ComplexSystems

9 ms·

1. 301.▲
   by ComplexSystems 4y ago

   What does "1/0.85" mean here? The reciprocal of 0.85 is ~1.176, so is that what you are recommending?
2. 302.▲
   by ComplexSystems 4y ago

   How are you getting such good results? Here's what I get, all with llama-cpp and 4 bit quantization: Here is LLaMA 7B, 4 bits quantized: "us presidents in chronological order: george washington,john adams, james madison, james mon
3. 303.▲
   by ComplexSystems 4y ago

   > It's great at a personal level that "founders" and startup employees didn't have to do without. But it's important to remember that they no longer automatically deserve any credit for taking risks and doing som
4. 304.▲
   by ComplexSystems 4y ago

   What can this do that iTerm2 can't?
5. 305.▲
   by ComplexSystems 4y ago

   How are you getting this to run fast? I'm on a top of the line M1 MBP and getting 1 token every 8 minutes.
6. 306.▲
   by ComplexSystems 4y ago

   I got one token every 8 minutes or so.
7. 307.▲
   Ask HN: Has anyone switched from Chrome to Edge to play with the new AI stuff?

   4 points by ComplexSystems 4y ago|2 comments
8. 308.▲
   by ComplexSystems 4y ago

   > Yes. As a judge can check what information was used and what algorithm applied to get to a conclusion. And it may declare that use of data or algorithm illegal. A ban on "AI social scoring" would be neither necessary nor suff
9. 309.▲
   by ComplexSystems 4y ago

   The smallest model (7B) is supposed to outperform GPT-3. Does anyone have any idea what hardware is needed to run this?
10. 310.▲
    by ComplexSystems 4y ago

    It is probably worth noting that you don't even need the user to click on anything. Bing will readily go and search and read from external websites given some user request. You could probably get Bing, very easily, to just silently tak
11. 311.▲
    by ComplexSystems 4y ago

    Good post. Tangentially related: I have thought recently that we really need to refresh the way that we explain these ideas to beginners in this new world of AI/ML doing everything. For instance, here is an excerpt from Scott Aaronson&
12. 312.▲
    by ComplexSystems 4y ago

    The "vast majority" of SAT instances would be something like a SAT instance you would draw uniformly at random from some kind of distribution on all possible circuits. Making this idea rigorous is what k-random SAT is all about. I
13. 313.▲
    by ComplexSystems 4y ago

    This seemed interesting. So if I get your idea correctly, rather than talking to the chatbot directly, you first run your prompts through some algorithm which increases the chances of the AI getting what you are asking and giving a successf
14. 314.▲
    by ComplexSystems 4y ago

    Do you have a link to the video?
15. 315.▲
    by ComplexSystems 4y ago

    This is a good point. I think that if you really want something like these named bools but which keeps the same lazy evaluation behavior, you'd have to use preprocessor macros then.
16. 316.▲
    by ComplexSystems 4y ago

    I switched from Homebrew to Macports, although I may switch back. Basically, Homebrew will completely crap out on you if you are ever more than two macOS versions out of date. The brew people simply don't keep the older bottles around
17. 317.▲
    by ComplexSystems 4y ago

    Is this the same GPT-3 which is available in the OpenAI Playground?
18. 318.▲
    by ComplexSystems 4y ago

    Wouldn't this still be about 36 terabytes of data?
19. 319.▲
    by ComplexSystems 4y ago

    Here is the company that made it: https://www.mismatch.media How much do you think it costs to actually create all this? If it's using GPT-3 I can see that being quite pricey.
20. 320.▲
    by ComplexSystems 4y ago

    Sometimes.
21. 321.▲
    by ComplexSystems 4y ago

    I was literally about to comment this as well. Actually, that's the biggest plus for ChatGPT right now.
22. 322.▲
    by ComplexSystems 4y ago

    What did OpenAI do, exactly?
23. 323.▲
    by ComplexSystems 4y ago

    Ok. So is that what you want to see? Just train the model on natural language processing of math word problems? Then it's capable of what you call "real" thought?
24. 324.▲
    by ComplexSystems 4y ago

    No. People are, if anything, underestimating GPT pretty significantly at this point. The entire conversation always involves this endless spiel about how GPT is just generating plausible-sounding text and doesn't "really understan
25. 325.▲
    What is HN's take on online Master's programs in Machine Learning/Data Science?

    1 points by ComplexSystems 4y ago|0 comments
26. 326.▲
    by ComplexSystems 4y ago

    Yeah yeah, but "an Google"?
27. 327.▲
    by ComplexSystems 4y ago

    It isn't that I'm changing "specified" to "known"; I'm pointing out that this is an article on a purported "encyclopedia of unknowns," and it's only being included there because the authors
28. 328.▲
    by ComplexSystems 4y ago

    I'm not sure any mathematician would agree with this page: https://wikenigma.org.uk/content/mathematics/irrational_numb... It says "An irrational number is a real number that cannot be expressed as a rat
29. 329.▲
    by ComplexSystems 4y ago

    "In a proposed class action filed on Wednesday, Christopher Lazazzaro said he would not have bought or would have paid less for Hershey's Special Dark Mildly Sweet Chocolate, Lily's Extra Dark Chocolate 70% Cocoa and Lily&#x2
30. 330.▲
    by ComplexSystems 4y ago

    I wish there were some kind of Easter egg a few minutes in. Perhaps a rogue "neigh" or something