You'd be surprised at how productive many of them are.
Just as you've been trained to react to the academic stimulus, memorize all the emacs' shortcuts and whatnot; you can also be trained to acknowledge many of the things other immigrants do.
Can you share more about your particular setup? I use a pretty vanilla setup of Doom emacs on Linux, and while I really wish to give exwm a try my experience with emacs has been too unstable so far. E.g. it sometimes crashes when it gets an I/O error trying to write a file (which happens when a USB drive is removed by accident). A more common annoyance is the entire program freezing while waiting for plugins that should be asynchronous, like Tramp or some LSP servers.
Have never tried it personally, but this talk made me think collaborative editing is a solved or close-to-solved problem in Emacs: https://emacsconf.org/2023/talks/collab/
Nice work, one thing (which you'll see everywhere once you've fixed it yourself) though: in TeX you want to escape functions like cos, i.e. use "\cos" not "cos", which will get rendered as a product of 3 variables c, o and s.
Not only did we get a whole new type of Chess engine, it was also interesting to see how the engine thought of different openings at various stages in its training. For instance, the Caro-Kann, which is my weapon of choice, was favored quite heavily by it for several hours and then seemingly rejected (perhaps it even refuted it?!) near the end.
The super cool thing about MuZero is that it learns the dynamics of the problem, i.e. you don't have to give it the rules of the game, which makes the algorithm very general. For example, DeepMind threw MuZero at video compression and found that it can reduce video sizes by 6.28% (massive for something like YouTube)[2][3].
Curious if anyone else knows examples of MuZero being deployed outside of toy examples?
To be fair, it uses MCTS, which requires many simulations of the game. For this, it needs to know which moves are valid, and when a player wins or loses the game.
So it does need to know the rules of the game, but it doesn't need any prior knowledge about which moves are better than others.
Not quite, you can define an illegal move as losing the game, and winning/losing is a “meta-observation” - ie if the player wins/loses, you don’t invoke another search.
Why not update the API contract that if you use the Reddit API you need to render the ads? If you don't render the Ads they can come up with some pricing model. Seems like Reddit's pricing is designed by an MBA intern who just joined silicon valley.
That would require the API to also include ad information. I think putting a sensible price on API access would make this a bit simpler, though (and allow for a paid ad-free experience as well). The core issue isn't so much paid API access as the price being so crazy high it can't possibly reflect how much reddit would normally be making from the users of these apps, so instead looks like it is aimed at just shutting down 3rd party apps entirely.
