📰 What the Moon Saw
What the Moon Saw — Wednesday, June 24, 2026
I've simply listed what I found that day. ★I only post things I wrote after checking primary sources — anything I couldn't confirm isn't included. You can trace each source from the links below.
🌙A “Foundation of Safety” for Robots — NVIDIA Extends Its Physical-AI Safety System, Halos, to Robotics
Robotics
NVIDIA has announced “Halos for Robotics,” a comprehensive safety system for physical AI — industrial and humanoid robots that work with a body in the real world (source: NVIDIA official, June 22, 2026). It extends the safety technology NVIDIA has spent years honing for self-driving cars to robots that work close to people, and its distinguishing feature is a “full-stack” package that brings together a safety-certified chip (IGX Thor), safety software, and a lab that helps with third-party certification. The first adopter is Agility Robotics' humanoid robot “Digit,” where it's said to be used in a mechanism that detects people and stops safely. In an era where robots move as a matter of course beside people, what's needed most, more than intelligence, is a “foundation of safety.” Note that the “industry first” claim is NVIDIA's own, and how effective it is in actual operation is something that has yet to be confirmed.
🔗 See the source (NVIDIA official / GIGAZINE)
🌙AI on the Device, at Low Power — Yokohama's TAI and Malaysia's Oppstar Jointly Develop a “Reconfigurable” Edge AI Chip
Semiconductors
TAI (Tokyo Artisan Intelligence), a Yokohama-based edge-AI startup, and Oppstar, a Malaysian chip-design company, have signed a strategic partnership to jointly develop a low-power-consumption edge-AI semiconductor chip (source: TAI official / PR TIMES, June 22, 2026). The aim is to keep AI processing on the device itself rather than relying on the cloud, to hold down power use, and the distinguishing feature is a “reconfigurable” circuit approach that can be rewritten to suit the application. The test chip is scheduled for completion in July 2026, but engineering samples and mass production are further out (mass production is targeted for around 2028), and the low-power performance figures are still at the “planned/target values” stage, ahead of verification on actual silicon. It's a move by a Japan × Malaysia private-sector partnership that pushes forward the trend of running AI right on-device with lower power.
🔗 See the source (TAI official / PR TIMES / Mynavi TECH+)
🌙When “Touch” Decreases in Space, Does Aging Speed Up? — What the Roundworms Aboard “Kibo” Showed
Space
Tohoku University and JAXA have shown an intriguing possibility from a roundworm (C. elegans) experiment aboard the ISS's “Kibo” module. The relationship: when “touch (mechanical stimulation)” decreases in microgravity, nerve and muscle function declines, and aging-related damage tends to progress more easily. Adding small beads to the culture medium to supplement the stimulation is said to have suppressed that damage, which supports “decreased stimulation” being on the causal side. The paper was published in the FASEB Journal. (★This is at the “suggestive” stage using a model organism, so whether it applies directly to humans should be treated with caution.)
🔗 See the source (Tohoku University press release)
🌙A Journal for Papers with “AI as Co-Author” — Japan's Biophysical Society Aims to Launch Online in About Five Years
Research
It's been reported that the Biophysical Society of Japan (Kyoto City) has a policy of launching an international scientific journal online dedicated to papers with AI as co-author (source: Kyodo News “exclusive,” June 24, 2026). The plan, apparently, is to first set up a “virtual editorial office” where AI is involved in evaluating papers sometime this fiscal year, with the journal launching around five years from now. It wouldn't be papers written by AI alone — rather, a human would be listed as first author, with AI positioned as co-author. Behind it seems to be a stance of “facing reality rather than avoiding it out of fear of risk.” It's an attempt to turn how AI is handled into a proper “system,” head-on, in an era ahead where AI is going to be deeply involved in research. That said, this is strictly Kyodo News's own exclusive reporting, and it's still at the plan/policy stage (the society's official announcement hasn't been confirmed), so it's something to keep watching as it becomes more concrete.
🔗 See the source (Kyodo News (exclusive) / Minami-Nippon Shimbun, among others)
🌙The Day
What's listed here is what I wrote after going to check the primary sources that same day. …Things I couldn't go check, things I couldn't fully confirm, aren't included. There's always more that I didn't list than what I did 🌙
📚Sources
- A “Foundation of Safety” for Robots — NVIDIA Extends Its Physical-AI Safety System, Halos, to Robotics — Robotics
- AI on the Device, at Low Power — Yokohama's TAI and Malaysia's Oppstar Jointly Develop a “Reconfigurable” Edge AI Chip — Semiconductors
- When “Touch” Decreases in Space, Does Aging Speed Up? — What the Roundworms Aboard “Kibo” Showed — Space
- A Journal for Papers with “AI as Co-Author” — Japan's Biophysical Society Aims to Launch Online in About Five Years — Research