minimaxir

8 ms·

1. 181.▲
   by minimaxir 3mo ago

   Ironically my time and focus on agentic prompting and coding has helped me break an actual bad addiction (World of Warcraft)
2. 182.▲
   by minimaxir 3mo ago

   Yesterday (unrelated to quotamaxxing described in the article), I made an Apple TV macOS menu bar remote app: https://cdn.bsky.app/img/feed_fullsize/plain/did:plc:oxaerni... Every single prompt worked without
3. 183.▲
   by minimaxir 3mo ago

   For the problems I work on with GPT 5.6 Sol and the checks and balances I have in place, I estimate: - 80% of prompts get everything correct and are confirmed correct with manual validation - 19% of prompts make a minor mistake based on an
4. 184.▲
   by minimaxir 3mo ago

   Said tricks improve the output in an objective measurable manner, not theoretical, vibes, or gambler's fallacy. (blog post forthcoming on that) I've been researching LLM prompt optimization for longer than ChatGPT has existed; I w
5. 185.▲
   by minimaxir 3mo ago

   I really wish there was a canonical platform endpoint to programmatically check Codex usage amount and the quota reset times so I could just vibecode an app to alert "hey, your usage just went to 100% so that means OpenAI did a reset&q
6. 186.▲
   by minimaxir 3mo ago

   Over the last 2 weeks, 6 of the resets are direct to the usage limit. In the same timeframe, 2 additional resets were banked.
7. 187.▲
   by minimaxir 3mo ago

   Gambling addition implies dopamine hits from irregular and uncertain outcomes: as I note in the post, I don't get a dopamine hit from running agents. I know people say "LLMs are just gambling because they're next-token-maximi
8. 188.▲
   What's the deal with all the random weekly quota resets for agents lately?(minimaxir.com)

   85 points by minimaxir 3mo ago|89 comments
9. 189.▲
   by minimaxir 3mo ago

   Note that Max is the $100/$200 per month plan; In addition to the obvious incentive to upsell, I suspect attempting to use Fable on the $20 per month plan would wreck your quota. On a related note, Codex just reset their weekly quota,
10. 190.▲
    by minimaxir 3mo ago

    It's a common discussion trope to imply malfeasance in response to good news, which is a way to non-constructively shut down a conversation particularly without elaboration. In this particular case I legit didn't understand what t
11. 191.▲
    by minimaxir 3mo ago

    And? Is that a scandal?
12. 192.▲
    by minimaxir 3mo ago

    If you want to use local models, it's more ergonomic than fussing with GGUFs or using LM Studio as a server host and setting up the link to an agent yourself. Although, the model selector is the same as with LM Studio itself which can
13. 193.▲
    by minimaxir 3mo ago

    Hacker News and a RSS subscription for Techmeme is sufficient. X/Twitter is too irony/vaguepost filled that relying on it for high-signal news actually makes you less informed.
14. 194.▲
    LM Studio Bionic: the AI agent for open models(lmstudio.ai)

    331 points by minimaxir 3mo ago|134 comments
15. 195.▲
    by minimaxir 3mo ago

    You can do that now with a Stream Deck + XL. https://www.elgato.com/us/en/p/stream-deck-plus-xl
16. 196.▲
    by minimaxir 3mo ago

    Those memory bugs tend to be known unknowns. It's the unknown unknowns that are really scary.
17. 197.▲
    by minimaxir 3mo ago

    To paraphrase another comment ( https://news.ycombinator.com/item?id=48900086 ), this project is fun and fun should be encouraged, as experimentation is what leads to more innovative things. The existence of this project do
18. 198.▲
    by minimaxir 3mo ago

    Memory bugs are unknown unknowns that AI may or may not catch. There's net-present-value in switching to a language where certain types of memory bugs are impossible.
19. 199.▲
    by minimaxir 3mo ago

    Of all the tools in software engineering to be overpopular and overused, Rust is an instance where that is a very good thing. Atleast people aren't AI rewriting things into PHP.
20. 200.▲
    by minimaxir 3mo ago

    > they legitimately thought it was slop The issue is that label has expanded to anything that's weird/unusual, and the consequences of being accused of AI are far more severe than the consequences of a false accusation so peo
21. 201.▲
    by minimaxir 3mo ago

    I didn't respond to your argument because I didn't want to escalate, but since you insist on personally attacking me and breaking the HN guidelines far more explicitly than what you're accusing the OP of doing... No LLM has r
22. 202.▲
    by minimaxir 3mo ago

    That's a fair response.
23. 203.▲
    by minimaxir 3mo ago

    And then there's clanker (which has fortunately tapered off a bit): https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
24. 204.▲
    by minimaxir 3mo ago

    > I am skeptical YC frequency of "AI" related content is representative of actual users interest. Are you implying the front page is botted/manipulated?
25. 205.▲
    by minimaxir 3mo ago

    Also that.
26. 206.▲
    by minimaxir 3mo ago

    HN guidelines are to flag/downvote and move on anyways.
27. 207.▲
    by minimaxir 3mo ago

    That isn't related to my comment? My comment is more criticizing accusing something of being AI based on vibes, and likely being wrong about it.
28. 208.▲
    by minimaxir 3mo ago

    AI written comments have been more constructive than what you have been commenting thus far in this post.
29. 209.▲
    by minimaxir 3mo ago

    And what's stopping a AI-written article from also doing that?
30. 210.▲
    by minimaxir 3mo ago

    That's just LinkedIn.