k__

5 ms·

1. 1.▲
   by k__ 6d ago

   Update: I'm using it right now and it's noticeably faster. I'd also say, it seems smarter, but I think that's because of some harness updates I installed. (I haven't used pi for almost a month)
2. 2.▲
   by k__ 7d ago

   So, while the throughput was 400-500tps in beta its now ~150tps on OpenRouter. I was hoping for a bit more, but it's still 100% faster for a very good price, so I won't complain.
3. 3.▲
   by k__ 7d ago

   Nice form factor. I have a Samsung Fold 4. The front screen is too narrow and the folding screen doesn't open all the way anymore. Not what I expected from a 1400€ phone.
4. 4.▲
   by k__ 7d ago

   At least they can use another provider or self-host.
5. 5.▲
   by k__ 7d ago

   Beta testers report >400 TPS. https://www.geeky-gadgets.com/deepseek-v4-1-flash-review/ I hope some of those speed increases will make it to production.
6. 6.▲
   by k__ 7d ago

   I used flash with pi and it worked pretty well. It built this whole IaC plugin from scratch: https://github.com/fllstck/nebius-alchemy
7. 7.▲
   by k__ 8d ago

   I don't know about animation, but it seems the UI system and their widget library was overhauled. "The Feathers widgets have migrated to BSN, Bevy's next-generation scene system. BSN is a better foundation for widgets than t
8. 8.▲
   by k__ 8d ago

   Yeah, I think only in crypto the non-us providers have a leg up, because running a node just needs an instance.
9. 9.▲
   by k__ 10d ago

   Yeah, there are plenty of slow burns. Yet, I wouldn't go so far as to say every novel is one. If you aren't an established author, publishers won't take on your manuscript if it doesn't hook them with the first chapter,
10. 10.▲
    by k__ 10d ago

    Interesting that he has this view. I would have taken him for mostly pro LLM.
11. 11.▲
    by k__ 13d ago

    Any signal that allows the model to see what's wrong helps. Checking code is (relatively) easy, you can use static type checks, linters, and execute it to see if it's correct. Fact checking is harder. A RAG can only check what&#x2
12. 12.▲
    by k__ 14d ago

    I had the impression that GLM 5.3 is just open weight and not open source.
13. 13.▲
    by k__ 14d ago

    I use a Google sheet.
14. 14.▲
    by k__ 14d ago

    Somehow RRS removed their search function recently. Whelp, as with all those aggregators, it was nice while it lasted.
15. 15.▲
    by k__ 19d ago

    I can't even login to languagereactor.
16. 16.▲
    by k__ 19d ago

    And with exponential, you mean sigmoid.
17. 17.▲
    by k__ 19d ago

    Cool idea! While I'm not so sure about learning a language with content from fictional sources, that should at least make it more relevant for the learners. And you can use it on podcasts too, where people talk normally. The biggest hu
18. 18.▲
    by k__ 20d ago

    The "marked" consolidated.
19. 19.▲
    by k__ 22d ago

    I'm always amazed by solutions for user interactions. I once built a Bomberman clone that had control issues. The first iteration of the controls were clunky, because I only checked if a player was touching a block in the direction the
20. 20.▲
    by k__ 28d ago

    I'm working remote since 2014. I built my whole life on that premise, traveling through the country on a weekly basis. I love the freedom.
21. 21.▲
    by k__ 29d ago

    Just a few days ago, I received one because of "the odour nuisance is no danger for the population" People were calling the emergency hotline because the wildfires 200km away got so big that their smoke engulfed the city.
22. 22.▲
    by k__ 1mo ago

    I had the opposite impression. LLMs can abstract pretty well. They write quality code for obscure frameworks no problem, as longs as you take the time to set up tight feedback loops.
23. 23.▲
    by k__ 1mo ago

    Yeah, None of them offers the same cache hit prices and my cache hits are >95%.
24. 24.▲
    by k__ 1mo ago

    I didn't get the impression that anyone competed with the old prices before.
25. 25.▲
    by k__ 1mo ago

    I get like 80.
26. 26.▲
    by k__ 1mo ago

    I wouldn't exactly call it snappy, but faster than Pro, yes.
27. 27.▲
    by k__ 1mo ago

    I tried the previous Pro model and in the end it was 50% more expensive than the previous Flash. Wasn't worth it.
28. 28.▲
    by k__ 1mo ago

    Around 5 percentage points better. (E.g., 87% instead of 82%)
29. 29.▲
    by k__ 1mo ago

    Had the same impression about TypeScript and Rust. Not as fun to write as Python and Nim, but I don't have to write it.
30. 30.▲
    by k__ 1mo ago

    Care to elaborate?