daemonologist

7 ms·

1.
by daemonologist 5d ago
Yeah, the blog post itself loads for me but everything else is really struggling.
2.
by daemonologist 9d ago
You're correct that it needs to run the full model to "verify" a token, but LLM inference benefits from batching - it's much faster to run twice in parallel than sequentially. So the draft model runs ~2-5 tokens ahead,
3.
by daemonologist 13d ago
Huh, it's true (well, technically the price was 100,000 hugging faces); good catch
4.
by daemonologist 13d ago
Model's not available to the public (or even to Simon, I guess) yet.
5.
by daemonologist 14d ago
and furthermore, this is because the drivetrain is ~always on the right side of the bike - if you want to inspect or admire a bicycle you look at the right side, as you might look under the hood of a car. (Why the drivetrain is on the right
6.
by daemonologist 19d ago
I doubt it - AWS hasn't added any non-western models since GLM 5 and MiniMax M2.5 in February, afaik. Might be a deal with OpenAI (GPT 5.4 was the first to be available via Bedrock, in April) or might just be that there isn't a l
7.
by daemonologist 19d ago
The no-Google-calendar-sharing-on-mobile is an absolute classic. I have personally seen multiple clumps of people huddled around a phone trying to figure out how to share a calendar with each other - fruitlessly until I butted in, becaus
8.
by daemonologist 20d ago
Thanks, but I was interested more in the cross-app communication aspect rather than the data formats ( https://codeberg.org/comaps/comaps/wiki/Content-Provider ). I once looked into building an app which pigg
9.
by daemonologist 20d ago
Neat, I was not aware of this mechanism. Is there any way to get map (tile) data from CoMaps into another app?
10.
by daemonologist 26d ago
You could probably score 100% on ARC 3 if you were motivated enough. I find some of the current problems to be kind of like Chess - mechanically simple, and ~solvable, but it's difficult to force myself to think at length about a mono
11.
by daemonologist 1mo ago
The Glacial National Park referenced by GP is in Montana; you were perhaps at Glacier Bay National Park. Regardless, some beautiful photos, thanks for sharing
12.
by daemonologist 1mo ago
Bedrock seems to have stopped adding new open-weights models, and mostly only has Anthropic and OpenAI stuff now. You can get Qwen 3.8 directly from Alibaba: https://www.qwencloud.com (proprietary variant) or from DigitalOcean
13.
by daemonologist 1mo ago
On the first point, I think it's a search of boat names , not boats - they're deduplicated.
14.
by daemonologist 1mo ago
Also (mentioned by Tom), this elastic-powered supersonic trebuchet is an excellent video: https://www.youtube.com/watch?v=gdXOS-B0Bus
15.
by daemonologist 1mo ago
Most phones won't boot without the battery. Obvious workaround is to connect a power supply (and large cap) directly to the BMS, but in my experience most phones will shut themselves down when they think they've drawn more than t
16.
by daemonologist 1mo ago
I don't think it's been shrunk, I think there just aren't enough high karma posts to fill the default (filtered) page. Old posts fall off due to karma decay and low quality posts get flagged or hidden by mods. I've see
17.
by daemonologist 1mo ago
It's here as well: https://api-docs.deepseek.com/quick_start/pricing/
18.
by daemonologist 1mo ago
With an unlimited budget, Kimi K3 (which is quite comparable to this Qwen Max imo). With a normal budget/a PC you might already have, probably Qwen 3.6 27B.
19.
by daemonologist 1mo ago
Going from 5.5 Instant (which was noticeably bad) to 5.6 Luna is a big jump as well. OpenAI is probably the most prominent among the general public - an advantage in some respects but they're giving away a lot of free inference and
20.
by daemonologist 1mo ago
Pricing: https://dev.meta.ai/docs/pricing-rate-limits Interesting that they have separate API pricing for "we can train on your data" (whereas iirc most of the big players either make that distinction only be
21.
by daemonologist 1mo ago
Service area: https://support.google.com/waymo/answer/9059119?hl=en#DAL
22.
by daemonologist 2mo ago
The linked qwencloud page has pricing; it's $2/6. https://www.qwencloud.com/models/qwen3.8-max
23.
by daemonologist 2mo ago
It seems to only be available in Kimi Code, via subscription, no there's no API pricing. The linked page says it consumes about half as much quota as the 1M version though.
24.
by daemonologist 2mo ago
Btw if you do need a permissively licensed model for this task, there are lots out there, e.g.: https://github.com/KupynOrest/s3od https://github.com/Tennine2077/PDFNet https://github.c
25.
by daemonologist 2mo ago
Agreed, and even SSR always felt like cheating imo. I wanna see someone build a primary radar on an esp 32.
26.
by daemonologist 2mo ago
My read is that they're announcing a plan to build non-folding wings to test the geometry they would eventually want with folding wings. Which makes sense but isn't very exciting.
27.
by daemonologist 2mo ago
It might be that a lot of the cost doesn't show up in the data because it comes out of your paycheck before it's "yours," in the form of your employer's contribution to health insurance premiums. (Anecdotally this
28.
by daemonologist 2mo ago
Parakeet is way faster (on Nvidia hardware) but not quite as accurate in my experience.
29.
by daemonologist 2mo ago
There was a fad a while back of building insanely long prompts - tens of thousands of tokens - including having models write prompts for themselves. I always thought it was counterproductive, especially if you're going to use the prom
30.
by daemonologist 2mo ago
I think they're saying that you should build a small business which nets you 85k, not find an employer that underpays you. (Whether this makes you more resistant to being "fired" is still up for debate of course.)

More ›