Palmik

8 ms·

31.
by Palmik 4mo ago
By requiring various forms of identification to use social media, it will be harder to criticize your leaders anonymously without fear of retribution.
32.
by Palmik 4mo ago
The company representative said that they report all users that use Graphene OS, without any additional qualifiers . Presumably after they've already uploaded their personal details. That's the egregious part.
33.
by Palmik 5mo ago
DeepSeek V4's KV cache is very efficient due to its heavily compressed and sparse attention architecture. DeepSeek V3.2 which uses DSA only (sparse attention, but without compression from HCA and CSA) is a smaller model but uses 10x mo
34.
by Palmik 5mo ago
I really hope Huawei ramps up Ascend production and DeepSeek open sources their optimized inference engine (they already open source a lot of their kernels -- kudos to them). This could shake things up.
35.
by Palmik 5mo ago
There are several things at play: Inference stack efficiency: Many of these providers take off the shelf sglang / vllm / trtllm and hope for the best. Meanwhile DeepSeek team is known for pushing the boundary of optimizations. Now
36.
House Committees Probe Cursor Parent, Airbnb over Chinese AI(semafor.com)
2 points by Palmik 5mo ago|0 comments
37.
by Palmik 5mo ago
Why was the title changed from "DeepSeek V4—almost on the frontier, a fraction of the price" to "DeepSeek V4—almost on the frontier"?
38.
by Palmik 6mo ago
Surely art also exists in textual realm.
39.
Anthropic Claude Code HERMES.md billing flaw(consumerrights.wiki)
1 points by Palmik 6mo ago|0 comments
40.
by Palmik 6mo ago
I don't think "friendly" and "publishing benchmarks" are at odds with each other. Model makers (both open and closed weight) typically publish benchmarks against other models and when they do not, people rightfully
41.
by Palmik 6mo ago
Similar article for vLLM: https://vllm-website-pdzeaspbm-inferact-inc.vercel.app/blog/... Bechmarks from InferenceX (they do not have apples-to-apples setups to compare the different engines for whatever reason): http
42.
DeepSeek V4 in vLLM: Efficient Long-Context Attention(vllm-website-pdzeaspbm-inferact-inc.vercel.app)
2 points by Palmik 6mo ago|0 comments
43.
by Palmik 6mo ago
Or there will be DSv4.1/2/3 ;)
44.
by Palmik 6mo ago
Misleading conclusion. This model is 8 times cheaper than Gemini for 1K images. Gemini is extremely overpriced. 1K image with Gemini is roughly $0.08 and only $0.01 with GPT Image.
45.
by Palmik 6mo ago
Did you enable thinking for your experiment? Are you sure you were on the 2.0 rather than 1.5 version?
46.
by Palmik 6mo ago
I do not think this is a good prompt or useful benchmark, but nonetheless, it seems to work better for me: https://chatgpt.com/share/69e88a94-ded8-8395-b5dc-abceb2f44d...
47.
NSA is using Anthropic's Mythos despite blacklist(axios.com)
485 points by Palmik 6mo ago|348 comments
48.
by Palmik 6mo ago
Could it be made even faster using some of the ideas from https://github.com/zerobootdev/zeroboot ?
49.
by Palmik 6mo ago
Official announcement: https://www.thunderbolt.io/announcing-thunderbolt
50.
Mozilla Announces "Thunderbolt" as an Open-Source, Enterprise AI Client(phoronix.com)
25 points by Palmik 6mo ago|11 comments
51.
Google, Pentagon discuss classified AI deal, the Information reports(reuters.com)
6 points by Palmik 6mo ago|2 comments
52.
OpenAI, Anthropic, Google Unite to Combat Model Copying in China(bloomberg.com)
3 points by Palmik 6mo ago|0 comments
53.
Speaking of Voxtral(mistral.ai)
19 points by Palmik 7mo ago|1 comments
54.
by Palmik 7mo ago
My email does mention it clearly: > Again, your organization's Copilot interaction data is not included in model training under this new policy, but we are excited for you to enjoy the product improvements it will unlock.
55.
by Palmik 7mo ago
Great work! There is maybe some bug. When you click on one of the 4 "opposing" countries (e.g. Czech Republic, Poland), it scrolls down and then shows that majority of the representatives from the country actually support it. Is t
56.
Nvidia Nemotron Coalition of Leading AI Labs to Advance Open Frontier Models(nvidianews.nvidia.com)
5 points by Palmik 7mo ago|0 comments
57.
by Palmik 7mo ago
What are your thoughts on this? https://www.anthropic.com/news/where-stand-department-war I am honestly unclear on the reasoning of people who flock from OpenAI to Anthropic, and doubly so of those who are not US citiz
58.
Sam Altman AMA on DoD Collaboration(twitter.com)
20 points by Palmik 7mo ago|3 comments
59.
by Palmik 8mo ago
Except most of the world's population, and in fact large fraction of the engineers and scientists working on these things, are not US citizens.
60.
Palantir Sues Magazine for Reporting That the Government Didn't Want Palantir(techdirt.com)
13 points by Palmik 8mo ago|0 comments

More ›