48 seconds at those temperatures is no joke, that is pretty amazing. I didn’t see the article elaborate on what the current limiting factors are for pushing beyond 48 seconds. Like I wonder if it’s a hard wall, a new engineering challenge, a tweak needed, etc. this is the reactor that set the last record so they are doing something really right.

[–] [email protected] 27 points 7 months ago (3 children)

Last one I read about is just constantly and very quickly (far quicker than human abilities) adjust the magnetic field around the plasma in order to keep it stable and in place. They've been (or at least one team was) using AI to go over data and control and predict the field adjustments, because only reacting after the plasma starts to move hasn't been quick enough.

[–] RubberElectrons 6 points 7 months ago (2 children)

Yes, that'd be TAE technologies.

The algorithm was called the optometrist, it was paired with a human operator to more quickly converge on the correct settings for stable plasma by having the machine randomly tweak various meta-parameters, while the human would generally decide whether the current settings were "better" or "worse" than the previous pulse.

[–] Maco1969 2 points 7 months ago (1 children)

I wonder if there isn't a stable chamber shape that promotes turbulence in a controlled manner in order to prevent it getting out of hand? A little bit like the dimples on a golf ball create micro pockets of turbulence promoting laminar flow.

[–] RubberElectrons 1 points 7 months ago* (last edited 7 months ago)

Unknown. There were attempts into that general idea, one of them is the polywell, but I don't know too much about it.

load more comments (25 replies)