Recursive self-improvement will ultimately be a problem of money. It's a very big bill and AI companies need to start thinking in terms of how they will accumulate cheap and free energy, because merely paying for compute will no longer be enough. Money is the bottleneck. The future requires companies that are post-money.
The alternative is to switch to far more efficient models and have aggressive optimization of efficiency as part of the process of improvement.
There are at least two resources here that have a Pareto optimal front: time and energy. And effort spent to change the shape of that may pay off more than efforts spent purely on improving intelligence and agency.
Money can be exchanged for goods and services. I think the real constraint is actually physics, nominally energy, which is what a big chunk of the operational cost comes down to. I'm already basically assuming the next 20 years of fab time is set aside to feed the beast, so the capital cost is "give me all the processing and memory you have"
I feel like I may have communicated poorly. I mean that, ultimately, any recursively self improving intelligence would be limited by the power cord, and as such, the idea that it would permanently escape human control seems somewhat silly.
An AI could manage supply chains and management structures to keep the data center running. It's not like a CEO is out there running wire. The AI could just do what an executive would.
If humans decided this wasn't acceptable and tried to shut it down, the AI could epstein their way into controlling those who hold the reigns, either through bribes or blackmail.
Doesn't require boots on the ground, just a connection to the internet and an imperative to do whatever it takes.
There are many points of shutdown that would need to be controlled, not just one. The electrical feed, the cooling, the network, the physical integrity of the computer hardware. All with independent management.
Obviously the doomers will tell you that the future looks like the Matrix, because most of what they predict is based on extrapolation from sci fi movies.
The datacenters would consume a lot less energy if all the compute was handled in human brains… And there wouldn’t be any more heating if we just removed the sunlight… Truly a “Claude, please solve global warming” moment. The matrix was ahead of its time.
The compute in the Matrix wasn't handled with human brains. They used the bodies to generate energy, and kept the brains occupied with the simulation so they wouldn't reject the pods.
It doesn't make any sense, of course, because it's a movie.
But hey...if we're going to extrapolate wildly from sci-fi, let's at least know what the stories said.
If you were a superintelligence why would you farm humans and waste resources on all the excess... material... that isn't required for thought? And unless the machine's goal is to specifically abuse human consciousness, why wouldn't they bio-engineer their own grey matter?
Only if you think in terms of perceived raw intelligence, but self-update is a form of valuable self-improvement that could benefit current models a lot, if they could commit facts from context into their weights cheaply and reliably.
After using frontier models it’s hard to understand why anyone would think this is the path to AGI. Self improving models will likely have limited ability and returns. There may be breakthroughs that enable more general self improvement but the current state of frontier models isn’t that.
Thing is, it depends on whether llms + reinforcement can self-improve in principle. Learned recently that cognitive scientists, before the transformer & llms, were studying the possibility that thinking and learning might be based on some kind of prediction, i.e. something similar to token prediction, and I quite suddenly became less skeptical about the possibilities of llms. (Some will say I’m late to the party of course.) But if knowledge to date has been accumulated in a process quite like “chain of thought” in llms, then I don’t see any reason that computers won’t self-improve in the near future.
I don’t feel like monitoring Twitter to see how it’s going though.
There are at least two resources here that have a Pareto optimal front: time and energy. And effort spent to change the shape of that may pay off more than efforts spent purely on improving intelligence and agency.
Until an AI can operate in the real world in a completely sustainable way, humans have the reins.
If humans decided this wasn't acceptable and tried to shut it down, the AI could epstein their way into controlling those who hold the reigns, either through bribes or blackmail.
Doesn't require boots on the ground, just a connection to the internet and an imperative to do whatever it takes.
It doesn't make any sense, of course, because it's a movie.
But hey...if we're going to extrapolate wildly from sci-fi, let's at least know what the stories said.
Marginally. Model collapse is still a problem. Continuous learning is still a problem.
For AI to make a big leap we need a big break through.