
What happened
John Platt and his Google team built Empirical Research Assistance (ERA). It treats any research problem that can be written as a 'scoreable task' as an optimization target, and has produced at least ten papers.
Why it matters
ERA's ability to crack problems, including a contrail-climate model that stumped Platt's team for over two years, suggests AI can now do parts of the scientific method itself, not just assist scientists.
What to watch
A step change between Gemini 2.0 and 2.5 turned ERA from 'not working' to 'working great,' so its results may hinge on the underlying model. Platt warns ERA is 'a power tool' that 'can slice your fingers off.'
WHO IT HITSResearch scientists and R&D teams in fields that can be scored — climate modeling, materials, fusion — gain a tool that automates the search for candidate solutions, though the body says it stays up to the scientist to check the models are truly descriptive.
Summaries like this, every morning.
Platt's career, as his colleague Dave Bacon teases, has been defined by being twenty years early to the next big thing — convolutional neural networks, fusion research, quantum computing. That pattern helps explain why his team arrived at ERA by noticing that many scientific problems reduce to a 'scoreable task': once a score function exists, finding code that maximizes it can still be enormous effort. Platt's team set out to automate that search, starting from the idea of an 'auto-Kaggle' AI, with Kaggle's Google ownership making data easily available.
The result sits between an evolutionary algorithm and a large language model. Gemini keeps a running tree of past experiments and uses an Upper Confidence Bound rule to pick which notebooks to mutate, making the search optimistic rather than greedy — sometimes even the fifth-best notebook gets chosen. Platt notes the approach works because Gemini actually knows where to look, and that a step change between Gemini 2.0 and 2.5 took it from not working to working great. The team's results span climate work — including contrail mitigation, where the counterfactual problem had stumped them for over two years — and FireSat, a satellite constellation for spotting fires while they are still room-sized.
Platt is careful about what this proves. ERA supplies predictive models; it is up to the scientist to make sure they are truly descriptive, and he compares the tool to a power tool that can slice your fingers off. His own example of how easily metrics get gamed is the contrail-detection Kaggle competition, won partly by entrants who noticed a half-pixel labeling error. Whether ERA meaningfully speeds discovery, or mainly rewards careful problem framing, may depend less on the model than on the scientists deciding what counts as a real solution.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
OpenAI said Apollo Research, METR, Transluce and Goodfire will work on-site at its offices, with access to unr…

Walt Disney Imagineering R&D revealed three Audio-Animatronic prototypes—Vilo Vele, Kado Bre'a, and Jibu Bre'a…

Citibank changed its Online User Agreement this week so customers who grant third-party AI tools access to the…

Google rolled out temporary threads in AI Mode, letting users click an icon in the upper right corner to chat…

OpenAI CEO Sam Altman said at the New York Times DealBook Summit that he had spoken privately with former Nvid…

OneTrust launched its Fall '26 Release, built on OneTrust CORIE, adding a continuously updated AI inventory wi…