AI ROBOTICS
Someone give the humans a chair
Sony AI has showcased “Ace”, a robotic table-tennis player that can compete under official table-tennis rules.
According to Reuters, The Guardian and a Nature study, Ace won three out of five matches against elite human players and stayed competitive in the others.
The robot uses nine cameras, an eight-jointed robotic arm, and a movable base.
It was trained with reinforcement learning, which means it learns through practice instead of being manually programmed for every shot.
That is important because table tennis moves fast.
It has to track the ball, read its spin, predict where it will go, and respond in real time.
The bigger story is not just that a robot can now ruin someone’s pub-table-tennis confidence.
It is that AI is getting better at controlling machines in fast, messy physical environments.
The main takeaways:
Ace beat elite players in three of five matches.
It uses cameras, robotics, and reinforcement learning to react in real time.
The tech could help industrial robots, but safety questions will matter.
Ace came to ruin lunch breaks
This could help in areas like manufacturing, where robots need to react quickly to moving objects.
But researchers also note that high-speed robotics could raise safety concerns if used outside controlled settings.
For now, Ace is still a specialised robot built for one task. It doesn’t mean general-purpose robots are suddenly ready for everything.
But it does show how far AI-powered robotics is moving.
Table tennis was already stressful. Sony added a metal opponent with cameras. Lovely.- MG