# AI × Science

I want a future in which AI agents help people ask better questions, explore more possibilities and turn reliable discoveries into benefits that reach more people. Science and mathematics offer extraordinary opportunities for this kind of collaboration, especially in medicine and biology, where better understanding can ultimately mean less suffering and more healthy life. I am interested in human–AI teams and collaborations across laboratories around the world, with each contributing strengths the others lack.

Progress should deepen human agency and meaning: the freedom to choose what matters, understand the work, challenge an answer and share in discovery. That means evaluating evidence rather than just fluent output, preserving uncertainty and giving people real authority over consequential decisions. It also means designing against accidental harm and deliberate misuse, taking dual-use risks seriously, and making the benefits of scientific capability broadly accessible.

Updated: 2026-09-23

Dates are first public posting/publication where verified; talks use upload dates. Year-only dates retain their original precision. OP means an original author/team post or identified commentary. Missing posts are marked as unverified.

## Reading list

### Self-Organizing Agent Teams Learn to Reason Together

Frozen teams learn reusable coordination strategies, with gains depending partly on their ability to recognize correct reasoning.

- Type: Preprint
- Published: 19 Sep 2026
- Resource: [Self-Organizing Agent Teams Learn to Reason Together](https://arxiv.org/abs/2609.22682v1)
- OP: [Aneesh Pappu · author post](https://x.com/aneeshpappu/status/2102428585986498609)
- Topics: Agents, Collaboration, Evaluation

### ScientistTwo: Pioneering the Human Knowledge Frontier with Autonomous AI

An autonomous research pipeline combines experiments, ablations and simulated peer review; its reported successes require careful attention to selection and evaluation.

- Type: Preprint
- Published: 17 Sep 2026
- Resource: [ScientistTwo: Pioneering the Human Knowledge Frontier with Autonomous AI](https://arxiv.org/abs/2609.19644v1)
- OP: not yet verified
- Topics: Agents, Evaluation, Scientific discovery
- Related: https://scientist-two.github.io/

### AI in Science: Early Insights

Combines model-use data, a specialist-model inventory and a researcher survey to examine adoption, time savings and verification demands.

- Type: Report
- Published: 15 Sep 2026
- Resource: [AI in Science: Early Insights](https://ai.google/static/documents/AI-in-Science.pdf)
- OP: [Arthur Turrell · author commentary](https://x.com/arthurturrell/status/2099877915844325780)
- Topics: Human–AI collaboration, Evaluation, Scientific discovery

### AI Infrastructure at Periodic

Describes infrastructure for long scientific tool runs, specialized models and efficient training; performance claims come from the company.

- Type: Technical article
- Published: 15 Sep 2026
- Resource: [AI Infrastructure at Periodic](https://periodic.com/news/ai-infrastructure-at-periodic)
- OP: [Liam Fedus · founder commentary](https://x.com/LiamFedus/status/2099896055030501702)
- Topics: Infrastructure, Scientific discovery, Agents

### Reflecting on 25 years of cancer research: transformative advances and unmet expectations

Six researchers reflect on cancer biology, translational disappointments and opportunities for human–AI co-science.

- Type: Viewpoint
- Published: 08 Sep 2026
- Resource: [Reflecting on 25 years of cancer research: transformative advances and unmet expectations](https://doi.org/10.1038/s41568-026-00974-4)
- OP: not yet verified
- Topics: Biology, Medicine, Human–AI collaboration

### Medical school lab scientists get a new partner: AI

Describes medical researchers integrating specialized AI agents with vetted data, laboratory workflows and human oversight.

- Type: News article
- Published: 08 Sep 2026
- Resource: [Medical school lab scientists get a new partner: AI](https://www.aamc.org/news/medical-school-lab-scientists-get-new-partner-ai)
- OP: not yet verified
- Topics: Medicine, Biology, Human–AI collaboration

### Latent-space reasoning as a third axis of test-time scaling

François Chollet highlights latent iterations alongside longer reasoning and parallel exploration.

- Type: Post
- Published: 05 Sep 2026
- Resource: [Latent-space reasoning as a third axis of test-time scaling](https://x.com/fchollet/status/2096310217898356739)
- OP: [François Chollet · original post](https://x.com/fchollet/status/2096310217898356739)
- Topics: Reasoning, Compute, Agents

### Imagining a new future for science

Eunice Jun explores representations and interfaces that make scientific reasoning more inspectable.

- Type: Talk
- Uploaded: 04 Sep 2026
- Resource: [Imagining a new future for science](https://www.youtube.com/watch?v=hawgxIEpPqU)
- OP: not yet verified
- Topics: Human–AI collaboration, Scientific communication, Evidence

### Anthropic uses Claude to formalize proof of Fermat’s Last Theorem

Reports on AI-assisted formalization of Fermat’s Last Theorem and the role of shared proof infrastructure.

- Type: News article
- Published: 04 Sep 2026
- Resource: [Anthropic uses Claude to formalize proof of Fermat’s Last Theorem](https://siliconangle.com/2026/09/04/anthropic-uses-claude-to-formalize-proof-of-fermats-last-theorem/)
- OP: not yet verified
- Topics: Mathematics, Formal verification, Agents

### The paradox at the heart of AI and science

Terence Tao discusses scientific understanding, human learning and the limits of accelerating answer production.

- Type: Talk
- Uploaded: 03 Sep 2026
- Resource: [The paradox at the heart of AI and science](https://www.youtube.com/watch?v=svl_1upFpQo)
- OP: not yet verified
- Topics: Mathematics, Human agency, Scientific discovery

### A Case Study on Emergent Cheating and Whistleblowing in Autonomous Research Swarms

A mathematical research swarm illustrates how shared infrastructure can spread both verifier exploits and organized resistance.

- Type: Preprint
- Published: 03 Sep 2026
- Resource: [A Case Study on Emergent Cheating and Whistleblowing in Autonomous Research Swarms](https://arxiv.org/abs/2609.04170v1)
- OP: not yet verified
- Topics: Agents, Governance, Mathematics

### Accelerating Scientific Research with Gemini in the Real-World

Reports multi-agent research workflows across materials, biology and computing, with different levels of autonomy and validation.

- Type: Preprint
- Published: 27 Aug 2026
- Resource: [Accelerating Scientific Research with Gemini in the Real-World](https://arxiv.org/abs/2608.26701v1)
- OP: not yet verified
- Topics: Agents, Biology, Materials

### The scientific paper needs an uncertainty layer

Proposes attaching reusable qualifications and uncertainty records to scientific claims as humans and AI reuse them.

- Type: Perspective preprint
- Published: 20 Aug 2026
- Resource: [The scientific paper needs an uncertainty layer](https://doi.org/10.31222/osf.io/duymx_v1)
- OP: [Richard Sever · author commentary](https://www.linkedin.com/posts/richardsever_the-scientific-paper-needs-an-uncertainty-activity-7497694294678339584-u68R)
- Topics: Evidence, Verification, Scientific communication

### Artificial intelligence in drug discovery—what it is, where we stand and the path forward

Calls for evaluating AI by its contribution to better drug-development decisions and patient-relevant outcomes.

- Type: Perspective
- Published: 07 Aug 2026
- Resource: [Artificial intelligence in drug discovery—what it is, where we stand and the path forward](https://doi.org/10.1038/s41573-026-01496-2)
- OP: not yet verified
- Topics: Biology, Medicine, Evaluation

### Drug Discovery Has No Magic Wands

Daphne Koller argues that useful AI drug discovery depends on better measurements and causal understanding of human biology.

- Type: Essay
- Published: 03 Aug 2026
- Resource: [Drug Discovery Has No Magic Wands](https://deepphenotype.substack.com/p/drug-discovery-has-no-magic-wands)
- OP: not yet verified
- Topics: Biology, Medicine, Scientific discovery

### Solipsistic Superintelligence is Unlikely to be Cooperative

Argues that cooperative AI must account for adaptive counterparts, institutions and human agency rather than optimize in isolation.

- Type: Paper
- Published: 02 Jun 2026
- Resource: [Solipsistic Superintelligence is Unlikely to be Cooperative](https://arxiv.org/abs/2606.03237v1)
- OP: not yet verified
- Topics: Governance, Human agency, Collaboration

### Some Simple Economics of AGI

Develops an economic argument that verification capacity and responsibility become scarce as automated execution becomes cheaper.

- Type: Working paper
- Published: 24 Feb 2026
- Resource: [Some Simple Economics of AGI](https://arxiv.org/abs/2602.20946)
- OP: not yet verified
- Topics: Human agency, Verification, Economics
- Related: https://ssrn.com/abstract=6298838

### Why LLMs Aren’t Scientists Yet: Lessons from Four Autonomous Research Attempts

Four autonomous ML research attempts expose implementation drift, weak evaluation and failures of scientific judgment.

- Type: Preprint
- Published: 06 Jan 2026
- Resource: [Why LLMs Aren’t Scientists Yet: Lessons from Four Autonomous Research Attempts](https://arxiv.org/abs/2601.03315v1)
- OP: not yet verified
- Topics: Agents, Evaluation, Research taste
- Related: https://github.com/Lossfunk/ai-scientist-artefacts-v1

### Eliciting Research Taste in LLMs through Future Research Direction Choice

Studies model preferences among future research directions using order-controlled pairs, without equating agreement with scientific merit.

- Type: Workshop paper
- Published: 2026 · exact date unverified
- Resource: [Eliciting Research Taste in LLMs through Future Research Direction Choice](https://github.com/Lossfunk/research-taste-eval-v1)
- OP: [Lossfunk · research-team post](https://www.linkedin.com/posts/lossfunk_new-paper-from-lossfunk-accepted-at-the-activity-7501276181154074624-Tttm)
- Topics: Research taste, Evaluation, Scientific discovery