Early science acceleration experiments with GPT-5
2025/11/20 by Sébastien Bubeck, Bubeck, Sébastien, Christian Coester +27 · 10 voices · 13 citations
Computer Science · Decision Sciences · Medicine · #Acceleration #Artificial Intelligence in Healthcare and Education #Computability, Logic, AI Algorithms #Data collection #Expert system #Frontier #Human research #Scientific Computing and Data Management #Scope (computer science) #cs.AI #cs.CL
paper · pdf · doi:10.48550/arxiv.2511.16072
published in arXiv (Cornell University) (Cornell University)
openalex publication_date 2025/11/20 · openalex created_date 2025/11/23 · openalex updated_date 2026/08/05
Abstract
AI models like GPT-5 are an increasingly valuable tool for scientists, but many remain unaware of the capabilities of frontier AI. We present a collection of short case studies in which GPT-5 produced new, concrete steps in ongoing research across mathematics, physics, astronomy, computer science, biology, and materials science. In these examples, the authors highlight how AI accelerated their work, and where it fell short; where expert time was saved, and where human input was still key. We document the interactions of the human authors with GPT-5, as guiding examples of fruitful collaboration with AI. Of note, this paper includes four new results in mathematics (carefully verified by the human authors), underscoring how GPT-5 can help human mathematicians settle previously unsolved problems. These contributions are modest in scope but profound in implication, given the rate at which frontier AI is progressing.
Cited by
Discussions
- People on BlueSky: AI is useless! A stochastic parrot! Mathematicians/biologists/physicists: It is already helping us do frontier technical research and in some cases solve open problems arxiv.org/pdf [bsky, 159 points, 67 comments]
- Early science acceleration experiments with GPT-5 [hn, 4 points, 1 comments]
- Early #science acceleration #experiments with GPT-5 #AI #largereasoningmodels #LLM 🧪 arxiv.org/pdf/2511.16072 [bsky, 4 points, 1 comments]
- Ganz so sinnfrei scheint KI in der Forschung nicht zu sein, zeigt diese Studie. Findet vergessene Ergebnisse, führt Dinge zusammen, auf die ein Mensch nur schwer kommt etc. Zeitgewinn bis zu Faktor 10 [bsky, 3 points, 1 comments]
- arxiv.org/abs/2511.16072 [bsky, 2 points, 0 comments]
- New paper from OpenAI: Early science acceleration experiments with GPT-5 arxiv.org/abs/2511.16072 [bsky, 2 points, 0 comments]
- Do you know anyone like this: "He can confidently make mistakes, ardently defend them, and confuse himself (and us) in the process"? This is the amusing characterization of ChatGPT by this paper that [bsky, 1 points, 1 comments]
- Research Paper (preprint): "Early #Science Acceleration Experiments with GPT-5" arxiv.org/abs/2511.16072 #AI #LLMs ||| Media Coverage via @thedecoder.bsky.social the-decoder.com/openai-repor... [bsky, 0 points, 1 comments]
- Early science acceleration experiments with GPT-5 arxiv.org/abs/2511.16072 [bsky, 0 points, 0 comments]
- And this is a fascinating paper (if you can get your head past the science - Early science acceleration experiments with GPT-5 arxiv.org/pdf/2511.16072 [bsky, 0 points, 0 comments]
Related