News for hackers. Stuff that matters.
Re:Show HN: Watch a neural net learn to play Snake by c1b on Friday May 15, 2026 @03:14PM (#48152630)
Yes it just collapses eventually — never stabilizes. The training process is flawed, I suspect it has to do with the fact that some weights blow up over time, you can see in “weights” tab.

But at around 4K avg score you should see it solve the env almost every time.

Just a demo :) optimized for speed over stability.

Reward structure: Step: -1 Dot: +100 Win: +1000 so ~4k is max theoretical score on 6x6; the model will eventually break if you keep training long enough.

Re:Show HN: Watch a neural net learn to play Snake by ticulatedspline on Friday May 15, 2026 @06:40PM (#48154881)
maybe because it doesn't understand "done"? perfect play is impossible, random variance will cause scores to drop even if the model plays well and "wins". feels like it would get stuck in a loop trying to improve what can't be improved.
Re:Show HN: Watch a neural net learn to play Snake by jesuo on Friday May 15, 2026 @07:59PM (#48155479)
feels like it would get stuck in a loop trying to improve what can't be improved.

That is the point, there is nothing on an intention that we cannot improve, the goal here is no more than 1 unique iteration of the same path

Re:Show HN: Watch a neural net learn to play Snake by fc417fc802 on Friday May 15, 2026 @07:05PM (#48155078)
The optimizer doesn't need to understand anything it's just an iterated mathematical construct. The author simply didn't bother to implement the necessary details to ensure numerical stability.

Alternatively it might be a problem with the scoring model in the end game.

Re:Show HN: Watch a neural net learn to play Snake by Tostino on Saturday May 16, 2026 @08:38PM (#48165091)
That is what I thought op was saying when he used the word "understood". No need to jump on people using every day language that is still easily understood in context IMO.