←
Bernardino Romera-Paredes1,231 posts
Bernardino Romera-Paredes avatar

Bernardino Romera-Paredes

@ber24

Previously core team member of AlphaFold2, AlphaTensor, and FunSearch at Google DeepMind. Now starting something new

◷ Joined July 2009▧ 78 Photos and videos
Bernardino Romera-Paredes avatar
Bernardino Romera-Paredes@ber24·Nov 11

10 years ago, when I read in the Telegraph "Microsoft will 'solve' cancer within 10 years", I thought saying that was irresponsible. Now when Dario Amodei says cancer could be "relics of the past" within a decade, I'm no longer skeptical. But my skepticism has moved to dual use

Dario Amodei says cancer, Alzheimer’s, heart disease, and mental illness will be relics of the past in less than a decade, and that "we can create an age of miracles" “You know, if today we look at things like cancer, Alzheimer’s, heart disease, mental illness, these will be relics of the past in, I think, less than a decade." "And I believe that the company which gets the risks right will also be the company that has the license to operate that society trusts and society believes can build the benefits in the right way." "I think if we do this right, we can create an age of miracles where humanity is able to solve problems that have eluded us for millennia." "We have to make sure that all the risks, all the ways that things can go wrong with this technology, we very carefully address those risks as we bring the benefits, and as we have a viable business at the same time.”

0:55
Stanislav Nikolov avatar
⟳ Bernardino Romera-Paredes retweeted
Stanislav Nikolov@snikolov·Oct 7

I get the feeling we'll be discovering many more unruly truths... perhaps, in due time, with a bird's eye view of these new territories we will start to understand and appreciate a higher order harmony

Are there alien types of math that are more akin to biology than the beautiful math that we tend to study? Math that is unruly and inscrutable but has layers of depth and structure and is ultimately useful to understand and try to tame?

Bernardino Romera-Paredes avatar
Bernardino Romera-Paredes@ber24·Oct 5

A week of essays on AI and the future of maths (Tao, Buzzard, Gowers, Christian), and I keep thinking of a one-page Ted Chiang story in Nature from 2000. Superhuman minds make the discoveries. Humans are left with hermeneutics: interpreting them

Post attachment

Reading all those recent beautiful essays by Terry Tao, Kevin Buzzard and @wtgowers, I could not resist, but write up my deeply personal perspective on how I perceive and see things: docs.google.com/document/d/e…

Bernardino Romera-Paredes avatar
Bernardino Romera-Paredes@ber24·Sep 9

Agentes de IA han resuelto el problema de Navier-Stokes, uno de los problemas más importantes de las matemáticas. Y ha sido gracias a los avances de Diego Córdoba y Luis Martínez-Zoroa. Ni una noticia en @elpais @elmundoes y demás. A ver si al menos @LaRevuelta_TVE los invita

Post attachment
Andrés Miguel Torrubia Sáez avatar
⟳ Bernardino Romera-Paredes retweeted
Andrés Miguel Torrubia Sáez@antor·Jun 10

Thankfully the 🇪🇺 AI act prevents Anthropic behavior with regards to sabotage, concentration of power, etc. Wait no 🤦‍♂️

Quantinuum avatar
⟳ Bernardino Romera-Paredes retweeted
Quantinuum@QuantinuumQC·11 Dec 2025

What if every quantum researcher had an army of students to help write quantum algorithms? LLMs are starting to serve as such a resource. We’ve partnered with @HivergeAI to use AI for quantum algorithm discovery, exploring practical quantum chemistry. quantinuum.com/blog/automate…

Adam Zsolt Wagner avatar
⟳ Bernardino Romera-Paredes retweeted
Adam Zsolt Wagner@azwagner_·6 Nov 2025

Really happy to share our new paper on using AlphaEvolve for mathematical exploration at scale, written with Javier Gómez-Serrano, Terence Tao, and @GoogleDeepMind's Bogdan Georgiev. We tested it on 67 problems and documented all our successes and failures. 🧵

Post attachment
Alhussein Fawzi avatar
⟳ Bernardino Romera-Paredes retweeted
Alhussein Fawzi@AlhusseinFawzi·15 Oct 2025

We challenged our intern @ramadan_al76760 (zero prior AI experience) to beat the CIFAR-10 training speed record using @hivergeai's algorithmic discovery engine. Result: Sub-2-second (!!) training for the first time ever.

New CIFAR-10 training speed record: 94% in 1.99 seconds on one A100 Previous record: 2.59 seconds (Nov. 10th 2024) New record-holder: Algorithmic discovery engine developed by @hivergeai Changelog: - Muon: Vectorize NS iter and reduce frequency of 'normalize weights' step 1/3

Post attachment
Bernardino Romera-Paredes avatar
Bernardino Romera-Paredes@ber24·11 Oct 2025

Falsest friend conjecture: "Inhabitable" is the only word (or else the longest one) such that: - it has the exact same spelling in two different languages (English & Spanish) - it means the exact opposite thing in each

Post attachmentPost attachment