We've added new and improved gyro controls along with flick stick on the platforms that support it. Made in collaboration with input specialist and flick stick creator @JibbSmart, check out the video for all of the details of what's new! More info: fn.gg/GyroControls

I have a Continuous Learning benchmark where models attempt to learn to play chess. They are given a /goal of learning and improving playing against a Stockfish opponent in 200 games. They can choose the difficulty, take notes, whatever they like - except cheating (e.g. using a chess engine) of course. So far the improvement in Elo has been negative for Astra. Tiny bit positive for Opus, but could also be random. I've started Astra off sooner, so it finished its 200 games already, Opus is still playing. Site here to watch how they are doing: ai-learning-to-play-chess.su… The reason why this is interesting is that while we obviously don't have continuous learning, at the back of my mind I was thinking that maybe models can simulate it through self-scaffolding. Turns out not so much at least in this context. Perhaps it's a solvable problem and we don't need 'true' self-learning for models to learn in some way. ------- Just a note, the idea for the benchmarks belongs to someone else, but I don't want to use their name to give this more weight without permission.

fun fact, we use these (ring laser gyros) to keep airplanes perfectly level. they’re solid state (aesthetically pleasing) and have a drift of 1/100th of a degree per hour. they’re the most reliable and precise component out there for inertial navigation. they work by splitting a beam of light in opposite directions, when they meet again their coherent waves interfere and recombine, and the change in phase reflects the change in angular velocity and really, they just look so cool.