Grant Sanderson
Creator of 3Blue1B · Math educator · Creator of 3Blue1Brown
aka 3Blue1B, 3Blue1Brown
“there's a spiky frontier to AI. Math is just right there in one of the spikes. But there's kind of a fractal nature to that spikiness, because when you zoom into the specific progress within math, you have some things that are a lot easier than others. So if we just think about imo, which is Old news at this point.”
“I think the dirty secret with the IMO is that you really can train for a lot of them.”
“We need the conjecture generator and then the definition generator. That's the premium tier mathematician. I don't understand how exactly you would make that a benchmark”
“What makes the Galois theory such an interesting example is you have literally this 100 year segment of, like, an idea that, like, flows through many different people's heads before it”
“There is a difference between proof and explanation.”
“I kind of used to think that AIs will become these automated theorem provers, but the role of the mathematicians is going to shift towards my job explain these things. I kind of suspect that actually they'll also be quite good at doing that”
“One interesting take that I've heard about what mathematicians will end up being is actually more analogous to art museum curators than anything else.”
“that's my guess on what most of the useful progress from these models will look like in the next five years is just really filling in that landscape of connections that you can draw.”
Grant Sanderson
One-line summary: Creator of 3Blue1B; tracked here for the June 2026 Dwarkesh conversation on AI and the future of math — IMO-as-trainable-spike, definition/conjecture as the un-benchmarkable premium tier, and the remaining human role as curation.
What they're known for
Brief factual context — fill in.
Why they matter to artificial-intelligence
Why this person's claims are tracked here — fill in.
Said
Speaker-attributed claims extracted from diarized sources. Each bullet mirrors one entry in quotes: frontmatter — keep them in sync.
-
On ai-math-capability-jaggedness, math-verification-loop-to-jagged-ai-progress:
"there's a spiky frontier to AI. Math is just right there in one of the spikes. But there's kind of a fractal nature to that spikiness, because when you zoom into the specific progress within math, you have some things that are a lot easier than others. So if we just think about imo, which is Old news at this point." — 2026-06-30-podcast-dwarkesh-podcast-grant-sanderson-ai-and-the-future-of-math (2026-06-30)
-
On ai-math-capability-jaggedness, agi-definitions-and-benchmark-saturation:
"I think the dirty secret with the IMO is that you really can train for a lot of them." — 2026-06-30-podcast-dwarkesh-podcast-grant-sanderson-ai-and-the-future-of-math (2026-06-30)
-
On ai-math-capability-jaggedness, can-llms-choose-the-right-research-question:
"We need the conjecture generator and then the definition generator. That's the premium tier mathematician. I don't understand how exactly you would make that a benchmark" — 2026-06-30-podcast-dwarkesh-podcast-grant-sanderson-ai-and-the-future-of-math (2026-06-30)
-
On ai-math-capability-jaggedness, automated-ai-research-llm-capability-boundary:
"What makes the Galois theory such an interesting example is you have literally this 100 year segment of, like, an idea that, like, flows through many different people's heads before it" — 2026-06-30-podcast-dwarkesh-podcast-grant-sanderson-ai-and-the-future-of-math (2026-06-30)
-
On ai-math-capability-jaggedness:
"There is a difference between proof and explanation." — 2026-06-30-podcast-dwarkesh-podcast-grant-sanderson-ai-and-the-future-of-math (2026-06-30)
-
On ai-math-capability-jaggedness:
"I kind of used to think that AIs will become these automated theorem provers, but the role of the mathematicians is going to shift towards my job explain these things. I kind of suspect that actually they'll also be quite good at doing that" — 2026-06-30-podcast-dwarkesh-podcast-grant-sanderson-ai-and-the-future-of-math (2026-06-30)
-
On ai-math-capability-jaggedness:
"One interesting take that I've heard about what mathematicians will end up being is actually more analogous to art museum curators than anything else." — 2026-06-30-podcast-dwarkesh-podcast-grant-sanderson-ai-and-the-future-of-math (2026-06-30)
-
On ai-math-capability-jaggedness, math-verification-loop-to-jagged-ai-progress:
"that's my guess on what most of the useful progress from these models will look like in the next five years is just really filling in that landscape of connections that you can draw." — 2026-06-30-podcast-dwarkesh-podcast-grant-sanderson-ai-and-the-future-of-math (2026-06-30)
Sources
Related
Cross-links — fill in.