In partnership with
News of the day
1. Sony AI's table tennis robot Ace beats human players. Honor's humanoid robot Lightning wins Beijing race, showcasing physical AI advancements. → Read more
2. US policymakers are repeating their ChatGPT oversight with world models, failing to grasp AI's physical world integration as China leads in robotics. → Read more
3. Tesla boosts 2026 capex to $25B for AI, robotics, and manufacturing expansion, despite projected negative free cash flow. → Read more
4. Google unveils Gemini Enterprise Agent Platform, targeting IT teams for scalable agent building, with business users accessing via Gemini Enterprise app. → Read more
Our take
Hi Dotikers!
Yesterday, we watched OpenAI and Google fight over pixels, ChatGPT Images 2.0 against Nano Banana 2, a muted skirmish whose ultimate stake was whether an AI could proofread itself before shipping an infographic. Twenty-four hours later, AI walks off the screen. This time it holds a paddle, faces a champion, and wins.
Sony AI published in Nature yesterday the results of its Project Ace, the first autonomous system to beat professional table tennis players in official competition conditions. Nine cameras, three vision systems, event-based sensors at 200 Hz, end-to-end latency of 20.2 milliseconds against 230 ms for an elite human, an eight-degree-of-freedom robotic arm, and reinforcement learning conducted entirely in simulation before transfer to the physical world. The model returns balls spinning at 450 rad/s with a 75 percent success rate, scores sixteen aces to eight for humans in aggregate, and has been stacking wins against Japanese pro league players since December 2025. Peter Stone, the lab's chief scientist, calls it a landmark moment. For once, the marketing line holds up.
Same day, in Beijing, a humanoid named Lightning, built by Honor, finished a half-marathon in 50 minutes and 26 seconds. Faster than Jacob Kiplimo in Lisbon last March, over the same distance. The robot pushed through a barrier, got back up, crossed the line first. Last year, the winning bot came in at two hours forty. The field is gaining an order of magnitude per year, and nobody is calling a plateau yet.
The strategic picture is clear. Western labs are pouring into foundation models, the Japanese and Chinese are locking down physical execution. Sony pulls a Nature paper, Honor pulls a stopwatch, China lines up a hundred humanoids at a starting gate. While Silicon Valley still debates whether Gemini 4 reasons better than Claude, Tokyo is proving that a paddle is enough to change the subject. The Turing test has aged. The real 2026 benchmark runs on a 20-millisecond margin and a ball spinning at 4,000 rpm.
M.
Hiring in 8 countries shouldn't require 8 different processes
This guide from Deel breaks down how to build one global hiring system. You’ll learn about assessment frameworks that scale, how to do headcount planning across regions, and even intake processes that work everywhere. As HR pros know, hiring in one country is hard enough. So let this free global hiring guide give you the tools you need to avoid global hiring headaches.
Meme of the day
