By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.

Sequoia Capital

David Silver and Lasse Espeholt of Ineffable Intelligence: A Superlearner for the Age of Experience

David Silver spent two years building AlphaGo on human data. When he removed human data altogether building AlphaGo Zero, the AI surpassed AlphaGo in under two days. He left DeepMind to run that same bet in pursuit of superintelligence. David and Lasse Espeholt, co-founders of Ineffable Intelligence, explain their plan to build a “superlearner”—an agent that starts knowing nothing and discovers science, mathematics, and language for itself by direct experience instead of through a pre-training run. Silver argues that human-language reasoning is a ceiling, not a foundation, that alignment via chain-of-thought monitoring is already reaching its limits, and that the science of how superintelligences coexist should be worked out in simulation before anything touches the real world. Espeholt describes the GPU cluster and the lineage from IMPALA and SEED RL that makes the experiment possible. Silver’s forecast: strong evidence within a year, superintelligence in three. Hosted by Sonya Huang, Sequoia Capital 0:00 – Introduction 1:09 – What held true from Go to today, and what the LLM era changed 3:35 – Why leave DeepMind to start a new lab 5:58 – Why "Ineffable" (and the secret second reason) 7:03 – Lasse's path: scaling RL at DeepMind and Google Brain 9:31 – Defining a super learner and first contact 12:35 – Super learners vs. LLMs that make discoveries 13:59 – Day to day: the RL stack, crucibles and human-journey milestones 16:13 – Don't breadcrumb the objective: simple rewards, emergent complexity 21:16 – What a crucible looks like 24:14 – Go: one agent is boring, two is infinite 25:41 – Will agents rediscover language? 27:22 – Removing human priors: AlphaGo to AlphaZero 32:05 – Keeping human bias out of the crucibles 33:35 – Confidence, timing and super forecasting 36:24 – Is compute the driver? 38:13 – The ceiling on simulation and Rich Sutton's objection 43:44 – Alignment and the science of superintelligence 47:45 – Monitoring agents that don't think in English 51:01 – Pacing the frontier 52:54 – The compute setup: Vera Rubin on GCP 55:18 – Training, inference, simulation and RL scaling laws 57:40 – IMPALA and SEED RL 1:00:58 – 20 people and the compute-per-person ratio 1:04:49 – A research-led culture 1:05:47 – Predictions

David SilverguestSonya HuanghostLasse Espeholtguest

Oct 13, 20261h 8mWatch on YouTube ↗

Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.

WHAT IT’S REALLY ABOUT

Ineffable’s bet: superintelligence via reinforcement learning from experience

David Silver says reaching superintelligence requires AI systems that learn from experience and autonomously discover knowledge, rather than relying on human language data as the core substrate.
Ineffable Intelligence is building a “super learner” defined by depth (starting from nothing and surpassing human knowledge) and breadth (learning across simulated, digital, and physical streams of experience).
The company is creating “crucibles,” simulated environments designed for emergent complexity, where simple objectives can produce surprising, open-ended discoveries instead of being guided by handcrafted task sequences.
Silver argues that alignment approaches centered on language monitoring and human-data priors will become insufficient as agents grow more capable, and that simulation can enable a science of superintelligence to inform safer real-world deployment.
The team emphasizes compute-intensive, GPU-centered reinforcement learning infrastructure and expects early evidence within a year, with Silver forecasting first contact with superintelligence in roughly three years.

IDEAS WORTH REMEMBERING

5 ideas

The core bet is that superintelligence requires learning-from-experience systems that discover knowledge for themselves.

He argues that humans collectively started "knowing nothing" and discovered languages, technology, science, and math through experience-driven discovery, and that AI needs the same capability to continually learn and discover new knowledge indefinitely.

Ineffable was created to pursue a non-LLM path to superintelligence with full focus.

He says Ineffable’s approach "doesn't actually go through LLM-based products as an intermediate point" and deserves a lab fully focused on this path rather than being a small part of an organization centered on LLM products.

A “super learner” is defined by both depth (from zero to beyond-human knowledge) and breadth (learning across many environments).

He defines depth as learning from scratch to beyond-human discoveries, and breadth as being able to drop the system into any stream of experience (simulated, digital, or physical) and have it keep learning.

They aim for emergence: simple objectives + complex worlds, rather than tightly specified task rewards.

He contrasts their approach with breadcrumbed objectives, saying they will give simple objectives in complex environments so surprising capabilities emerge—analogizing to evolution optimizing simple fitness signals that produce complex life.

Crucibles are built to support open-ended discovery through emergent complexity, not fixed task suites.

They describe “crucibles” as simulated environments designed for emergent complexity (e.g., via many interacting objects with physics engines or many interacting agents), not a handcrafted set of puzzles or benchmark tasks.

WORDS WORTH SAVING

5 quotes

Collectively, you know, we started with this kind of gray mush in our head. Um, we somehow started out knowing nothing at all, and somehow discovered, like, everything that humans know today was, like, this process of collective discovery, like of, of languages and technology and science and mathematics and, uh, all the rest of it.

— David Silver

And so to me, that is the essence of what it will take to get to superintelligence, like a system that can just keep learning and discovering knowledge forever.

— David Silver

And in the end, I, I realized that there's this other lamppost that right now no one is looking under, and where if, if I actually put my effort there, has a much higher chance of really having a, a, a massive impact on the field. 'Cause if this works, it's gonna change the whole of AI.

— David Silver

When we got AlphaZero right, you know, we gave up on the human data and finally focused on what would it be like to build an algorithm that could discover for itself. Um, so once we got it right, that algorithm, it took just a few hours.

— David Silver

I think we will achieve something which is a true superintelligence, like the, according to the definition I gave earlier of something which, you know, discovers things which are far beyond what humans can achieve in, across y- y- you know, everything from science, mathematics, technology, languages, et cetera, et cetera.

— David Silver

Learning from experience vs human-data trainingDefinition of a “super learner” (depth and breadth)Crucibles and emergent complexity in simulationSimple objectives vs reward shaping and breadcrumbingRemoving human priors (AlphaGo to AlphaZero analogy)Compute, GPUs, and scaling reinforcement learning systemsAlignment research in simulation and gating real-world deployment

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

Read along on your phone, search it beside the video, or ask your favorite AI app what was said.