Caleb Miller

1.5K posts

Caleb Miller banner
Caleb Miller

Caleb Miller

@cmiller__

Graphics programmer and general UI guy. Addicted to JavaScript and caffeine. Pretty good at plugging in USB Type-A cables.

Katılım Aralık 2015
322 Takip Edilen353 Takipçiler
Caleb Miller
Caleb Miller@cmiller__·
@ryanflorence I was so excited for pointer events to land in browsers (crazy that was well over 10 years ago now). I remember there being implementation pushback at the time over real differences between input modalities. I'm glad we got them, but yeah, there are still quirks to handle...
English
0
0
1
182
Ryan Florence
Ryan Florence@ryanflorence·
Man I was pumped about pointer events and then I remembered that pointers are how you scroll on touch and some pointerdown stuff felt bad on touch screens but great on mouse so many of my plans fell apart this afternoon UX details man, I'm sweatin' 'em
English
5
0
22
3.5K
Caleb Miller
Caleb Miller@cmiller__·
So there's no way to disconnect natural scrolling between the mouse scroll wheel and the trackpad? These two settings under two different sections are inextricably linked? This only makes sense if you're using a magic mouse, but I don't know anyone who does.
Caleb Miller tweet media
English
0
0
0
35
Sindre Sorhus
Sindre Sorhus@sindresorhus·
Having popular free apps requires extreme amounts of self restraint
English
7
2
94
8.7K
Caleb Miller
Caleb Miller@cmiller__·
Sped up (12x) descent footage from a recent flight. The volumetric rendering looks so good!
English
0
1
0
87
Liam Egan
Liam Egan@liamegan·
@cmiller__ 100% The urge to gatekeep only rears when the ignorant and overly-confident hyperbole starts.
English
1
0
1
10
Caleb Miller
Caleb Miller@cmiller__·
I guess connecting a grid of points with springs counts as "hyper-realism" now. The result speaks for itself... my condolences to everyone who actually works on cloth simulation.
Guillermo Rauch@rauchg

Created a flag on the moon based with hyper-realistic physics with @v0: v0-moon-flag.vercel.app. The flag is simulated as a grid of particles connected by springs (cloth simulation). It's influenced by gravity, wind, air resistance. The initial approach v0 one-shotted was quite simplistic, but giving it references to more sophisticated approaches in C++ and Rust quickly improved it. It was mind-blowing it to see it deeply understand the physics and port new ideas across languages. It simulated a real sun for lighting, and procedurally generated moon-like terrain with craters. I found a sick free texture on Reddit that I uploaded and it layered on top. Another epic moment was watching it improve performance. It seamlessly moved the physics simulation to a Web Worker. (Kinda cool to see given we just improved support for them in Turbopack 😉)

English
1
1
6
249
Caleb Miller
Caleb Miller@cmiller__·
@liamegan I bet. To be fair though, technically accurate over-explaining at least shows some level of understanding, and teaching vocab is great. It's the ignorant hyperbole that gets to me haha
English
1
0
1
14
Liam Egan
Liam Egan@liamegan·
@cmiller__ Haha. Having worked with many people who have found their entry into graphics programming via AI, this sounds pretty on-point. "Oh you derive the velocity of the particles each frame from their current and prev position for greater accuracy? You mean Verlet integration?"
English
1
0
1
21
Caleb Miller
Caleb Miller@cmiller__·
@liamegan "simulated a real sun" – oh you added a directional light?
English
1
0
1
16
Caleb Miller
Caleb Miller@cmiller__·
@liamegan Lol, great screenshot. There's clipping even when the flag is mostly straightened out. I wouldn't normally be so snarky over someone's excitement for a fun toy they built, but the hype-speak is out of control here...
Caleb Miller tweet media
English
1
0
1
22
Caleb Miller
Caleb Miller@cmiller__·
@wwwtyro Very cool. Molecular models seem like a great fit for path tracing, and it's awesome to see it being done proper.
English
0
0
1
22
Rye Terrell
Rye Terrell@wwwtyro·
@cmiller__ Thanks! Yeah, more than a decade ago I worked on computational chemistry stuff, and the gamedev in me couldn't accept the visualization tools available so every once in a while I throw a new one together. 😁
English
1
0
1
33
Caleb Miller
Caleb Miller@cmiller__·
@wwwtyro (btw, I was traveling the other week and came across Guns Blazing via an old bookmark, and lost the next 30 minutes 😅)
English
0
0
1
19
Caleb Miller
Caleb Miller@cmiller__·
@wwwtyro What the heck man, this is gorgeous. Original was cool but this is outstanding. Is this just scratching a personal itch?
English
2
0
1
36
Peter Schmidt-Nielsen
Peter Schmidt-Nielsen@ptrschmdtnlsn·
@Alientrap I really like this! I want to see blocks rejoin at staggered times. I think it should work like: if remaining_time - bonus_for_being_really_aligned < 0 { rejoin } so if a block is about to rejoin anyway, and happens to be in a great spot right now it rejoins a bit early.
English
1
0
20
2K
Lee Vermeulen
Lee Vermeulen@Alientrap·
Tried out a voxel physics idea where it breaks into individual cube bodies, then adds back to the voxel grid after
English
77
38
2.8K
255.4K
Caleb Miller
Caleb Miller@cmiller__·
Went through DC Metro Center yesterday, and the national guard was stationed above the escalators. Their demeanor was calm and relaxed, chatting with each other, but also certainly paying attention to surroundings. Honestly, it's the safest I've ever felt walking through there.
English
0
0
0
34
Caleb Miller
Caleb Miller@cmiller__·
Devs writing common web code are surprised that people still write any code by hand, or even look at most of the code they're given. Limited perspective. These models will outright lie to you and fabricate all types of bs. You need to have trust issues with LLMs.
Sebastian Aaltonen@SebAaltonen

LLM explaining why it added yet another branch in perf critical inner loop: > "defensive fallback for callers that might render an unsynchronized scene" There's no such thing as "unsynchronized scene".

English
0
0
0
49
Caleb Miller
Caleb Miller@cmiller__·
I work on a compiler of sorts. One stage is ~600 lines, with a 4,200-line test file (95% hand-written). It's big, but focused and well-contained. The pros outweigh the cons. I don't get the AI 'rule' that files must be <500 lines – it'd make a mess out of real cases like this.
English
0
0
0
51
Caleb Miller
Caleb Miller@cmiller__·
@devongovett Awesome that this works at build time. Seems like there's no true replacement to what you did here unless CSS supported this somehow.
English
1
0
0
912
Devon Govett
Devon Govett@devongovett·
Seems like a good opportunity to share a trick I used on the React Aria website to make headings scale to fit the viewport on mobile, even with SSR. I used fontkit, a library I wrote 10 years ago to parse fonts and layout glyphs, to measure the width of the heading at build time in the "em" unit. width = font.layout('DateRangePicker').advanceWidth / font.unitsPerEm This is a size independent measurement using the font's internal metrics. When multiplied by any font size, you get the pixel width of the text. To calculate the font size, set this as a CSS variable and divide the target width (100vw minus padding). Clamp to make sure it doesn't get too big or small. font-size: clamp(35px, (100vw - 32px) / var(--text-width), 55px) And tada! You have perfectly scaling text no matter the viewport size. 🎉
English
10
32
594
47.7K
Caleb Miller
Caleb Miller@cmiller__·
@gfodor @Duderichy also it's looking more likely that LLMs will help medicine advance through cheap personalization, making the "one-off" argument even more silly
English
0
0
2
26
gfodor.id
gfodor.id@gfodor·
@Duderichy if you are able to do something once there is always at least some reason to think you can do it again you're just assuming that the only variable we are looking at is the specific intervention, not the process
English
2
0
10
359
the Rich
the Rich@Duderichy·
the thing about curing cancer in one person is just that you’ve cured cancer in one person and there’s no reason to believe it’ll work nearly as well in others
Massimo@Rainmaker1973

A single dose of a new cancer drug made a brain tumor almost disappear – in just five days. Doctors at Massachusetts General Hospital reported “dramatic and rapid” tumor regression in the first patients treated with a next-generation form of CAR T-cell therapy for glioblastoma, one of the most aggressive brain cancers known. The therapy, called CARv3-TEAM-E, was developed to overcome a major hurdle in treating solid tumors: their ability to hide from the immune system. The personalized treatment reprograms a patient’s immune cells to attack the tumor, and in one extraordinary case, nearly eliminated the cancer within just five days. This novel therapy is designed to target multiple features of the tumor at once, a strategy that may help overcome the common challenge of treatment resistance in solid tumors like glioblastoma. Although the tumors eventually returned, the early outcomes were described as unprecedented. One patient saw a 60% reduction in tumor size that lasted for half a year—an impressive result in a cancer known for its aggressiveness. The trial’s success marks a major step forward for immunotherapy in brain cancer and raises new hopes for long-term control or even a cure. Researchers are now working to refine the treatment and extend its effects, with the ultimate goal of turning a once-terminal diagnosis into a survivable condition.

English
12
1
39
5.4K
Caleb Miller
Caleb Miller@cmiller__·
It seems like Apple has something against these back/forward buttons. They don't work in Finder, Safari, or Apple Music. They work as expected in Chrome. These are common buttons on many popular mice. What gives? Even Windows Explorer supports them...
Caleb Miller tweet media
English
0
0
0
57