@AnthropicAI just highlighted us in their latest research on AI-driven scientific discovery. This is where things are heading. 🧵

Dario: “The biggest thing to watch is this issue of AI systems building AI systems.” Today, Autoscience is announcing $14M in funding by General Catalyst, Perplexity Fund and Toyota Ventures to create autonomous AI research labs. Human AI researchers can't keep up with the pace of new AI research. They’re out of time to run the experiments they want. So we’re building an autonomous AI lab that can. The era of human-scale R&D is over. Machine-scale development has begun. 🚀

AI systems can write working code … but can they earn a medal on a live Kaggle competition with $50k in prizes?

Human textual edits to this 7-page paper were limited to formatting and citations. We’re very concerned about academic integrity, so we chose to manually edit citations to provide the highest possible standard of academic attribution.

First Full-Length AI generated paper passes peer review! Until now, the longest AI generated paper to pass peer review was 4 pages. The Bi-Align Long Track (9 pages) just accepted our paper, building on our previous success of the first AI-generated paper to pass peer review. 🧵👇 autoscience.ai/blog/ai-resea…

Academics must take responsibility for their citations, even if the work is AI generated. That’s why when we submitted this accepted AI generated paper to ICLR workshops, the only human textual edits were to citations: Check out our accepted AI generated papers here: xitter.cf/AutoScienceAI/status/1…

Solving multi-step coding tasks is hard. Open-ended code generation for novel research experiments is even harder. We built Carl in a three-stage process that allows the system to ideate, execute, and present ideas in an iterative manner.

At Autoscience, we’re solving autonomous AI research. We just released a technical report on Carl, the first AI system to create an end-to-end AI research paper accepted past peer review. 🧵