[00:00]
Artificial intelligence companies are now asking a question that used to belong to science fiction writers and slightly over-caffeinated bloggers: what happens when AI begins to improve AI? Because it is beginning to happen.
[00:16]
I'm always looking for new products that you might find useful. And today's sponsor Consensus is one that you should definitely have a look at because it solved a big problem for me.
[00:28]
My problem with using AI for research is that if you ask a chatbot about science, it'll happily invent a source, present it with total confidence and sound completely convincing,
[00:40]
except the paper doesn't exist. This is why Consensus works the other way around. It searches 220 million peer-reviewed papers first and then uses AI to summarize what it actually found. That
[00:56]
way, every claim links to a real verifiable study. I particularly like the deep search function, which shows you the consensus across the whole literature. You can see where studies agree,
[01:09]
where they conflict, and how the research has changed over time. It also integrates with Claude. Use the link in the description for a full week of Consensus free. And now back to the science news.
[01:23]
That AI improves itself is widely regarded as the beginning of the singularity, aka the intelligence explosion, after which AI will see a rapid self-powered improvement
[01:35]
that will either turn the world into a paradise or hell, depending on who you ask. Jimmy Ba, the cofounder of xAI, recently tweeted that: “Recursive self-improvement loops likely do
[01:48]
live in the next 12 months.” And that “2026 is gonna be insane and likely the busiest and most consequential year for the future of our species.” Dario Amodei, CEO of Anthropic,
[02:02]
one of the world’s leading AI firms, also thinks it’s the biggest thing to watch out for right now, and I agree. [Dario Amodei:] “I think the biggest thing to watch is this issue of AI systems building AI systems. How that goes, whether that goes one way or another,
[02:19]
that will determine, you know, whether it's a few more years until we get there, or if we have wonders and a great emergency in front of us that we have to face.”
[02:33]
Indeed, Anthropic just created an internal think tank as a “fire drill” to prepare for the intelligence explosion and to look for “early warning signals for recursive
[02:45]
self-improvement”. They called it the “Anthropic Institute”. It’s led by co-founder Jack Clark who put the odds above sixty percent that by twenty twenty-eight an AI system could
[02:57]
be told to make a better version of itself, and then do it autonomously. However, the closer we have come to AI self-improvement, the more obvious it’s become that it’s an incredibly vague and useless term. One can argue that if an AI researcher uses AI
[03:15]
in their research, then that counts as a sort of self-improvement of AI. Though seeing how things are going with the scientific literature, I think that’s an intelligence implosion.
[03:27]
But I digress. My point is that there are many ways AI can self-improve in some sense and yet it’s not the sense that anyone had in mind when they were thinking of the intelligence explosion. It was supposed to be a runaway effect, caused by AI improving their own code,
[03:44]
that leads to exponential or super exponential intelligence increase. I already talked last year about the beginning of self-improving AI with Large Language Models that
[03:57]
tune their own hyperparameters or, on the more applied side, AI that has improved microchips used for AI training. Self-improvement, yes, but not the runaway effect we’re waiting for.
[04:11]
But, these examples have multiplied in the past months, and some of these AIs do now indeed write new AI code. The clearest recent example is probably Andrej Karpathy’s
[04:25]
AutoResearch. It is a small Python setup where an artificial intelligence agent modifies language-model training code, trains the model for 5 minutes, checks whether the result improved, keeps or discards the change, and repeats. Meta’s HyperAgents work like this,
[04:42]
too. These are LLM-based research agents that write code, run experiments, debug failures, keep what works and toss what doesn’t. These are not yet models directly rewriting
[04:55]
their own brains. But it is artificial intelligence developing new artificial intelligence, albeit on a small scale. We are indeed getting closer to the self-recursive part.
[05:07]
METR, a nonprofit institute for Model Evaluation and Threat Research is keeping track of this closely. In November they asked whether we have already reached the point where AI can
[05:20]
do AI research well enough to speed up the next AI? To answer this question, they tested GPT-5.1-Codex-Max on some research-engineering tasks, like fixing a damaged small language
[05:34]
model. In November their answer was: we have not yet reached the self-acceleration point. But that was in November, and at the pace that things are going, that's basically the stone age. Already in February, one METR researcher published a model predicting that,
[05:51]
if current trends continue, AI could automate more than 99 percent of AI research and development around 2032. Personally, I think that it will happen much sooner than that.
[06:03]
So this is no longer just “someone on the internet says the singularity is near”. The people who work in the field are now discussing when and how the recursive loop is going to close and how to prepare for that event.
[06:19]
That said, so far, artificial intelligence has not replaced scientists. It has merely automated the part where we try 700 things that don’t work and pretend this was the plan all along.
[06:33]
Thanks for watching. See you tomorrow.