A new clip from an interview with Demis Hassabis, co-founder of DeepMind and a leading figure in artificial intelligence, wades into one of science’s most entertaining barroom arguments: Which legendary physicist ranks highest — Isaac Newton, Albert Einstein, or Richard Feynman?

The clip, shared online with the hashtag #physics and #ai, presents the question as a "fun debate" that eventually pivots to a broader conversation about how AI might reshape the process of scientific discovery. The full interview is linked in the pinned comment, according to the briefing provided by the editorial desk.

The debate that never gets old

Ranking Newton, Einstein, and Feynman is a favorite pastime among physicists and science enthusiasts. Newton laid the foundations of classical mechanics and universal gravitation. Einstein overturned Newtonian notions of absolute space and time with relativity. Feynman reimagined quantum mechanics and made complex ideas accessible to generations of students.

Each represents a different kind of genius. Newton was the solitary systematizer, Einstein the intuitive theorist, and Feynman the iconoclast who thought in diagrams and bongos. Ask a room full of physicists to rank them, and you’ll get a dozen different orderings — each defended with passionate, often contradictory reasoning.

What makes this particular debate interesting is who is leading it. Hassabis is not merely a commentator; he built a company that produced AlphaGo, AlphaFold, and a string of AI systems that have tackled problems from protein folding to game playing. His perspective on scientific methodology carries weight.

Where AI enters the conversation

The clip itself does not reveal Hassabis’s personal ranking of the three scientists, at least not in the material provided. What the briefing makes clear is that the ranking debate serves as a launching point for a larger discussion about AI’s role in future scientific breakthroughs.

That framing is telling. Hassabis has long argued that AI can accelerate the pace of scientific discovery by handling the combinatorial explosion of possibilities that human researchers can’t easily explore. AlphaFold, which predicts protein structures, is the clearest example: it solved a 50-year-old grand challenge in biology by brute-force pattern recognition across millions of sequences.

If the debate asks who was the greatest scientist of the past, the follow-up implicitly asks: Will the next great leap come from a machine?

The limitations of the debate

It’s worth noting that the clip is short — a "quick clip" according to the briefing — and intentionally light. The ranking question is posed as entertainment, not as a rigorous historical evaluation. The real weight of the interview likely lies in the part that follows, which the briefing says is "the full deep dive" with Hassabis.

Still, the mere fact that a prominent AI researcher is using the comparison to talk about future discovery says something about how the field sees itself. DeepMind’s culture has long venerated scientists like Feynman — the company’s research blog often cites his famous phrase about not being able to understand something well enough to explain it. But AI today does not explain its reasoning. It produces answers without a narrative. That tension between human-style insight and machine-generated results lies at the heart of the conversation.

What the debate reveals about scientific progress

Ranking past geniuses is a way of asking what qualities matter most in science. Is it the breadth of foundational work (Newton), the depth of a single revolutionary idea (Einstein), or the ability to reframe a field and communicate it (Feynman)? The answer depends on who you ask and what era you value.

AI complicates that question. A machine that can propose novel hypotheses, design experiments, and interpret results at scale does not have personal genius in the human sense. It has optimization. Yet it may produce discoveries that rival or exceed those of any single human.

That doesn’t make the ranking debate obsolete — it makes it more interesting. By starting with a human-centered comparison, Hassabis can draw a contrast between the kind of brilliance we romanticize and the kind of capability AI might bring.

The broader industry context

SysCall News has covered the rise of AI-driven research tools extensively. From language models that write scientific papers to systems that generate candidate molecules for drugs, the trend is clear: AI is becoming a co-discoverer rather than just a calculator.

Hassabis’s own AlphaFold is a paradigm case. It didn't just solve one problem — it opened up structural biology to computational methods that previously required years of lab work. If the conversation in this clip moves from a three-way human ranking to what AI can do next, it mirrors a shift happening across the scientific community.

What comes next

The full interview with Demis Hassabis likely delves deeper into specific examples: where AI is making progress now, what obstacles remain, and whether the next Newton or Einstein might be a system rather than a person. The pinned comment on the original post provides the link, and for anyone following the intersection of AI and fundamental science, it’s probably worth watching.

For now, the clip does what a good teaser should: it provokes thought without giving away the conclusion. The ranking of Newton, Einstein, and Feynman may never be settled, but the question of what will surpass them is the one that matters most.