behnamoh

37 ms·

1. 871.▲
   by behnamoh 2y ago

   > Bing CoPilot Tangent: it annoys me so much that there's a persistent useless tiny horizontal scroll on the Bing page! when scrolling down, it rattles the page left and right.
2. 872.▲
   by behnamoh 2y ago

   > If Google can somehow undo literally everything about SEO, then it can become useful again So, Google should De-Google itself?
3. 873.▲
   by behnamoh 2y ago

   Nah, most people have moved on to Gemma, Qwen, Mistral Small/Nemo variants.
4. 874.▲
   by behnamoh 2y ago

   is sglang an LLM engine or does it use vLLM/llama.cpp under the hood? and while we're at it, has anyone done a comparison of LLM engines? I've also heard of Mistral.rs, LLM MLC, and obviously HF transformers library and its k
5. 875.▲
   by behnamoh 2y ago

   > My top choice for this kind of thing in 2025 is GitHub, using GitHub Pages. It's free for public repositories and I haven't seen GitHub break a working URL that they have hosted in the 17+ years since they first launched. Doe
6. 876.▲
   Gemini Pretraining [slides] [pdf](vladfeinberg.com)

   2 points by behnamoh 2y ago|0 comments
7. 877.▲
   by behnamoh 2y ago

   > What made you choose memgraph? ... And then Memgraph showed me the cost. That kind of sold me for time for us to be able to do that. It's an ad post about memgraph.
8. 878.▲
   by behnamoh 2y ago

   Often times, I don't find all the icons I need in one icon package. I wish there was a way to find icons of various sources (and thus, various styles) and make them more consistent to be used in a program.
9. 879.▲
   by behnamoh 2y ago

   macOS keyboard shortcuts situation is weird. it shines where Windows is lacking, and falls behind where Windows is ahead. Example: I still don't know how to install an app (*.pkg) just using keyboard on macOS. On Windows, it's sim
10. 880.▲
    by behnamoh 2y ago

    Exactly my thoughts when I first saw the comments!
11. 881.▲
    by behnamoh 2y ago

    am I the only one who finds the unnecessarily positive vibes of OpenAI realtime voices unrealistic, too much, and borderline creepy?
12. 882.▲
    Haskell comprehension measured through WTF/min(ripplingbrainwaves.blogspot.com)

    6 points by behnamoh 2y ago|0 comments
13. 883.▲
    Hell: Shell Scripting Haskell Dialect(chrisdone.github.io)

    12 points by behnamoh 2y ago|0 comments
14. 884.▲
    Ask HN: Is there any programming language that doesn't have WTFs?

    5 points by behnamoh 2y ago|6 comments
15. 885.▲
    by behnamoh 2y ago

    half the amount of those dots is what it takes. but also, why compare a 27B model with a +600B? that doesn't make sense.
16. 886.▲
    by behnamoh 2y ago

    > ... velocity/error quotas start to inverse ... Could you please elaborate? Is this how management (at least in your company) looks at code—as a ratio of how fast it's done over how many tests it passes?
17. 887.▲
    by behnamoh 2y ago

    This is what local LLMs need—being treated like first-class citizens by the companies that make them. That said, the first graph is misleading about the number of H100s required to run DeepSeek r1 at FP16. The model is FP8.
18. 888.▲
    by behnamoh 2y ago

    You don't have all the context. The language serves a serious novel purpose. You seem to be bitter about someone uploading a research paper (I didn't say it was a white paper) to arXiv.
19. 889.▲
    by behnamoh 2y ago

    arXiv has one of my papers on hold for a long time because their team couldn't believe I—someone without a CS degree—was able to create a programming language from scratch on my own.
20. 890.▲
    by behnamoh 2y ago

    > ... now have to make choices in a place where they used to be fine ... That seems to be the customer's problem though, not the business owner. The cafe owner has no obligation to stick to the old way of doing things forever. Compa
21. 891.▲
    by behnamoh 2y ago

    > People are bored, insecure, or just looking for something to latch onto. So we pay more for a label that makes us feel seen. It’s not about taste—it’s about signaling. And cafés know that. They exploit it. Ruthlessly. This is a poor ta
22. 892.▲
    by behnamoh 2y ago

    I have this convention: <group-name><object-name><object-type> So in your case: ToDoPromise if it's the only todo.
23. 893.▲
    JavaScript for Automation Cookbook (2014)(github.com)

    3 points by behnamoh 2y ago|0 comments
24. 894.▲
    by behnamoh 2y ago

    > if it fails validation once, I route it to a stronger model just for that tool call. the problem the GP was referring to is that even the large model might fail to notice it's struggling to solve a task and keep trying more-or-les
25. 895.▲
    by behnamoh 2y ago

    > ... lua is a game changer... Am I the only one who absolutely hates Lua? Its syntax, the fact that variables are global by default unless you local them, `foo = nil` isn't variable assignment but instead variable unassignment!, et
26. 896.▲
    by behnamoh 2y ago

    He has more content with figures on another platform: https://typefully.com/DanHollick
27. 897.▲
    by behnamoh 2y ago

    Wait, did they just kill YOLO, at least for time-insensitive tasks?
28. 898.▲
    by behnamoh 2y ago

    So their method of sandboxing Python code is to spin up a JS runtime (deno), run Pyodide on it, and then run the Python code in Pyodide. Seems a lot of work to me. Is this really the best way to create and run Python sandboxes?
29. 899.▲
    Ask HN: Why do people buy Nvidia RTX 5090 at x2 the MSRP?

    11 points by behnamoh 2y ago|10 comments
30. 900.▲
    by behnamoh 2y ago

    OpenAI be like: o1, o1-mini, o1-pro, o3, o4-mini, gpt-4, gpt-4o, gpt-4-turbo, gpt-4.5, gpt-4.1, gpt-4o-mini, gpt-4.1-mini, gpt-4.1-nano, gpt-3.5-turbo