- Sabine's Newsletter
- Posts
- America’s Genesis Mission, Artificial Neurons, AI-powered Papers, and the End of an Era
America’s Genesis Mission, Artificial Neurons, AI-powered Papers, and the End of an Era
This week’s science bits from SWTG
Marcus Gee-Lim
January 07, 2026
|
Happy New Year, readers! Hopefully you had a restful holiday break with family and friends, and I’m looking forward to sharing all of 2026’s best science headlines with you all!
US Government Embraces AI Science with New “Genesis Mission”
The US government has fully leaned into AI science by launching the Genesis Mission in late 2025. In their own words, it’s “a dedicated, coordinated national effort to unleash a new age of AI‑accelerated innovation and discovery that can solve the most challenging problems of this century.” It tasks the Department of Energy’s (DOE) 17 National Labs to build an integrated AI-driven discovery platform, in partnership with industry and academia. As part of this mission, Google DeepMind agreed to provide DOE scientists early access to its frontier AI tools. It’s a smart move, and I am somewhat distressed (though not overly surprised) that the EU has no comparable initiative.
This week’s episode of Science News is about a new era for physics. It’s a new year, and according to a new paper, physics is headed into a new era. Over the past few years, the field of physics has seen a major shift in research priorities – let’s take a look at that shift, why it’s happened, and where physics is headed next.
Artificial Neurons See Dramatic Improvement
Figure: Zhao et al, Nature Electronics 8, 1211 (2025)
Researchers from the University of Southern California have shown that many key behaviours of a biological brain cell can be reproduced using just three tiny electronic parts: a switch called a “memresistor,” a standard transistor, and a resistor. Earlier artificial neurons relied on dozens of components and consumed tens to hundreds of picojoules per signal. Digital neuromorphic chips often use even more energy. The new device is not only much simpler, it operates at just about one picojoule per signal, one to two orders of magnitude lower than many previous hardware neurons. The authors estimate that further miniaturisation could push this down to the attojoule range, which would place it below the estimated energy used by a biological neuron. They also ran simulations which show that networks built from these artificial neurons can still perform pattern recognition tasks with high accuracy. This could become a route toward much more energy-efficient brain-inspired computing hardware.
Paper here.
Learn how to make AI work for you
AI won’t take your job, but a person using AI might. That’s why 1,000,000+ professionals read The Rundown AI – the free newsletter that keeps you updated on the latest AI news and teaches you how to use it in just 5 minutes a day.
Scientists Who Use AI Become Stunningly More Productive
Figure: Kusumegi et al, Science 390, 6779, 1240 (2025)
In a new paper that just appeared, the authors found that researchers who use large language models (LLMs) become stunningly more productive. An analysis of more than 2 million papers revealed that those who adopt AI increase their paper output on average by 40%. For non-native English speakers this productivity jump is even larger, up to 80%. The scary part is that this is data only until July 2024, when scientists used LLMs mostly for writing. It is foreseeable that in the next few years the AI-adoption rate will go to nearly 100% across all scientific disciplines, resulting in more papers than ever and putting significant strain on peer review. Paper here.
