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

John Carmack's new AGI company, Keen Technologies, has raised a $20M round

twitter.com

Post reply on HN
Print view View original

241–250 of 643 posts

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#241

Post by heavenlyblue »

kromem wrote: post #240
Jack000 wrote: post #116

I'm very optimistic for near-term AGI (10 years or less). Even just a few years ago most in the field would have said that it's an "unknown unknown", we didn't have the theory or the models, there was no path forward and so it was impossible to predict. Now we have a fairly concrete idea of what a potential AGI might look like - an RL agent that uses a large transformer. The issue is that unlike supervised training y…

There's a hardware component here too though. I think hybrid photonic AI chips handling some of the workload are supposed to hit in 2025 at the latest, and some of the research on gains is very promising. So we may see timelines continue to accelerate as broader market shifts occur outside just software and models.

Which research?

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#242

Post by fatherzine »

Jack000 wrote: post #116

I'm very optimistic for near-term AGI (10 years or less). Even just a few years ago most in the field would have said that it's an "unknown unknown", we didn't have the theory or the models, there was no path forward and so it was impossible to predict. Now we have a fairly concrete idea of what a potential AGI might look like - an RL agent that uses a large transformer. The issue is that unlike supervised training y…

Not sure if "optimistic" is the proper word here. Perhaps "scared senseless in the end-of-mankind kind of way" is more appropriate?

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#243

Post by dqpb »

mellosouls wrote: post #87

I am not saying the intention here is the same, but the headline doesn't inspire confidence. There's something incredibly creepy and immoral about the rush to create then commercialise sentient beings. Let's not beat about the bush - we are basically talking slavery. Every "ethical" discussion on the matter has been about protecting humans, and none of it about protecting the beings we are in a rush to bring into lif…

If AGI is possible, it’s immoral not to create it.

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#244

Post by version_five »

xtracto wrote: post #143
LudwigNagasena wrote: post #74

That’s crazy money for a vaporware seed round, isn’t it?

Most of VCs funding seed rounds do it mainly for the Team. As long as the team has OK credentials and the idea is not a hard stop (illegal or shady stuff) most will likely provide money. Given the John Carmack name... I can see why ANYONE would love to throw money to a new entrepreneurship idea.

> Most of VCs funding seed rounds do it mainly for the Team.

I know this to be true and it makes a lot of sense for the average VC backed startup with some founders that are not famous but have a record of excellence in career/academy/open source or whatever.

I'd be curious to see how it translates to superstar or famous founders, that have already had a success in the 99.99th percentile (or whatever the bar is to be a serious outlier). I doubt it does, but I have no data one way or the other.

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#245

Post by blibble »

semi-extrinsic wrote: post #139
Tepix wrote: post #7

I wonder if Carmacks moral compass is in order. First he sticks around at Facebook, now he endangers humanity with AGI. And i'm only half joking.

The meme that AGI, if we ever have it, will somehow endanger humanity is just stupid to me. For one, the previous US president is the perfect illustration that intelligence is neither sufficient nor necessary for gaining power in this world. And we do in fact live in a world where the upper echelons of power mostly interact in the decidedly analog spaces of leadership summits, high-end restaurants, golf courses and c…

none of these things are air-gapped once you have the ability to coerce people

if you want a fictional example: watch Colossus: The Forbin Project

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#246

Post by TrainedMonkey »

ianceicys wrote: post #28

Artificial general intelligence (AGI) — in other words, systems that could successfully perform any intellectual task that a human can. Not in my lifetime, not in this millennium. Possibly in the year 2,300. Weird way to blow $20 million.

Do you have a rationale for that? I get a feeling progress in both machine learning and understanding biological intelligence is fairly rapid and has been accelerating. I believe two primary contributing factors are cheaper compute and vast amount of investment poured into machine learning, see https://venturebeat.com/ai/report-ai-investments-see-largest...

Now, the question of whether we are going to have AGI is incredibly broad. So I am going to split it into two smaller ones: - Are we going to have enough compute by year X to implement AGI. Note that we are not talking about super intelligence or singularity here. This AGI might be below human intelligence and incredibly uneconomical to run. - Assuming we have enough compute, will we a way to get AGI working.

The compute advancements scale with new Chip Fabs linearly and tech node improvements exponentially. I think it is reasonable for compute to get cheaper and more accessible throughout at least 2030. I expect this because TSMC is starting 3nm node production, Intel is decoupling fabing and chip design (aka TSMC model), and the strategic investments into into chap manufacturing driven by supply chain disruptions. See https://www.tomshardware.com/news/tsmc-initiates-3nm-chips-p...

How much compute do we need? This is hard to estimate, but amount of human connections in human brain is estimated at 100 trillion, that is 1e14. Current largest model has 530B parameters, that is 5.3e11: https://developer.nvidia.com/blog/using-deepspeed-and-megatr... . That is factor of 500 or 9 doublings off. To get there by 2040 we would need a doubling every 2 years. This is slower that recent progress, but past performance does not predict future results. Still, I believe getting models with 1e14 parameters by 2040 is possible for tech giants. I believe it is likely that a model with 1e14 parameters is sufficient for AGI if we know how structure and train it.

Will we know how to structure and train it? I think is mostly driven by investment into the AI field. More money means more people and given the venture beat link above the investment seems to be accelerating. A lot of that investment will be unprofitable, but we are not looking to make a profit - we are looking for breakthroughs and larger model sizes. Self-driving, stock trading, and voice controls are machine learning applications which are currently deployed in the real world. At the very least it is reasonable to expect continuous investment to improve those applications.

Based on the above I believe we would need to mess things up royally to not get AGI by 2100. Remember this could be below human and super uneconomical AGI. I am rather optimistic, so my personal prediction is that we have 50% chance to get AGI by 2040 and 5-10% chance of getting there by 2030.

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#247

Post by Isinlor »

sailingparrot wrote: post #227
Jack000 wrote: post #116

I'm very optimistic for near-term AGI (10 years or less). Even just a few years ago most in the field would have said that it's an "unknown unknown", we didn't have the theory or the models, there was no path forward and so it was impossible to predict. Now we have a fairly concrete idea of what a potential AGI might look like - an RL agent that uses a large transformer. The issue is that unlike supervised training y…

> we didn't have the theory or the models, there was no path forward and so it was impossible to predict. Now we have a fairly concrete idea of what a potential AGI might look like - an RL agent that uses a large transformer. Who is we exactly? As someone working in AI research I know no one that would agree with this statement, so im quite puzzled by that statement.

Nando de Freitas Research Director at @DeepMind. CIFAR. Previously Prof @UBC & @UniofOxford. made a lot of headlines:

https://twitter.com/NandoDF/status/1525397036325019649

Someone’s opinion article. My opinion: It’s all about scale now! The Game is Over! It’s about making these models bigger, safer, compute efficient, faster at sampling, smarter memory, more modalities, INNOVATIVE DATA, on/offline, … 1/N

Solving these scaling challenges is what will deliver AGI. Research focused on these problems, eg S4 for greater memory, is needed. Philosophy about symbols isn’t. Symbols are tools in the world and big nets have no issue creating them and manipulating them 2/n

https://twitter.com/NandoDF/status/1525397036325019649

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#248

Post by ml_basics »

sailingparrot wrote: post #227
Jack000 wrote: post #116

I'm very optimistic for near-term AGI (10 years or less). Even just a few years ago most in the field would have said that it's an "unknown unknown", we didn't have the theory or the models, there was no path forward and so it was impossible to predict. Now we have a fairly concrete idea of what a potential AGI might look like - an RL agent that uses a large transformer. The issue is that unlike supervised training y…

> we didn't have the theory or the models, there was no path forward and so it was impossible to predict. Now we have a fairly concrete idea of what a potential AGI might look like - an RL agent that uses a large transformer. Who is we exactly? As someone working in AI research I know no one that would agree with this statement, so im quite puzzled by that statement.

I work in AI and would roughly agree with it to first order.

For me the key breakthrough has been seeing how large transformers trained with big datasets have shown incredible performance in completely different data modalities (text, image, and probably soon others too).

This was absolutely not expected by most researchers 5 years ago.

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#249

Post by est31 »

jollybean wrote: post #226
ryanSrich wrote: post #184

Earlier quoted context omitted.

> AGI Idk what prompted you to say this, but is there a version of AGI that isn't "real" AGI? I don't know how anyone could fake it. I think marketing departments might say whatever they want, but I don't see any true engineers falling for something masquerading as AGI. If someone builds a machine that can unequivocally learn on it's own, replicate itself, and eventually solve ever more complex problems that humans c…

This is upside down. First - we already have software that can unequivocally do the things you just highlighted. Learn? Check. Replicate? Trival. But what does that have to do with AGI? Solve Problems Humans Cannot. Check. So we already have 'AGI' and it's a simple computer program. Thinking about 'AGI' as a discrete, autonomous system makes no sense. We will achieve highly intelligent systems with distributed system…

> Replicate? Trival. But what does that have to do with AGI?

If you see it as copying an existing model to another computer, yes it is trivial. But an AGI trying to replicate itself in the real world has to also make those computers.

Making modern computer chips is one of the most non-trivial things that humans do. They require fabs that cost billions, with all sorts of chemicals inside, and extreme requirements on the inside environment. Very hard to build, very easy to disable them via an attack.

Re: John Carmack's new AGI company, Keen Technologies, has raised a $20M round

#250

Post by semi-extrinsic »

HL33tibCe7 wrote: post #182
semi-extrinsic wrote: post #139

Earlier quoted context omitted.

The meme that AGI, if we ever have it, will somehow endanger humanity is just stupid to me. For one, the previous US president is the perfect illustration that intelligence is neither sufficient nor necessary for gaining power in this world. And we do in fact live in a world where the upper echelons of power mostly interact in the decidedly analog spaces of leadership summits, high-end restaurants, golf courses and c…

It sounds like you haven’t really thought through AI safety in any real detail at all. Airgapping and the necessity of human input are absolutely not ways to prevent an AGI gaining access to a system. A true, superintelligent AGI could easily extort (or persuade) those humans. If you think concerns over AGI are “stupid”, you haven’t thought about it enough. It’s a massive display of ignorance. The Computerphile AI sa…

So what if the AGI discovers a bunch of zero days on the internet? We can just turn the entire internet off for a week and be just fine, remember?

And exactly how does the AGI extort or persuade humans? What can it say to me that you can't say to me right now?

Post reply on HN