ziofill

22 ms·

391.
by ziofill 2y ago
I have to WFH 100% of the time because my company’s HQ is three time zones away. I think I would be more productive if I could go to the office once or twice a week and have face to face interactions with my colleagues rather than always on
392.
by ziofill 2y ago
> Putting the processes aside, if you black box a human and a language model and put us head to head on reasoning tasks, sometimes you're going to get quite similar results. I cannot believe this is true. LLMs are awful at whatever
393.
by ziofill 2y ago
Yes it does, look at the caption of Fig. 1: "Photons we receive that were emitted by objects beyond the Hubble sphere were initially receding from us (outward sloping lightcone at t <∼ 5 Gyr). Only when they passed from the region o
394.
by ziofill 2y ago
https://arxiv.org/abs/astro-ph/0310808
395.
by ziofill 2y ago
Actually, the parts of the universe receding from us faster than the speed of light can still be causally connected to us. It’s a known “paradox” that has the following analogy: an ant walks on an elastic band toward us at speed c, and we s
396.
by ziofill 2y ago
I assumed it was a soundless video and I played it while my 2 year old was falling asleep. Not asleep now :facepalm:
397.
by ziofill 2y ago
Maybe I am not understanding something: what if the surface is a plane? The wobble can’t go away in that case right?
398.
by ziofill 2y ago
A professor I know once gave me the most beautiful example of a zero knowledge proof: she said “Imagine I want to prove to you that I can count the leaves on a tree, but I don’t want to reveal how I do it. I start by telling you there are e
399.
by ziofill 2y ago
But while keeping the data constant, adding more and more parameters is a strategy that works, so what gives? Are the functions getting somehow regularized during training so effectively you could get away with fewer parameters, it's j
400.
by ziofill 2y ago
You can fit any data with enough parameters. What’s tricky is to constrain a model so that it approximates the ground truth well where there are no data points. If a family of functions is extremely flexible and can fit all kinds of data ve
401.
by ziofill 2y ago
I don't find this hard to believe. I'm no brain scientist, but dance puts together several senses and proprioception: music, rhythm, one's position in 3D space, balance, and physical exercise...
402.
by ziofill 2y ago
“I was excited. I almost gave up fishing for the day.” I love how he almost gave up fishing.
403.
by ziofill 2y ago
Is it standard practice to put lots of code in the `__init__.py` files?
404.
by ziofill 2y ago
I’m taking a crack at ARC-AGI. Very curious to see where I land.
405.
by ziofill 2y ago
https://www.zdnet.com/article/20-years-later-real-time-linux... This article explains well why PREEMPT_RT is a big deal and why it was so much work to get in into the kernel.
406.
by ziofill 2y ago
>"not more than a Tea Spoonful," according to his diary — Franklin poured it onto the agitated water. The oil spread rapidly across the surface, covering "perhaps half an Acre" of the pond and rendering its waters &qu
407.
by ziofill 2y ago
Sure, but that is because there's a lot of context that is tucked away inside the formalism and corresponding syntax.
408.
by ziofill 2y ago
Yeah it’s hard to use everyday language to exactly describe these things… And even more difficult to decide where to draw the line where some hand waving is acceptable.
409.
by ziofill 2y ago
You are exactly right at the end when you say that particles are what makes a detector go click. Let me try to clarify further. The "particleness" is not about localization, but about energy quantization. In fact, a single photon
410.
by ziofill 2y ago
Physicist here. You are right that the mental picture we get when we use the term “particle” is a little ball or something like that. It is unfortunately a confusing name… You need to begin with a field, like the electromagnetic field for i
411.
by ziofill 2y ago
Doesn’t this assume that the probability of a correct answer is iid? It can’t be that simple.
412.
by ziofill 2y ago
Is this effectively some sort of knowledge distillation?
413.
by ziofill 2y ago
I think it just blasts off into the universe
414.
by ziofill 2y ago
A few years ago I learned Python and RL by coding a high frequency trading agent for cryptocurrencies. In backtesting it was phenomenal. While executing though, trading fees, slippage and other factors negated all the advantages. It was qui
415.
by ziofill 2y ago
Unfortunately not ever, because you cannot change the statistics at one side of an entangled pair of systems by acting on the other.
416.
by ziofill 2y ago
A galaxy is so large (e.g. the milky way is ~ 1 billion billions km in diameter!), and bodies are so far from each other that it is far from intuitive that there should even be a magnetic field, and it's even neatly aligned along the a
417.
by ziofill 2y ago
Question for those who do have access: how is it?
418.
by ziofill 2y ago
I'm a theoretical physicist and I swear electrical stuff is so hard to understand! I have a lot of respect for electrical engineers ^^' (and electricians)
419.
by ziofill 2y ago
Is there a place where one can buy pre-built leg parts or even pre-assembled legs? That seems like the most difficult part, then attaching them onto stuff I can do too.
420.
by ziofill 2y ago
the .sick file extension is a nice touch ^^

More ›