Charlie Loyd @vruba@everything.happens.horse

This hacky lightweight 4× super-resolver has learned that I want things to look sharp but not too sharp, so it sort of scatters edges around the image like spice in a sauce.

Turns out the derivative of sqrt(x) does some colorful things around x=0. I prefer not to pry, but it seems not to want to be bothered around there. I’m going to leave it be.

Spent several days working out a cool optimization for a denoising diffusion model (addressing the noise continuously, modeling its full distribution at every step of t instead of working with quantized cumulative sums) and have now discovered why people don’t do it this way.

I’m not trying to forgive this commute, because I think it’s bad. But I think it’s a teachable moment. We recognize it as ludicrous, but it’s within what we accept in other contexts for no good reason.

Commuting from LA to Berkeley by plane is horrifying from an emissions point of view and a chilling indictment of our transit infrastructure, our housing crisis, our remote learning culture, etc., etc.

Granting all this, if you do the math, what he’s doing is well within the emissions range of the normal car commutes of millions of Americans. There are many, many people driving cars to work in the ordinary way who are doing more climate harm.

https://www.berkeleyside.org/2023/05/15/uc-berkeley-la-plane-commute-grad-student

Getting my daily steps in (going to the burrito truck a couple blocks away).

Another day, another long phone call about EPSG codes.

Good month to be an aerosol. https://botsin.space/@dscovr_epic/110341110053167051

I kid, but if you broke down how we do it now at a similar level of detail (which is subjective, granted), I think the diagram would be about as complex.

They don’t want you to know, but this is how it still works most of the time. https://vis.social/@maarten/110033143162477225

Nodding every time the loss goes down, frowning every time it goes up, pursing my lips and making a note if it’s the same for two batches in a row.

Hard to believe it’s elk rendang season again.

I appreciate a discussion among a bunch of people who are all, so to speak, capable of flying space shuttles, yet understand why not every interface should look like the space shuttle cockpit. https://mstdn.social/@kissane/110300512767913104

This paper is 5 years old but I just happened across it and have been talking it up to like four different friend groups because it’s so fun. I particularly enjoyed the retrograde klimarübe (you know, the rutabaga that represents an idealized continent for climate delineation purposes … you know).

Mikolajewicz et al. (2018): “The climate of a retrograde rotating Earth” – https://esd.copernicus.org/articles/9/1191/2018/

I work in the field of remote sensing but the only thing the name “remote sensing” has going for it is that it sounds like something that a gutta-percha magnate established an as-yet–unclaimed £5,000 prize to prove the existence of in 1926.