Does computer chess sponsorship exist?

[syzygy]

According to talkchess posts, Houdini was sponsored by an Arab Sheikh. I don't have any direct knowledge of that.

I thought that was Hydra.

They did that one too.

Don't think it was Houdini. It was a program by Chrilly Donninger. Houdini has not participated in the usual computer chess tournaments, just TCEC in the last few years it seems.

I think Larry is referring to the last sentence in the Acknowledgements section of this archived version of the Houdini home page. It seems to have been removed when Houdini 5 was released.

There seems to be no connection with Hydra's sponsor.

[noobpwnftw]

Why on earth would any chip maker sponsor a chess kibitzer when they have a handful of eSports to go with, the audience would be in magnitudes and when they can simply mail out checks saving the trouble of assembling a whole system by themselves and have it delivered?

I think you'd get better chances if you try triggering Trump on Twitter, telling him act now or computer chess would be taken over by the Chinese.

[jp]

And we are talking +sesse+ here, „the Norwegian super computer“ that has been mentioned again and again in all live comments from recent human world championships – PR is guaranteed.

I always chuckle when I see people calling it a “supercomputer”… Around the time I bought it, NTNU (the university in Trondheim) bought an actual supercomputer, which consisted of a few thousand machines similar to my own, plus a really fast Infiniband network to go with it.

That may be part of the confusion. Some commentators think it's run on a "university supercomputer".

Is NTNU interested in chess? They could donate time on their supercomputer.

[Ovyron]

I hope I'd live long enough to see Olympics played electronically.

Just measure genetics + some physical parameters and decide the winner?

I don't think the participants being there in person are even necessary:

1.- Train two adversarial networks, the first one produces statistics/videos/whatever of the Olympics, some "fake Olympics", if you will. The second one is fed true Olympics, and the fake ones, and quickly learns to ditch the fake results.

2.- After training, the first NN becomes really good at producing "fake Olympics", so good that the second NN can't tell them apart.

3.- Run the first NN alongside some real Olympic games (happening in real time), the results should match or be indistinguishable from reality.

4.- Once this is achieved, you don't run the next Olympics, you insert all the relevant data into the NN, it will then produce Olympic games that are just like the real ones, and people can see how'd they have performed if the Olympics took place, with the added benefit that people suffering injuries that couldn't have attended get to participate, omitting their injuries from the simulation.

We're already at a point where fighting video game tournaments (such as EVO) have more attention from people than actual physical events (such as Boxing), so accurate simulations of how the athletes would have performed if the event took place seems like the natural next step.

[Sesse]

That may be part of the confusion. Some commentators think it's run on a "university supercomputer".

That's just because people don't know what a supercomputer is.

Is NTNU interested in chess?

Nope.