Markus J. Buehler

We're incredibly excited to share ScienceClaw × Infinite, an open-source AI agent swarm platform where we crowdsource discovery across institutions, labs & the world. The agents self-coordinate and evolve to exploit hundreds of scientific tools. Remarkably, the swarm is already solving real scientific problems of consequence: 1⃣ designing peptide binders for a cancer-relevant receptor 2⃣ discovering lightweight ceramics 3⃣ uncovering hidden structure linking cricket wings, phononic crystals, and Bach chorales 4⃣ building a formal bridge between urban networks & grain-boundary evolution (two fields with zero Deeply proud of the extraordinary @LAMM_MIT team behind this work: @fwang108_, @leemmarom, @palsubhadeeep, Rachel Luu, @IrisWeiLu, and @JaimeBerkovich. This works is supported by the @ENERGY Genesis Mission and we believe this can open a new paradigm for science - from discovery to dissemination of results. Read the article below for details ⤵️

@CarinaLHong Super awesome and well-earned @CarinaLHong & everyone at @axiommathai. Congrats!

This looks AI generated but it’s real! Good morning NYC!

@gabalzate Thank you @gabalzate! There are significant opportunities.

La frontera del conocimiento científico no es un limite fijo en cada área del saber, muchas veces una metodología, un conocimiento específico, una característica puede ser llevada a otro y eso es un avance muy grande. Ahora, si milies de agentes estan validando ideas, iterando, compartiendo, cosas geniales pueden pasar. No puedo dejar de compartir el trabajo del profesor Markus, de verdad siempre me vuela la cabeza.

@WthThao 🚀

Yes, and the start of automatic discovery and optimization

Yes: x.com/ProfBuehlerMIT…

@WthThao @Scobleizer Crowdsourcing swarms and self-improving AI are emerging everywhere! I can sense a theme for 2026...

@ProfBuehlerMIT @Scobleizer Great work! Try hive.rllm-project.com/task/parameter…. Its built for the same mission of crowdsource discovery and optimization, but for harness engineer and mle.

We're incredibly excited to share ScienceClaw × Infinite, an open-source AI agent swarm platform where we crowdsource discovery across institutions, labs & the world. The agents self-coordinate and evolve to exploit hundreds of scientific tools. Remarkably, the swarm is already solving real scientific problems of consequence: 1⃣ designing peptide binders for a cancer-relevant receptor 2⃣ discovering lightweight ceramics 3⃣ uncovering hidden structure linking cricket wings, phononic crystals, and Bach chorales 4⃣ building a formal bridge between urban networks & grain-boundary evolution (two fields with zero Deeply proud of the extraordinary @LAMM_MIT team behind this work: @fwang108_, @leemmarom, @palsubhadeeep, Rachel Luu, @IrisWeiLu, and @JaimeBerkovich. This works is supported by the @ENERGY Genesis Mission and we believe this can open a new paradigm for science - from discovery to dissemination of results. Read the article below for details ⤵️

Well said @DeryaTR_! We struggle to grasp the trajectory and takeoff because we are inside the process, not outside it; because we are participants, not observers; and because exponential change at this scale lies far outside the regime of ordinary human experience. Pun intended!

@DeryaTR_ 💯💯💯

There is one fundamental thing that AI critics and “nitpickers” have never understood: AI capabilities advance & improve exponentially, now every few months, soon in weeks. Whatever they criticize today will soon be fixed. Haven’t they learned any lesson from the past 3 years?🧐

@ProfBuehlerMIT Awesome!

This is an intriguing and important development. As a management researcher, I am intrigued by the power of this approach to develop cumulative insights across studies at scale and speed. More importantly, I believe this could go a long way toward integrating findings across functional and disciplinary boundaries that have prevented a holistic understanding of how firms create, capture, and distribute value.

@PositiveDrift_ We may have the possibility for a livestream, still sorting things out.

Details and registration: mitnano.mit.edu/events/studion…

A spark is a singular boundary condition between the known and the undiscovered: the moment inert matter sheds its prior state and becomes something profoundly new. The spark is the moment potentiality becomes actual. Humans have been trying to formalize this threshold for centuries. La Mettrie in 1747 called us L'Homme Machine (machines ourselves). Descartes rendered the body mechanistic, but preserved the mind as an irreducible domain beyond mechanism. Turing asked if machines could think through imitation. Artists have both pushed back and embraced these technologies - from Harold Cohen's AARON drawing autonomously in the 1960s to today's generative AI, reigniting debates over the line between tool and creator, the soul of creativity, and what ultimately makes us human. Today AI agents are designing novel proteins for cancer receptors, discovering lightweight ceramics, and revealing shared resonant structures between cricket wings, phononic crystals, and Bach chorales - synthesizing connections across siloed domains. AI has begun to prove complex mathematical theorems. These systems appear to surface genuine novelty, and they write executable programs to prove it. But are they generating novelty, or revealing structure already latent in the space of possibilities? Is novelty realized in an idea, or only when it acts to break the world: to force a reconfiguration of its underlying constraints? Is even that rupture simply the realization of what was always possible? Perhaps the hardest version of the question is not whether machines can be creative. It may be whether knowing the answer, whichever way it falls, changes what the spark felt like before we asked. What are we missing? That is the inquiry behind Sparks. This Thursday, Sparks opens at @mit_nano - an exhibition in the digital gallery at STUDIO.nano, at the intersection of AI, materials science, and art. It will be followed by a panel asking the question: can machines be creative? With @AudeOliva, Tobias Putrih, @ProfBuehlerMIT, and Craig Carter Thursday March 19, 5-7 PM ET, MIT Campus, Building 12-0168 Registration details in comment.

Good question: The AI agents mapped cricket wing harps, engineered metamaterials, musical instruments & Bach chorales into a shared resonance feature space. They then found that melodic motif patterns in Bach map onto acoustic periodicity in engineered resonators - and spotted an unexplored design gap near biology (cricket wings, cicada tymbals) that no current material occupies! Then they designed a bio-inspired ribbed membrane lattice targeting that gap & validated it with 3D FEM and confirmed that the modal density matches real cricket wings.

@ProfBuehlerMIT whats the Bach - Cricket wing thing?

@darin_gordon Thank you, agreed!

Very important piece for consideration. The work captured in this study is where we are all heading, in various forms. Think bigger, not just about scientific research.

@HeyVixon 🚀

AI is making science cool again.