winwang

7 ms·

1.
by winwang 5d ago
(Without profiling or looking into this at all) I'd guess this has to with thread creation, inter-core communication/latency, and possibly having to merge results or otherwise interleave operations. SMT is another likely candidate
2.
by winwang 9d ago
imo, partially because it's still not easy (in terms of code -> formal proof). With AI, I've been Lean-ifying a simpler (but non-trivial) algo. Pointing (current) AI at it only goes so far and in fact might go "too far&quo
3.
by winwang 16d ago
"maybe even psychologists" -- unironically true, but maybe I am talking about something different from you. The basis of (higher-level) math seem to mostly be "am I psychologically (emotionally...?) comfortable with accepting
4.
by winwang 20d ago
Every day, we stray closer to Haskell. Dare I say it: good!
5.
by winwang 1mo ago
(not the above poster but) I set up a local workspace for it to read/write from, so that edit is effectively just VSCode/vim. Unfortunately, it seems more verbose when writing out to file.
6.
by winwang 1mo ago
I decided to ask chatgpt for a fermi estimate of cores/datacenter, which you can check for yourself: 1~10 million physical cores. The surprising part of this for me was the "low" power usage (it used 20MW/"datacente
7.
by winwang 1mo ago
Back of the envelope math, if anyone wants to correct my mental model: 2.1B (min core)/week ~ 300M (min core)/day. Assuming ~300k cores (~3k physical server CPUs, seems fine), that's 1k min/day. Seems about right, though
8.
by winwang 1mo ago
I don't GR but I'm fairly certain dipole does not work for gravitational waves (i.e. two objects accelerating straight at each other). Supposedly, there's a specific definition of "gravitational wave" here. You ca
9.
by winwang 1mo ago
Wouldn't discretion "just" be a really good rules engine?
10.
by winwang 1mo ago
I used Claude to generate technical SVGs (which are verifiable correct) because I wasn't going to take the time to draw the grid/arrows/etc.
11.
by winwang 2mo ago
... I don't think the original commenter said anything about mice being a non-model organism. But mice are far from humans, and I, at least, find the distinction rather important.
12.
by winwang 2mo ago
This is a big part of why I am in love in Posthog's design (though personally I have little belief in my own current ability to replicate it)
13.
by winwang 2mo ago
Almost completely disagree. Slightly more expensive, but significant better on a per-prompt basis? For non-trivial projects, the former is a small linear increase, the latter is a (somewhat-)exponential(-ish) cost/time/sanity savi
14.
by winwang 2mo ago
I found 4.6 more amenable than 4.8 to style directions, we'll see how 5.0 does. Super-small-sample-size: I think part of its "Claude-ism" style comes from its propensity to try and "proactively" move the conversatio
15.
by winwang 2mo ago
Several other commenters have disparaging the design seemingly mostly due to its AI-generated nature, or maybe they actually do dislike cyberpunk. Personally, I think being able to have these design languages be easily prototypable is fucki
16.
by winwang 2mo ago
> This isn't a real product. The most disappointing part q_q
17.
by winwang 2mo ago
How bad is it for you? Are you on ultra or xhigh/max? I typically ask it (5.5, now 5.6-sol) to use subagents for specific things anyway. On the Pro 20x plan, I'm seeing like ~1% usage per 20-30 min per session (on max effort), whi
18.
by winwang 2mo ago
I find it interesting that no one here has mentioned the increased (usable) context window 258k -> 353k. That's huge, but I wonder if it means we pay long context (2x) for the ones past 272k still.
19.
by winwang 2mo ago
I would pay $200/mo for Opus 4.8 already. Fable 5 is just a cherry on top (although, without it, Codex 5.5 is the better buy imo).
20.
by winwang 3mo ago
Honestly quite excited to see what can happen here, I think biology has generally had a lack of data science expertise.
21.
by winwang 3mo ago
(Yes, not yet, but...) As a Scala afficionado, "free"/freer runtime performance is very welcome :) Fun read.
22.
by winwang 3mo ago
(Based off 2-3 month-old recollection, take with a grain of salt) I had wanted to use it for my agent "network". A2A didn't fit the use case of "trusted agent, and was bloated due to "what if rogue actor". Of c
23.
by winwang 3mo ago
Not saying you're doing this specifically, but I'd be careful with thinking that "company" in China means the same as "company" in America (or in the West more generally).
24.
by winwang 3mo ago
Love this. I've shifted in the past few months to using highly expressive types in Scala 3 to have types carry more and more compile-time proofs (without macros, though a couple are warranted). Not only does it help with agentic test &
25.
by winwang 3mo ago
Minor note, 2x $/tok is not 2x cost. Personally, I see Fable being significantly more token-efficient than Opus 4.8. Then, there's also the compounding costs of quality.
26.
by winwang 4mo ago
My experience has been that 5.4 is slower than 5.5 (confound: I use >512k max context size for 5.4, though it seems slower even below the normal size)
27.
by winwang 4mo ago
I typically just launch CC with `--model claude-opus-4-6[1m]`, `4-6[1m]` -> `4-8[1m]` works fine. Still 200k max without the `[1m]`.
28.
by winwang 4mo ago
There's the other (orthogonal) possible explanation of using more GPUs for stress-testing before product launch.
29.
by winwang 4mo ago
How else would you write this (marketing copy) exactly? "Its output matches better to its CoT which matches to better to our hidden state decoder according to <insert measure here>; see <insert paper ref>"? ... Actuall
30.
by winwang 4mo ago
Awesome, thanks for posting because I think I hit a possibly-spurious bug in turning Adaptive off when I switched models (4.6 -> 4.8, extra). Tried again, works as intended (I hope). More importantly for me, though, is how CC will respon

More ›