fancy_pantser

6 ms·

31.
by fancy_pantser 9mo ago
When OpenAI announced the Triton language, I was worried I'd be confused one day while reading something because of Nvidia's open-source Triton inference server. I made it quite a long time, but it finally happened today! I was so
32.
by fancy_pantser 10mo ago
Mods usually apply [Dupe] to later submissions if a recent (last year or so) one had a fair amount of discussion.
33.
by fancy_pantser 10mo ago
Student: Look, there’s hundred dollar bill on the ground! Economist: No there isn’t. If there were, someone would have picked it up already. To wit, it's dangerous to assume the value of this idea based on the lack of public implementa
34.
by fancy_pantser 10mo ago
> The team then created VeriGen, the first specialized AI model trained solely to generate Verilog code. Perhaps it's the first open one. I was an eng manager at a hyperscaler helping one of our clients, a large semiconductor design
35.
by fancy_pantser 10mo ago
> Prices are entirely hidden Recent legal changes have made pricing more transparent. In 2020, the federal government issued the "transparency in coverage" final rule under the Federal No Surprises Act. This limited the expense
36.
by fancy_pantser 10mo ago
> Attention is quadratic Exactly. Standard Multi-Head Attention uses a matrix that grows to 4B parameters for a 64K sequence as a starting place. FlashAttention v2 helps slightly, but as you grow to 128K context length, you still need ov
37.
by fancy_pantser 10mo ago
Software like i2 Analyst's Notebook.
38.
by fancy_pantser 11mo ago
Have you given using MCPs to provide documentation and examples a shot? I always have to bring in docs since I don't work in Python and TS+React (which it seems more capable at) and force it to review those in addition to any specifica
39.
by fancy_pantser 1y ago
Would love to see this connect to a smartphone running a matching app, doing the inferencing on the phone so you wouldn't have to bill for increments. It will be a few more years until that can be done in a low-enough latency way (impr
40.
by fancy_pantser 1y ago
I scratched a similar itch and found local LLMs plus Whisper worked really well to listen in and "DJ" a soundtrack while playing tabletop RPGs with a group. If you want to check it out: https://github.com/sean-publ
41.
by fancy_pantser 1y ago
Claimify from MS research aims in this direction. There's a paper and video explainer from a few months ago. https://www.microsoft.com/en-us/research/blog/claimify-extra...
42.
by fancy_pantser 1y ago
There are utilities to help, waybackpack comes to mind, but I haven't looked in a while. https://github.com/jsvine/waybackpack
43.
by fancy_pantser 1y ago
As OP, I think I'm supposed to arrange for the card to be collected... oh hey you're Kirk! It's Sean formerly of Amino as well :)
44.
Intel CEO Letter to Employees(morethanmoore.substack.com)
292 points by fancy_pantser 1y ago|554 comments
45.
by fancy_pantser 1y ago
Are you looking for a tool like Barman?
46.
by fancy_pantser 1y ago
I've been using the unofficial builds via scoop for the last two months. It's working great so far. I use it on a Macbook as well and I haven't found any features that are missing or buggier on Win11. Really enjoying the new
47.
by fancy_pantser 1y ago
Maybe their privacy pass is useful then? https://help.kagi.com/kagi/privacy/privacy-pass.html
48.
by fancy_pantser 1y ago
Been having a great time with postgresml for this exact kind of thing. If you don't need a complex DAG but have a simple pipeline or work queue that can be easily represented in postgres anyway, it's very straightforward to work w
49.
by fancy_pantser 2y ago
You can use the number with your local area code just about anywhere at the pump to get a gas discount as well (a common loyalty reward program benefit).
50.
by fancy_pantser 2y ago
A very similar card game played by deciding when to stop flipping cards from a deck where red is $1 and black is −$1 as described in Timothy Falcon’s quantitative-finance interview book (problem #14). Gwern describes it and also writes code
51.
by fancy_pantser 2y ago
have you checked daily.co pricing for audio-only? they do SIP
52.
by fancy_pantser 2y ago
> specifying and visualizing DAGs Do you mean like Airflow or Pachyderm? I am also very interested in new tooling in this space that has these features.
53.
by fancy_pantser 2y ago
Recipe Filter: https://chromewebstore.google.com/detail/recipe-filter/ahlcd...
54.
by fancy_pantser 2y ago
here ya go ;) https://news.ycombinator.com/item?id=24665282
55.
by fancy_pantser 2y ago
apple glazing:apple::navel gazing:orange
56.
by fancy_pantser 2y ago
Look into PocketBook readers. They have a full lineup with color and grayscale in various sizes. They run linux, you can ssh into it and install koreader, etc. They have a good privacy story and little vendor lock-in. The experience is a lo
57.
by fancy_pantser 2y ago
I wondered this myself and found some. I've played: When Rivers Were Trails, Thunderbird Strike, Dialect, Zulu Dawn, This Land Is My Land. There are a lot of options in boardgames (maybe more than with video games?), such as Burn the F
58.
by fancy_pantser 2y ago
I see Command-R+ but not Command-R marked for tool use. The model is geared for it, much easier to fit on commodity hardware like 4090s, and Ollama's own description for it even includes tool use. I think it's just not labeled for
59.
by fancy_pantser 2y ago
No fine-tuning is necessary. You can use something reasonably good at RAG that's small enough to run locally like the Command-R model run by Ollama and a small embedding model like Nomic. There are dozens of simple interfaces that will
60.
by fancy_pantser 2y ago
Perhaps it was John Koza? http://www.genetic-programming.com/johnkoza.html

More ›