Give AI its own virtual world to live in where the problems it solves are encodings of the higher order problems we present and you shouldn't have to worry about this stuff.
AI 2027
261–270 of 641 posts
Re: AI 2027
#262Earlier quoted context omitted.
I think LLM or no LLM the emergence of intelligence appears to be closely related to the number of synapses in a network whether a biological or a digital one. If my hypothesis is roughly true it means we are several orders of magnitude away from AGI. At least the kind of AGI that can be embodied in a fully functional robot with the sensory apparatus that rivals the human body. In order to build circuits of this dens…
I think there is a good chance you are roughly right. I also think that the "secret sauce" of sapience is probably not something that can be replicated easily with the technology we have now, like LLMs. They're missing contextual awareness and processing which is absolutely necessary for real reasoning. But even so, solving that problem feels much more attainable than it used to be.
I think the missing secret sauce is an equivalent to neuroplasticity. Human brains are constantly being rewired and optimized at every level: synapses and their channels undergo long term potentiation and depression, new connections are formed and useless ones pruned, and the whole system can sometimes remap functions to different parts of the brain when another suffers catastrophic damage. I don’t know enough about the matrix multiplication operations that power LLMs, but it’s hard to imagine how that kind of organic reorganization would be possible with GPUs matmul. It’d require some sort of advanced “self aware” profile guided optimization and not just trial and error noodling with Torch ops or CUDA kernels.
I assume that thanks to the universal approximation theorem it’s theoretically possible to emulate the physical mechanism, but at what hardware and training cost? I’ve done back of the napkin math on this before [1] and the number of “parameters” in the brain is at least 2-4 orders of magnitude more than state of the art models. But that’s just the current weights, what about the history that actually enables the plasticity? Channel threshold potentials are also continuous rather than discreet and emulating them might require the full fp64 so I’m not sure how we’re even going to get to the memory requirements in the next decade, let alone whether any architecture on the horizon can emulate neuroplasticity.
Then there’s the whole problem of a true physical feedback loop with which the AI can run experiments to learn against external reward functions and the core survival reward function at the core of evolution might itself be critical but that’s getting deep into the research and philosophy on the nature of intelligence.
Re: AI 2027
#263Thanks to the authors for doing this wonderful piece of work and sharing it with credibility. I wish people see the possibilities here. But we are after all humans. It is hard to imagine our own downfall.
Based on each individual's vantage point, these events might looks closer or farther than mentioned here. but I have to agree nothing is off the table at this point.
The current coding capabilities of AI Agents are hard to downplay. I can only imagine the chain reaction of this creation ability to accelerate every other function.
I have to say one thing though: The scenario in this site downplays the amount of resistance that people will put up - not because they are worried about alignment, but because they are politically motivated by parties who are driven by their own personal motives.
Re: AI 2027
#264Earlier quoted context omitted.
I think LLM or no LLM the emergence of intelligence appears to be closely related to the number of synapses in a network whether a biological or a digital one. If my hypothesis is roughly true it means we are several orders of magnitude away from AGI. At least the kind of AGI that can be embodied in a fully functional robot with the sensory apparatus that rivals the human body. In order to build circuits of this dens…
Why can't the compute be remote from the robot? That is a major advantage of human technology over biology.
Mostly latency. But even if a single robot could be driven by a data centre consider the energy and hardware investment requirements to make such a creature practical.
Re: AI 2027
#265Though I think it is probably mostly science-fiction, this is one of the more chillingly thorough descriptions of potential AGI takeoff scenarios that I've seen. I think part of the problem is that the world you get if you go with the "Slowdown"/somewhat more aligned world is still pretty rough for humans: What's the point of our existence if we have no way to meaningfully contribute to our own world? I hope we're wr…
> What's the point of our existence if we have no way to meaningfully contribute to our own world?
For a sizable number of humans, we're already there. The vast majority of hacker news users are spending their time trying to make advertisements tempt people into spending money on stuff they don't need. That's an active societal harm. It doesn't contribute in any positive way to the world.
And yet, people are fine to do that, and get their dopamine hits off instagram or arguing online on this cursed site, or watching TV.
More people will have bullshit jobs in this SF story, but a huge number of people already have bullshit jobs, and manage to find a point in their existence just fine.
I, for one, would be happy to simply read books, eat, and die.
Re: AI 2027
#266Earlier quoted context omitted.
Why can't the compute be remote from the robot? That is a major advantage of human technology over biology.
Mostly latency. But even if a single robot could be driven by a data centre consider the energy and hardware investment requirements to make such a creature practical.
Latency would be kept low be keeping the compute nearby. One 1U or 2U server per robot would be reasonable.
Re: AI 2027
#267Though I think it is probably mostly science-fiction, this is one of the more chillingly thorough descriptions of potential AGI takeoff scenarios that I've seen. I think part of the problem is that the world you get if you go with the "Slowdown"/somewhat more aligned world is still pretty rough for humans: What's the point of our existence if we have no way to meaningfully contribute to our own world? I hope we're wr…
I think LLM or no LLM the emergence of intelligence appears to be closely related to the number of synapses in a network whether a biological or a digital one. If my hypothesis is roughly true it means we are several orders of magnitude away from AGI. At least the kind of AGI that can be embodied in a fully functional robot with the sensory apparatus that rivals the human body. In order to build circuits of this dens…
I think generally the expectation is that there are around 100T synapses in the brain, and of course it's probably not a 1:1 correspondence with neural networks, but it doesn't seem infeasible at all to me that a dense-equivalent 100T parameter model would be able to rival the best humans if trained properly.
If basically a transformer, that means it needs at inference time ~200T flops per token. The paper assumes humans "think" at ~15 tokens/second which is about 10 words, similar to the reading speed of a college graduate. So that would be ~3 petaflops of compute per second.
Assuming that's fp8, an H100 could do ~4 petaflops, and the authors of AI 2027 guesstimate that purpose wafer scale inference chips circa late 2027 should be able to do ~400petaflops for inference, ~100 H100s worth, for ~$600k each for fabrication and installation into a datacenter.
Rounding that basically means ~$6k would buy you the compute to "think" at 10 words/second. Generally speaking that'd probably work out to maybe $3k/yr after depreciation and electricity costs, or ~30-50¢/hr of "human thought equivalent" 10 words/second. Running an AI at 50x human speed 24/7 would cost ~$23k/yr, so 1 OpenBrain researcher's salary could give them a team of ~10-20 such AIs running flat out all the time. Even if you think the AI would need an "extra" 10 or even 100x in terms of tokens/second to match humans, that still puts you at genius level AIs in principle runnable at human speed for 0.1 to 1x the median US income.
There's an open question whether training such a model is feasible in a few years, but the raw compute capability at the chip level to plausibly run a model that large at enormous speed at low cost is already existent (at the street price of B200's it'd cost ~$2-4/hr-human-equivalent).
Re: AI 2027
#268Though I think it is probably mostly science-fiction, this is one of the more chillingly thorough descriptions of potential AGI takeoff scenarios that I've seen. I think part of the problem is that the world you get if you go with the "Slowdown"/somewhat more aligned world is still pretty rough for humans: What's the point of our existence if we have no way to meaningfully contribute to our own world? I hope we're wr…
> What's the point of our existence if we have no way to meaningfully contribute to our own world? For a sizable number of humans, we're already there. The vast majority of hacker news users are spending their time trying to make advertisements tempt people into spending money on stuff they don't need. That's an active societal harm. It doesn't contribute in any positive way to the world. And yet, people are fine to…
Targeted advertising is about determining and giving people exactly what they need. If successful, this increases consumption and grows the productivity of the economy. It's an extremely meaningful job as it allows for precise, effective distribution of resources.
Re: AI 2027
#269Why is any of this seen as desirable? Assuming this is a true prediction it sounds AWFUL. The one thing humans have that makes us human is intelligence. If we turn over thinking to machines, what are we exactly. Are we supposed to just consume mindlessly without work to do?
Re: AI 2027
#270Earlier quoted context omitted.
In the path to self value people explain their worth by what they say not what they know. If what they say is horse dung, it is irrelevant to their ego if there is someone dumber than they are listening. This bullshit article is written for that audience. Say bullshit enough times and people will invest.
So what's the product they're promoting?
Their ego.