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

Ilya Sutskever's SSI Inc raises $1B

reuters.com

Post reply on HN
Print view View original

661–670 of 803 posts

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

#661

Post by steveBK123 »

dr_dshiv wrote: post #630
steveBK123 wrote: post #440

Earlier quoted context omitted.

How is this any different than the (lack of) business model of all the voice assistants? How good does it have to be, how many features does it have to have, how accurate does its need to be.. in order for people to pay anything? And how much are people actually willing to spend against the $XX Billion of investment? Again it just seems like "sell to AAPL/GOOG/MSFT and let them figure it out".

Congrats on 10k karma :) One could ask: how is this different from automatic call centers? (eg “for checking accounts, push 1…”) well, people hate those things. If one could create an automated call center that people didn’t hate, it might replace a lot of people.

Now call centers, not sexy, but the first rational achievable use case mentioned for LLMs in an HN response I've seen in a while!

The global call center market is apparently $165B/year revenue, and let's be honest even the human call center agents aren't great. So market is big and bar is low!

However, we are clearly still quite far from LLMs being a) able to know what they don't know / not hallucinate b) able to quickly/cheaply/poorly be trained the way you could a human agent c) actually be as concise and helpful as an average human.

Also it is obviously already being tried, given the frequent Twitter posts with screenshots of people jailbreaking their car dealership chat bot to give coding tips, etc.

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

#662

Post by Obscurity4340 »

robertlagrant wrote: post #642
__loam wrote: post #618

Earlier quoted context omitted.

Maybe I'm naive but there seems to be way too much financial incentive in this space for CUDA to continue to be the top dog. Just like microprocessors, these devices are going to get commodified, standardized, open sourced, etc. Nvidia making massive profits is a sign of huge market inefficiency and potential opportunities for competition.

> Nvidia making massive profits is a sign of huge market inefficiency and potential opportunities for competition. Yep, but just as the first reading glasses were only available to the wealthy, and now anyone can have them, the inefficiency takes time to work out. It'll take a long time, especially given how vertically integrated Nvidia are.

How could people who actually like NEEDED glasses be productive or get around without them prior to glasses being developed and commodified?

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

#663

Post by brigadier132 »

Yizahi wrote: post #640
brigadier132 wrote: post #405

Earlier quoted context omitted.

Why are you assuming this hypothetical intelligence will have any motivations beyond the ones we give it? Human's have complex motivations due to evolution, AI motivations are comparatively simple since they are artificially created.

Any intelligence on the level of average human and for sure on the level above it will be able to learn. And learning means it will acquire new motivations, among other things. Fixed motivation thing is simply a program, not AI. A very advanced program maybe, but ultimately just a scaled up version of the stuff we already have. AI will be different, if we will create it.

> And learning means it will acquire new motivations

This conclusion doesn't logically follow.

> Fixed motivation thing is simply a program, not AI

I don't agree with this definition. AI used to be just "could it solve the turing test". Anyway, something with non-fixed motivations is simply just not that useful for humans so why would we even create it?

This is the problem with talking about AI, a lot of people have different definitions of what AI is. I don't think AI requires non-fixed motivations. LLMs are definitely a form of AI and they do not have any motivations for example.

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

#664

Post by trashtester »

robertlagrant wrote: post #641
trashtester wrote: post #639

Earlier quoted context omitted.

My guess is that whoever develops superintelligence first will not release it to the public, but rather use it for their own purposes to gain an edge. They may still release AI products to the public that are good enough and cheap enough to prevent competitors from being profitable or receive funding (to prevent them from catching up), but that's not where the value would come from. Just as an example, let's say xAI…

> whoever develops superintelligence first will not release it to the public, but rather use it for their own purposes to gain an edge We have no evidence that superintelligence will be developed. There's no "first". There's only "remote possibility".

> We have no evidence that superintelligence will be developed.

Fundamentally, we have no evidence of anything that will happen in the future. All we do is to extrapolate from the past through the present, typically using some kind of theory of how the world operates.

The belief that we will eventually (whether it's this year or in 1000+ years), really only hinges on the following 3 assumptions:

1) The human brain is fully material (no divine souls is necessary for intelligence)

2) The human brain does not represent a global optimum for how intelligent a material intelligence-having-object can be.

3) We will eventually have the ability to build intelligence-having-objects (similar or different from the brain) that not only can do the same as a brain (that would be mere AGI), but also surpass it in many ways.

Assumptions 1 and 2 have a lot of support in the current scientific consensus. Those who reject them either do not know the science or they have a belief system that would be invalidated if one of those assumptions were true. (That could be anything from a Christian belief in the soul to an ideological reliance of a "Tabula Rasa" brain).

Assumption 3 is mostly techno-optimism, or an extrapolation of the trend that we are able to build ever more advanced devices.

As for WHEN we get there, there is a fourth assumption required for it to happens soon:

4. For intelligence-having-objects to do their thing, they don't need some exotic mechanisms we don't yet need to build. For instance, there is no need to build large quantum computers for this.

This assumption is mostly about belief, and we really don't know.

Yet, given the current rate of progress, and if we accept assumptions 1-3, I don't think assumption 4 is unreasonable.

If so, it's not unreasonable to assume that our synthetic brains reach roughly human level intelligence when their size/complexity becomes similar to that of human brains.

Human brains have ~200 trillion synapses. That's about 100x-1000x more than the latest neural nets that we're building.

Based only on scale, current nets (GPT-4 generation) should have total capabilities similar or slightly better than a rat. I think that's not very far off from what we're seeing, even if the nets tend to have those capabilities linked to text/images rather than the physical world that a rat navigates.

In other words, I think we DO have SOME evidence (not conclusive) that the capabilities of a neural net can reach similar "intelligence" to animals with a similar number of synapses.

So IF that hypothesis holds true, and given assumptions 1-3 above, there is a fair possibility that human level intelligence will be reached when we scale up to about 200 trillion weights (and have the ability to train such nets).

And currently, several of the largest and most valuable companies in the world are making a huge gamble on this being the case, with plans to scale up nets by 100x over the next few years, which will be enough to get very close to human brain sized nets.

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

#665

Post by HarHarVeryFunny »

accountnum wrote: post #634
HarHarVeryFunny wrote: post #388

Earlier quoted context omitted.

> And again, human brains are clearly limited in the number of steps it can compute without writing something down No - there is a loop between the cortex and thalamus, feeding the outputs of the cortex back in as inputs. Our brain can iterate for as long as it likes before initiating any motor output, if any, such as writing something down.

You seem to repeatedly insist that hidden computation is a distinction of any relevance whatsoever. First of all, your understanding of the architecture itself is mistaken. A transformer can iterate endlessly because each token it produces allows it a forward pass, and each of these tokens is postpended to its input in the next inference. That's the autoregressive in autoregressive transformer, and the entire reason…

Sure - a transformer can iterate endlessly by generating tokens, but this is no substitute for iterating internally and maintaining internal context and goal-based attention.

One reason why just blathering on endlessly isn't the same as thinking deeply before answering, is that it's almost impossible to maintain long-term context/attention. Try it. "Think step by step" or other attempts to prompt the model into generating a longer reply that builds upon itself, will only get you so far because keeping a 1-dimensional context is no substitute for the thousands of connections we have in our brain between neurons, and the richness of context we're therefore able to maintain while thinking.

The reasoning weakness of LLMs isn't limited to "some domains" that they had less training data for - it's a fundamental architecturally-based limitation. This becomes obvious when you see the failure modes of simple problems like "how few trips does the farmer need to cross the river with his chicken & corn, etc" type problems. You don't need to morph the problem to require out-of-distribution knowledge to get it to fail - small changes to the problem statement can make the model state that crossing the river backwards and forwards multiple times without loading/unloading anything is the optimal way to cross the river.

But, hey, no need to believe me, some random internet dude. People like Demis Hassabis (CEO of DeepMind) acknowledge the weakness too.

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

#666

Post by jowdones »

trashtester wrote: post #664
robertlagrant wrote: post #641

Earlier quoted context omitted.

> whoever develops superintelligence first will not release it to the public, but rather use it for their own purposes to gain an edge We have no evidence that superintelligence will be developed. There's no "first". There's only "remote possibility".

> We have no evidence that superintelligence will be developed. Fundamentally, we have no evidence of anything that will happen in the future. All we do is to extrapolate from the past through the present, typically using some kind of theory of how the world operates. The belief that we will eventually (whether it's this year or in 1000+ years), really only hinges on the following 3 assumptions: 1) The human brain is…

>> Fundamentally, we have no evidence of anything that will happen in the future.

Yeah, by this line of thought Jesus will descend from Heaven and save us all.

By the same line of fantasy, "give us billions to bring AGI", why not "gimme a billion to bring Jesus. I'll pray really hard, I promise!"

It's all become a disgusting scam, effectively just religious. Believe in AGI that's all there is to it. In practice it's just as (un) likely as scientists spontaneously creating life out of primordial soup concoctions.

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

#667

Post by steveBK123 »

ithkuil wrote: post #605
steveBK123 wrote: post #520

Earlier quoted context omitted.

I do worry that unless some of these LLMs actually find a revenue model soon, the self reinforcing bubble is going to pop broadly. GPUs, servers, datacenters, fabs, power generation/transmission, copper/steel/concrete.. All to train models in an arms race because someone somewhere is going to figure out how to monetize and no one wants to miss the boat.

If the bubble pops, would that bring the price for at least part of that hardware down and thus enable a second round of players (who were locked out from the race now) to experiment a little bit more and perhaps find something that works better?

Maybe?

My outsider observation is that we have a decent number of players roughly tied at trying to produce a better model. OpenAI, Anthropic, Mistral, Stability AI, Google, Meta, xAI, A12, Amazon, IBM, Nvidia, Alibaba, Databricks, some universities, a few internal proprietary models (Bloomberg, etc) .. and a bunch of smaller/lesser players I am forgetting.

To me, the actual challenge seems to be figuring out monetizing.

Not sure the 15th, 20th, 30th LLM model from lesser capitalized players is going to be as impactful.

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

#668

Post by steveBK123 »

damanoman wrote: post #532
steveBK123 wrote: post #517

Earlier quoted context omitted.

I guess the questions then are - why is it 2x the competing price, why do you willing pay 2x, and how many people are willing to pay that 2x? Consider they are competing against the Lyft/Uber asset-light model of relying on "contractors" who in many cases are incapable of doing the math to realize they are working for minimum wage...

All those businesses are predatory. It’s so crazy.

I used to call it the re-intermediation economy.

Taking a cut of existing businesses/models (taxi, delivery, b&b, etc).

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

#669

Post by trashtester »

dcchambers wrote: post #336
gigatexal wrote: post #306

This being ycombinator and as such ostensibly has one or two (if not more) VCs as readers/commentators … can someone please tell me how these companies that are being invested in in the AI space are going to make returns on the money invested? What’s the business plan? (I’m not rich enough to be in these meetings) I just don’t see how the returns will happen. Open source LLMs exist and will get better. Is it just tha…

I also don't understand it. If AGI is actually reached, capital as we know it basically becomes worthless. The entire structure of the modern economy and the society surrounding it collapses overnight. I also don't think there's any way the governments of the world let real AGI stay in the hands of private industry. If it happens, governments around the world will go to war to gain control of it. SSI would be nationa…

> If AGI is actually reached, capital as we know it basically becomes worthless.

If ASI is reached but is controlled by only a few, then ASI may become the most important form of capital of all. Resources, land and pre-existing installations will still be important, though.

What will truly suffer if the ASI's potential is realized, is the value of labor. If anything, capital may become more important than before.

Now this MAY be followed by attempts by governments or voters to nationalize the AI. But it can also mean that whoever is in power decides that it becomes irrelevant what the population wants.

Particularly if the ASI can be used to operate robotic police capable of pacifying the populace.

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

#670

Post by trashtester »

morgante wrote: post #432
convolvatron wrote: post #424

Earlier quoted context omitted.

if agi is commoditized and labor is useless, what does anyone need capital for? paying ad time on monopolized social media channels?

Commoditized doesn't mean 0 capex. Literal commodities can in fact be very capital intensive (ex. offshore oil rigs). In this case, you need capital to stockpile the GPUs.

I've put some of my savings in commodities (mines, etc).....

If ASI and the ability to build robots becomes generally available and virtually free (and if the exponential growth stops), the things that retain their value will be land and raw materials (including raw materials that contain energy).

Post reply on HN