pizza

21 ms·

151.
Diffusion Models Learn Low-Dimensional Distributions via Subspace Clustering(arxiv.org)
1 points by pizza 2y ago|0 comments
152.
Latent computing by biological neural networks: A dynamical systems framework(arxiv.org)
2 points by pizza 2y ago|0 comments
153.
Federal prisons prep to move trans inmates as early as next week(npr.org)
19 points by pizza 2y ago|58 comments
154.
by pizza 2y ago
For years I tried and never got it. Though these days, I only remember, or have, a dream once every few months.
155.
by pizza 2y ago
Serious questions: are the country's institutions strong enough to provide for checks and balances against the executive branch as it grows unchecked, in reality? If so, how does that play out? Are the timescales of the checks and bala
156.
by pizza 2y ago
What's wrong with evidence?
157.
Analysis of fiber optic wired drones in Ukraine(twitter.com)
2 points by pizza 2y ago|0 comments
158.
An Experimental Analysis of RowHammer in HBM2 DRAM Chips [pdf](people.inf.ethz.ch)
7 points by pizza 2y ago|0 comments
159.
by pizza 2y ago
You might be able to try the lynx terminal web browser
160.
by pizza 2y ago
This almost surely took place somewhere in the past week alone, just with a lawyer being the mediating human face.
161.
by pizza 2y ago
You might be able to load up all the txt files into the embedded vector db in `llm` so that you could also query them semantically.
162.
by pizza 2y ago
I think bluesky might fit what you're looking for, in the sense that either you can get chronological mode (the default feed), or use an recommendation algorithm of your own design.
163.
Automatic Stochastic Differentiation(jiha-kim.github.io)
1 points by pizza 2y ago|0 comments
164.
by pizza 2y ago
General advice? Collect data, train a model, note the mistakes in the model, mistakes in the data, and think critically about what it is that you're ending up teaching. Repeat many, many, many times.. For some tasks, don't be surp
165.
by pizza 2y ago
This is the guy you're equivocating with the average government employee: https://krebsonsecurity.com/2025/02/teen-on-musks-doge-team-...
166.
by pizza 2y ago
Not sure how true this all is, but it's a centralized discussion of different systems of thinking I've seen associated with the current administration's tack https://www.politico.com/news/magazine/20
167.
by pizza 2y ago
Have you considered that DOGE is an opaque, unaccountable bureaucracy that is motivated to optimize for ragebait? Who don't seem to know the consequences of their actions, or have any skin in the game if they make mistakes? Who have be
168.
Curriculum Interleaved Online Behavior Cloning for Complex RL Applications(researchgate.net)
3 points by pizza 2y ago|0 comments
169.
Is my "red" your "red"? Color similarity judgments using unsupervised alignment(cell.com)
3 points by pizza 2y ago|2 comments
170.
A vector-based strategy for olfactory navigation in Drosophila(biorxiv.org)
1 points by pizza 2y ago|0 comments
171.
by pizza 2y ago
I guess then, why is what that solution lacking physical meaning versus the normal one - in some sense that means what it corresponds to is something other than a part of physics, but then what is that?
172.
by pizza 2y ago
The Minkowski metric is Δs = sqrt(-Δt^2 + Δx^2 + Δy^2 + Δz^2) One aspect of this is that, if you sub `t -> -t'`, that's just as good a solution too. Which would suggest any solution with a positive time direction can h
173.
by pizza 2y ago
I'd suggest spending more time revisiting your calendar/remembering your current work, and/or finding what's not worth it any more. That way, you can either better say 'no' at the moment of new proposals, or ma
174.
by pizza 2y ago
Ah, you know what, I misunderstood the situation: it got moved to an archived copy of the past version of whitehouse.gov: https://www.reddit.com/r/rust/comments/1i7m431/whitehouse_pr...
175.
by pizza 2y ago
I know you're being facetious but they literally deleted the Biden administration's recommendation to develop software using memory-safe languages.
176.
by pizza 2y ago
I guess there's 3 players in the games that often get invoked in discussions about AI cogsec: - the corporation running the AI - the user - the AI, sometimes - depending on the particular conversation's ascription of agency to the
177.
by pizza 2y ago
This is going to be incredibly useful for a lot of people!
178.
by pizza 2y ago
I got the impression from somewhere that they used the simplest machine learning techniques (just fitting regressions to data), but that it was "the 'what' that they decided to fit" that was the secret sauce.
179.
by pizza 2y ago
Just realized that the Mixture of Million Experts paper from last year is similar in some respects to this tiny pointers idea
180.
Generalized Transformers from Applicative Functors(glaive-research.org)
2 points by pizza 2y ago|0 comments

More ›