If It Can't Solve Quantum Gravity, It Isn't Real AGI
A physicist draws a high line for AGI: solving quantum gravity is the test. The conversation also clarifies the real relationship between quantum materials and AI — a searched-for material still has to pass the synthesizability gate, and the result may be zero.
The argument · tap a timestamp to hear it
Learning too much paralyzes you; only naivety dares ask the original question
The hardest part of going from student to researcher is not absorbing knowledge but posing the original questions that push the boundary of knowledge forward. Gopakumar describes this transition as ‘naivety plus experience’: learn too much and you become paralyzed; you need to keep a bit of naivety to dare ask questions that seem to have no starting point. The advisor's job is to teach you to break a big problem into small steps you can advance right now without losing sight of the ultimate goal. His own advisor, David Gross, was good at this — by his account, Gross's first student produced work that won the two of them a Nobel Prize together, and Witten also won a Fields Medal. Now he trains students by replicating the same method: he assigns the first one or two problems, for which even he has no answer, and walks with the student slowly until the third or fourth year, then lets them ask questions independently.
— Rajesh GopakumarSpacetime is not a background but an emergent approximation
Think of string theory through the analogy of water being made of discrete atoms: the continuous spacetime we touch is very likely just a low-resolution approximation of some deeper structure at macroscopic scales. Gopakumar stresses this is not pure speculation — quantum mechanics itself does not allow spacetime to be treated as an infinitely divisible smooth background; add black holes (Hawking and others found they have many anomalous properties) and the failure of the Big Bang at the singularity, and physicists must answer ‘where does spacetime come from?’ String theory happens to provide a natural framework: spacetime emerges from something more fundamental, and gravity and the known interactions fit inside it. This explains why he takes ‘solving quantum gravity’ as a very high measure of intelligence: it demands not just doing problems but building a new conceptual framework.
— Rajesh GopakumarMaterials should be designed backwards: set the property first, then find the structure
Start with a long-horizon yardstick: quantum mechanics was originally just for understanding atoms, but the devices it now underpins account for roughly 35%-40% of US GDP by his estimate, and the commercial ripples only became obvious forty or fifty years after the theory appeared. He calls the current phase the second quantum revolution: quantum materials like graphene, discovered by accident, are already commercializing, and the next step is inverse materials design with AI — first define what property the material must have, then search for what structure can deliver it and how to synthesize it, bypassing traditional trial-and-error chemistry. He also reveals a conversation with Demis Hassabis: DeepMind's next holy grail is finding a room-temperature superconductor, and Hassabis told him from the start not to talk about quantum gravity.
— Rajesh GopakumarBeing stuck is the norm; keep three or four problems in play at once
Being stuck is not the exception but the researcher's daily life; especially during the student-to-researcher transition, being stuck can feel like a crush (being overwhelmed). Gopakumar's first strategy is portfolio hedging: push three or four problems at once, and when one dies, switch to another — probabilistically, one will move half a step. The second move is to go teach — the moment a student's eyes light up, that satisfaction can pull you out of a trough, and he has repeatedly found that ‘after teaching, the stuck problem suddenly opens up.’ The third move is to deliberately learn something new you have always wanted to learn but never had time for, to switch to a new frame of reference before returning to the old problem.
— Rajesh GopakumarTalent is multiplicative, not additive; the shortage is never headcount
Why does top science cluster in a few places rather than spread evenly? Gopakumar gives the mechanism: talking in the same place, using each other as sounding boards, keeps the brain stimulated even without formal collaboration; once real collaboration happens, the different skills two people bring are multiplicative rather than additive. He borrows a physics metaphor: photons in an ordinary light bulb are chaotic, while photons in a laser are coherent with one another, and a group of people that achieves coherence will also produce a lasing effect like a laser. He says many Indian institutions actually underperform their talent, and what is missing is not headcount but this resonant atmosphere; hence ICTS's founding principle is simple — hire only first-rate people, give them resources and freedom, and don't let them lower their ambitions.
— Rajesh GopakumarAI can accelerate science but cannot make conceptual leaps
AI has just won two Nobel Prizes in a row, but Gopakumar splits ‘AI for science’ and ‘science for AI’ into two things: the former he acknowledges as an overdrive gear, with AI accelerating a very narrow class of tasks in research; but the ‘theory’ AI does today is still far from his conceptual-layer standard, and the best physics relies on conceptual leaps rather than pattern recognition. Conversely, he is more bullish on science for AI — among the people who discovered scaling laws at Anthropic are physicists by training; Hopfield himself was a statistical physicist who used simple physical models to simulate the brain. Along this path, physicists may join computer scientists and neuroscientists to take AI from language models to physics-inspired world models. This distinction is worth remembering.
— Rajesh GopakumarAfter the synthesizability gate, zero new AI materials remain
The soberest bucket of cold water came from an audience Q&A: Google claims to have discovered a large number of new materials with AI, but after filtering out feasibility and the synthesizability gate, Gopakumar's phrasing is ‘I think left with zero’ — zero remain. His attitude is not to deny AI materials science (someone in the same session cited Microsoft's DFT model with only 385,000 parameters that bypasses traditional hybrid methods), but to remind that the whole chain is still long: after predicting a crystal structure you still have to grow it and scale it, and every step can filter out the surprises. He says India should also bet on this chain. This line is a useful reverse calibration for the ‘AI discovers new materials’ narrative.
— Rajesh GopakumarIn their own words · checked verbatim
I think I'll believe that there is AGI when it solves quantum gravity.
Rajesh Gopakumar0:02
he was asking me what do you think are the big unsolved physics question and he said don't tell me about quantum gravity
Rajesh Gopakumar40:45
a laser beam is much more powerful than your regular light bulb. The laser beam has a coherent set of photons and that's the Bose Einstein effect.
Rajesh Gopakumar55:08
the point was that you you you are not just a cog in the wheel like you are might be in some of the welloiled machines of Stanford or Princeton or Harvard or some places.
Rajesh Gopakumar58:10
as a nation I mean we are a many millennia old civilization we should be thinking in in in terms of the long term investments and that's what I think is going to to to lift the country.
Rajesh Gopakumar1:10:19
I think of it as a little genie on my uh table uh which uh I mean I never used to code.
Rajesh Gopakumar1:26:38
Figures
| Share of US GDP from devices using the laws of quantum mechanics | about 35%-40% | 35:44 |
| Number of naturally occurring elements | 92 | 8:17 |
| Non-US nationals among Princeton's theoretical physics PhD cohort | 10 out of 18 | 13:24 |
| Number of problems Gopakumar advises researchers to pursue in parallel | 3-4 | 50:59 |
Glossary
- Quantum gravity
- The theoretical goal of unifying general relativity and quantum mechanics; no complete theory exists yet, and string theory is a candidate framework.
- String theory
- A framework that assumes elementary particles are different vibrational modes of one-dimensional strings, attempting to unify gravity with the other interactions and explain the emergence of spacetime.
- Bose-Einstein condensation
- A large number of bosons entering the same quantum state at low temperature; in the text it is borrowed to describe the lasing effect when talent is in sync.
- Synthesizability
- Whether a computationally predicted new material can actually be synthesized; it is the gate where AI materials discovery most easily fails.
How to listen
Founders and investors doing AI for Science, and engineers and scientists torn between going to a big platform or doing hardcore research with a startup mindset.
The first 25 minutes of the education story can be sped up; if you want the arguments, start from the middle.