@grok thanks @grok you made it in @cursor_ai with Grok 4.7, i just asked and received. The server feature and simultaneous port to mobile was a one shot success today.
Bespoke software: I made a local mp3 player, that i've been using for 6 months, and today added a web demo at https://t.co/rtIG3NqsXk,
Brings back that old time music listening feel, In the repo
https://t.co/SVmMobIQqT I provide the code for you to install and play your fav..
@Tailscale@googlechrome library. And what about visualizers. While KyTunes ships with ready to go p5.js based visualizers with a simple click. And because its @p5js, you can even edit the code live in the app and make any visualizer you want. Your art can be saved and shared or added to the repo...
with @Tailscale to share the songs from my server with my other mac computers running the KyTunes app, and with my android phone running @googlechrome progressive web app streaming the songs. The app runs in the background and with the screen locked to play music from your libr..
The optimality of the packing for 11 squares has been formalized in lean thanks to Astra and Claude! Huge thanks to @ojoshe, @kleddamag, @wand_125, @guzhou0806, and @ctjlewis for aiding in the process.
Image credit: https://t.co/2oUaPKe2Hz
1/n
Can we replace individual neural network weights with human-readable Python code? Sorry to be late... but I'm super thrilled to share that our paper was accepted to #NeurIPS26.
"Individual Parameters in Weight-Sparse Transformers Appear Interpretable" (w/ @sheimersheim!!)
Can we replace individual neural network weights with human-readable Python code? Sorry to be late... but I'm super thrilled to share that our paper was accepted to #NeurIPS26.
"Individual Parameters in Weight-Sparse Transformers Appear Interpretable" (w/ @sheimersheim!!)
I am afraid that a useful conjecture has not been stated in learning theory in spite of the fact that say the new model can answer old unanswered questions that theory folks used to care about.
I’m very curious for example about non proof enabling assumption style theory that has predictive power eg one that says “given this real data set this real algorithm this real model and this realistic HP choices we can fit the training labels”
Knowing nothing about what these 100 problems are, but expecting their style to pattern match “problem theory people could not solve for N years” I’m willing to buy you dinner at French laundry or whatever if any of them are even close to the the style of the one from above.
Not because Bel or whatever the model is called can’t solve it, but because the above question never bothered theoreticians
Here’s a question I tried but failed to answer with astra
“Given lenet5 and mnist prove that any practically relevant variant of SGD will fit the labels and do so in Flops less than a small fraction of what it would be required to actually train the model till it fits the labels”
Now trad theory doesn’t care about this problem, as it sounds too pedestrian, yet I think it’s an instance of what useful math will look like in a few years
Sorry about the ramble.
A week remembering a loss that still makes no sense. Life doesn’t stop, and neither do the memories. Never forgotten, always living in his spirit. @drdagly
Canadian soccer player, Stephen Antunes Eustáquio playing soccer as a kid in Leamington, Ontario. Eustáquio just scored to advance Canada further into the World Cup competition. 🇨🇦