Hacker News Reader

Comment by t3rabytes | Hacker News Reader

I haven't seen a particularly great explanation that explains why this is such a positive discovery that is copy/paste/digestable. No MSM outlet is going to run this story unless they can say an impact.

View on HN

There are a ton of them on Twitter.

Yes and many of them are completely wrong.

I keep seeing people saying "CPU's that don't generate heat". How exactly would that work? When a transistor turns off/ switches to 0, where does it dump the electrons? Hint: into heat

Sorry, can you explain? My understanding is that transistors don't "dump" electrons anywhere. The gate controls the voltage, which in the `0` state forces resistance to be high enough s.t. current flow through that transistor stops.

As the Veritasium video explains [0], current flow is not a literal flow of electrons, but a state of the electromagnetic field (or something to that effect... it's been many years since my electromagnetism university courses).

[0]: https://youtu.be/bHIhgxav9LY

current flow isn't always a literal flow of electrons, but in a wire or transistor it is

you probably should trust your electromagnetism textbook more than a youtube video

There will literally always be heat from computation, that's Landauer. The point is that LK-99 could be or lead to a breakthrough in making computation produce much, much less heat because you're getting rid of most resistivity losses. Which is important, because heat is actually a serious constraint on a bunch of use cases.

Reply on news.ycombinator.com