EnPissant

7 ms·

1. 91.▲
   by EnPissant 9mo ago

   touche.
2. 92.▲
   by EnPissant 9mo ago

   > People have a stark reaction to the visual of a guy being 200 meters above the ground but the reality is if you're a circus acrobat and you're 10 meters up the air you're also likely dead or maimed if something goes wron
3. 93.▲
   by EnPissant 9mo ago

   You need to be specific! There are 2 lawyers with completely opposite hand sizes.
4. 94.▲
   by EnPissant 9mo ago

   They say they do not compress the user messages, but yeah, it's purpose is to do very lossy compression of everything else. I'd expect it to be small.
5. 95.▲
   by EnPissant 9mo ago

   Are you sure? For reasoning, encrypted_content is for sure actually encrypted.
6. 96.▲
   by EnPissant 9mo ago

   I'm guessing that's ~26 decode tokens/s for 2-bit or 3-bit quantized Minimax-m2.1 at 0 context, and it only gets worse as the context grows. I'm also sure your prefill is slow enough to make the model mostly unusable, ev
7. 97.▲
   by EnPissant 9mo ago

   Maybe it's a model fine tuned for compaction?
8. 98.▲
   by EnPissant 9mo ago

   > I think the delta may be an overloaded use of "turn"? The Responses API does preserve reasoning across multiple "agent turns", but doesn't appear to across multiple "user turns" (as of November, at le
9. 99.▲
   by EnPissant 9mo ago

   Thanks. That's really interesting. That documentation certainly does say that reasoning from previous turns are dropped (a turn being an agentic loop between user messages), even if you include the encrypted content for them in the API
10. 100.▲
    by EnPissant 9mo ago

    Thanks. I consider myself an insignificant or contemptible person or thing. Also, everyone loves chess puns.
11. 101.▲
    by EnPissant 9mo ago

    I don't think this is true. I'm pretty sure that Codex uses reasoning.encrypted_content=true and store=false with the responses API. reasoning.encrypted_content=true - The server will return all the reasoning tokens in an encrypte
12. 102.▲
    by EnPissant 9mo ago

    Have you tried lower reasoning levels?
13. 103.▲
    by EnPissant 9mo ago

    Do you own a dog house?
14. 104.▲
    by EnPissant 9mo ago

    Yeah, they are really impressive for 4" drivers.
15. 105.▲
    by EnPissant 9mo ago

    I currently have the 8020D + 7040A. It's pretty much perfect sound with room correction applied, but the 7040A is kind of big and ugly. I'd be willing to give up some low-end to simplify the setup. Also the sub performs the crosso
16. 106.▲
    by EnPissant 9mo ago

    Not OP, but I have 2 8020D + a subwoofer. If I had to do it all over again I would have 2 8030C and no subwoofer.
17. 107.▲
    by EnPissant 9mo ago

    You are depicting this as some dramatic shift, but the Cato institute has always been anti-Trump.
18. 108.▲
    by EnPissant 9mo ago

    It depends on your OS. Linux and Windows are significantly better for both 1440p and 4k monitors. Both Linux and Windows have subpixel rendering and configurable font hinting for 1440p. And they both have fractional scaling UIs for 4k. macO
19. 109.▲
    by EnPissant 9mo ago

    I used to have the same complaint, and recently swapped to 4k monitors. I thought that would solve my zed font problems, but text presentation is still bad. In zed, it feels like there is significantly more spacing between each line compa
20. 110.▲
    by EnPissant 9mo ago

    I know what a post history is. Obviously I was not asking someone to link to my profile.
21. 111.▲
    by EnPissant 9mo ago

    My post history?
22. 112.▲
    by EnPissant 9mo ago

    What exactly is my post history doing?
23. 113.▲
    by EnPissant 9mo ago

    I’ve seen this claim a lot, but I’m skeptical. Has anyone actually published benchmarks showing a big speedup from using the NPU for prefill? AMD’s own marketing numbers say the NPU is about 50 TOPS out of 126 TOPS total compute for the pla
24. 114.▲
    by EnPissant 9mo ago

    Then again, I have a RTX 5090 + 96GB DDR5-6000 that crushes the spark on prompt processing of gpt-oss-120b (something like 2-3x faster), while token generation is pretty close. The cost I paid was ~$3200 for the entire computer. With the cu
25. 115.▲
    by EnPissant 9mo ago

    I don’t think your problem is RGB instead of BGR. That’s just the compositor’s work area and your monitor never sees it (it includes an alpha channel). Have you tried KDE Plasma? It sounds like KWin uses 10-bit planes by default when availa
26. 116.▲
    by EnPissant 9mo ago

    Was it an 8k cable? Are you on wayland?
27. 117.▲
    by EnPissant 9mo ago

    I asked claude to investigate: Your Display Configuration Both monitors are outputting 10-bit color using the ABGR2101010 pixel format. | Monitor | Connector | Format | Color Depth | HDR | Colorspace |
28. 118.▲
    by EnPissant 9mo ago

    It's not obvious how to interpret the output. I pasted it into chatgpt and it thinks I am using "Format: ABGR2101010" for both monitors (only 1 has HDR on) so I don't trust it. EDIT: See my sibling comment.
29. 119.▲
    by EnPissant 9mo ago

    I don't think this is true. I can go into my display settings in kde plasma and enable HDR and configure the brightness. I have a nvidia blackwell card.
30. 120.▲
    by EnPissant 10mo ago

    I've never seen any significant difference in linear vs affine types. To me it just seems like Rust has Linear types, and the compiler just inserts some code to destroy your values for you if you don't do it yourself. I guess the