TB Tom Brown Tom Brown 32 year old Londoner; a lover of a knitted polo, an Old Fashioned and, of course, Formula 1.

“We believe that this can accelerate even further as we approach recursive self-improvement. Claude will be able to discover compute multipliers entirely by itself, allowing the number of compute multipliers to grow faster and faster.”

AI Explained Official Podcast · AI Research & Frontier Labs · November 2026

“We believe that this can accelerate even further as we approach recursive self-improvement. Claude will be able to discover compute multipliers entirely by itself, allowing the number of compute multipliers to grow faster and faster.” — Tom Brown, AI Explained Official Podcast

Context

A clip from Anthropic's IPO roadshow video, introduced as co-founder Tom Brown. The host had just defined compute multipliers: discoveries that let you train the same model with a fraction of the compute.

Transcript

AI Explained Official Podcast Around 02:47 into the episode
Speaker 1

…AI research itself. Indeed, Anthropic co-founders boast about how good their models are at recursive self-improvement. In time for their IPO debut today, I watched the 40-minute IPO promo video and look at what Tom Brown, one of the co-founders of Anthropic, says about self-improvement.

Speaker 2

…you can use 25 units of compute to train the exact same model, or you can use those 100 physical units to train an even more intelligent model that would have previously required 400 units. Compute multipliers create a competitive advantage for us that's hard to replicate. They build on each other. Each new multiplier we discover has a positive effect on the multipliers before it. They stack. We believe that this can accelerate even further as we approach recursive self-improvement. Claude will be able to discover compute multipliers entirely by itself, allowing the number of compute multipliers to grow faster and faster.

Speaker 1

Which brings me to the third part of this video, the race to the top that Dario Amadei promised. The idea is somewhat simple. Some in Anthropic call it, you can't do safety from second place. But the idea is that Anthropic rush to be in the lead, rush to create an automated AI researcher. Then they…

Speaker 3

So the way that we've observed is the AI doing what we want it to do is through chain of thought monitoring. And when we first developed that technology, we said we have to make sure this stays faithful. But as the models have got more capable, they can do more kind of within just the forward path…

Speaker 1

Dare I say it, but for me, OpenAI are starting to seem more open in their communications. Far from a race to the top led by Anthropic, OpenAI's report yesterday says this. They admit that to maintain human oversight of systems, we will need new breakthroughs. Basically, on our current trajectory…

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

More from AI Explained Official Podcast