mtone

6 ms·

1.
by mtone 2mo ago
Firefox's AI Chatbot feature only proposes one open model provider (Mistral) and zero local option. They're not walking the talk.
2.
by mtone 3mo ago
It was trained from scratch by Cohere. They're the only Canadian AI lab - I'm glad they're releasing open weights and I wish them luck catching up!
3.
by mtone 3mo ago
I am using the `voipmonitor/vllm:lucifer` docker from the RTX6K discord community discussed at the same link the other commenter posted. It is based around this PR https://github.com/vllm-project/vllm/pull&#x2
4.
by mtone 3mo ago
Here's a DeepSeek-V4-Flash benchmark on 2X RTX Pro 6000: - Prefill: ~10K tok/s - Decode: 190 | 375 | 980 tok/s (for 1 | 4 | 16 concurrent requests) - GPU power draw during benchmark: Average: 585W | Max: 849W | Limit:
5.
by mtone 4mo ago
You're forgetting a critical factor: concurrency. If a given hardware serves a single request at 150 tokens/s, it can also serve 20-30 requests at 100 tokens/s. Suddenly your $5K becomes $100K/month, enough to recoup the
6.
by mtone 6mo ago
Thanks! There's a heading control to include rotation in link: https://earth.google.com/web/@3.63731074,-23.1618975,-2690.7...
7.
by mtone 7mo ago
> Does privacy of Netflix ratings matter? The issue is not “Does the average Netflix subscriber care about the privacy of his movie viewing history?,” but “Are there any Netflix subscribers whose privacy can be compromised by analyzing
8.
by mtone 7mo ago
For this type of repetitive application I think it's common to "fine-tune" a model trained on your business problem to reach higher quality/reliability metrics. That might not be possible with this chip.
9.
by mtone 9mo ago
Good question. That was for my MS account/licenses and some Azure stuff. I use Google Authenticator for most things. Thanks for the link, I'll take a look. I might just move it to a secondary device first.
10.
by mtone 9mo ago
Just looked - Microsoft Authenticator doesn't appear to work. I might be able to get off of it but it will take some prep. My banks are supported so that's good.
11.
by mtone 9mo ago
> if you could exclude all of the R&D and training costs LLMs have a short shelf-life. They don't know anything past the day they're trained. It's possible to feed or fine-tune them a bit of updated data but its world
12.
by mtone 11mo ago
Recently llama.cpp made a few common parameters default (-ngl 999, -fa on) so it got simpler: --model and --context-size and --jinja generally does it to start. We end up fiddling with other parameters because it provides better performance
13.
by mtone 1y ago
Do you really need a H200 for this? Seems like something a consumer GPU could do. Smaller models might be ideal [0] as they don't require extensive world knowledge and are much more cost efficient/faster. Why can't you build
14.
by mtone 1y ago
Related discussion on Anubis: https://news.ycombinator.com/item?id=43427679
15.
by mtone 2y ago
I use Readdle Documents to sync PDF folders with my server PC via FTP. Free version supports PDF highlighting & simple annotations, basic file management, and automatically syncs back everything.
16.
by mtone 3y ago
Assuming it's correct, I think this answer explains it well: https://stackoverflow.com/a/31996121/283879 Basically, yes it may write snapshots per-file (never per-commit) locally but there is a separate routi
17.
by mtone 3y ago
There was talk about introducing type syntax as valid but ignored in the JS language, making TS valid JS. It would take forever to become mainstream but if node and major browsers started to support this tomorrow, along with ESM modules we
18.
by mtone 3y ago
They're turing-complete and modular so it's not really about what they can or cannot do. Testability, tooling and the open-source ecosystem and either bad or non-existent. Writing PL/SQL is the worst environment I've wor
19.
by mtone 3y ago
Just upgraded to Windows 11 (for its HDR features) and weather is now part of the "Widgets" bombarding me with ads and poor news sources. I genuinely tried to customize my "feed" but it's all junk -no reputable so
20.
by mtone 4y ago
In a monorepo with project references where each symlinked (not paths) package has a src/ and output lib/, I've always had issues with VsCode incorrectly proposing to import from src/ over lib/ from other packages.
21.
by mtone 4y ago
We're not on 18 yet but things are clear, thanks!
22.
by mtone 4y ago
I notice that ReactDOM.render is also deprecated. In our embedded/plugin component scenario where we are given a <div> to load in, it appears we should replace our current pattern ReactDOM.render(React.createElement(... with c
23.
by mtone 4y ago
> it has by far the best vertical tab implementation of all browsers (someone please copy this) TIL. I think it's still far behind TreeStyleTabs (TST) on Firefox. Missing for me: - No nested groups. I know it's called vertical
24.
by mtone 4y ago
I close a door or put a box in its way. No app.
25.
by mtone 4y ago
When I check my recommended feed it's all stuff very closely related stuff to my subscriptions and recent views, and some useless shorts. It's just an echo chamber at this point. If I'm looking for potentially interesting con
26.
by mtone 4y ago
Both bigger than bog-standard CRUD apps, but there is: https://github.com/dotnet-architecture/eShopOnWeb - Demo app from MS https://github.com/jellyfin - A plex alternative
27.
by mtone 4y ago
Recently Akka [0], a 12K star project. [0]: https://github.com/akka/akka/pull/31561
28.
by mtone 4y ago
Thanks. I really like the presentation/animation style.
29.
by mtone 4y ago
I add stuff in FF Group Speed Dial [0] under a variety of headings. I try to stick to ~100-150 links in there, deleting entire headings once past their usefulness. They're no good as a long term knowledge base however -- search is too
30.
by mtone 4y ago
> and probably seems [..] foreign Yes, it's French for "free" after all.

More ›