an interview.” A couple years ??

They’d need a year or two just to rebuild a ChatGPT-level LLM, and they want to go way beyond that.

a current-day* ChatGPT-level LLM

At a time when things are advancing at breakneck speed. Where is the goalpost going to be in 2 years time?

Re: Ilya Sutskever's SSI Inc raises $1B

#208
post #66
post #48

Earlier quoted context omitted.

Talent attracts capital. Ilya is a legendary visionary, with a proven track record of turning billions into hundreds of billions. Of course he can raise unlimited money.

There is so much talent in the world that didn’t join PayPal and get silicon valley investors and go on to make billions of dollars and found other companies. The PayPal mafia includes Elon Musk, Peter Thiel, etc. They now parlayed that capital into more platforms and can easily arrange investments. Heck Peter Thiel even works with governments (Palantir) and got J D Vance on Trump’s ticket, while Elon might be in his…

Those people weren’t handed that success. You are acting as if they were born billionaires, which is far from true.

It’s not personally my goal to amass immense wealth and start giant companies (I would rather work minimally and live hedonically) but I am impressed by those that do so.

Re: Ilya Sutskever's SSI Inc raises $1B

#209
post #61

Earlier quoted context omitted.

This feels like a situation with a sold out train to a popular destination, where people are already reselling their tickets for some crazy markup, and then suddenly railway decides to add one more train car and opens flash ticket sale. Investors feeling missing out on OpenAI and others are now hoping to catch this last train ticket to the AI.

Isn't that what happened to Evergrande

Evergrande imploded because of massive amounts of debt that they had been rolling for years. Continually rolling this massive debt was working till property demand slowed and their revenues couldn't keep up adequately to qualify them to issue new debt.

Re: Ilya Sutskever's SSI Inc raises $1B

#210

Earlier quoted context omitted.

In 2022 Ilya Sutskever claimed there wasn't a distinction: > It may look—on the surface—that we are just learning statistical correlations in text. But it turns out that to ‘just learn’ the statistical correlations in text, to compress them really well, what the neural network learns is some representation of the process that produced the text. This text is actually a projection of the world. ( https://www.youtube.co…

Which basic counting tests do they still fail? Recent examples I've seen fall well within the range of innumeracy that people routinely display. I feel like a lot of people are stuck in the mindset of 10 years ago, when transformers weren't even invented yet and state-of-the-art models couldn't identify a bird, no matter how much capabilities advance.

> Recent examples I've seen fall well within the range of innumeracy that people routinely display.

Here's GPT-4 Turbo in April botching a test almost all preschoolers could solve easily: https://substackcdn.com/image/fetch/f_auto,q_auto:good,fl_pr...

I have not used LLMs since 2023, when GPT-4 routinely failed almost every counting problem I could think of. I am sure the performance has improved since then, though "write an essay with 250 words" still seems unsolved.

The real problem is that LLM providers have to play a stupid game of whack-a-mole where an enormous number of trivial variations on a counting problem need to be specifically taught to the system. If the system was capable of true quantitative reasoning that wouldn't be necessary for basic problems.

There is also a deception is that "chain of thought" prompting makes LLMs much better at counting. But that's cheating: if the LLM had quantitative reasoning it wouldn't need a human to indicate which problems were amenable to step-by-step thinking. (And this only works for O(n) counting problems, like "count the number of words in the sentence." CoT prompting fails to solve O(nm) counting problems like "count the number of words in this sentence which contain the letter 'e'" For this you need a more specific prompt, like "First, go step-by-step and select the words which contain 'e.' Then go step-by-step to count the selected words." It is worth emphasizing over and over that rats are not nearly this stupid, they can combine tasks to solve complex problems without a human holding their hand.)

I don't know what you mean by "10 years ago" other than a desire to make an ad hominem attack about me being "stuck." My point is that these "capabilities" don't include "understands what a number is in the same way that rats and toddlers understand what numbers are." I suspect that level of AI is decades away.

Post reply on HN