munchler

22 ms·

1. 361.▲
   by munchler 2y ago

   > Yann Lecun ... argued that because language models generate outputs token-by-token, and each token introduces a new probability of error, if we generate outputs that are too long, this per-token error will compound to inevitable failur
2. 362.▲
   by munchler 2y ago

   Betteridge's law of headlines applies: Any headline that ends in a question mark can be answered by the word no . (If the answer was yes , that would be the headline instead.)
3. 363.▲
   by munchler 2y ago

   I don't buy that. The string "Null" is different from the keyword null in programming, so `if $var = null` would be false when $var is the string "Null". Note that when interpolated into SQL, the contents of $var mu
4. 364.▲
   by munchler 2y ago

   This is exactly how I felt as a photographer when smart phones took over the world 15 years ago. I try to remind myself that democratization is good, even if it isn’t necessarily pretty.
5. 365.▲
   by munchler 2y ago

   The subtle irony of this long article on organic beauty that slowly turns into a discussion of… consumer product design.
6. 366.▲
   by munchler 2y ago

   People are still going to be writing papers like this after AGI has taken over the world. Denial is a strong impulse.
7. 367.▲
   by munchler 2y ago

   I would credit F# even more. It's been a main source of new C# features for a while now.
8. 368.▲
   by munchler 2y ago

   Chaos is still deterministic in classical physics. If you start out with the same initial conditions, the system always evolves in the same perfectly deterministic (but chaotic) way.
9. 369.▲
   by munchler 2y ago

   You've got it backwards. Newtonian theory does not rule out perfect spheres (or any other mathematically perfect shape), hence the possibility of non-determinism.
10. 370.▲
    by munchler 2y ago

    Norton's Dome isn't spherical, so it's a bit more sophisticated than you imply. In classical physics, you could roll a ball up the dome so that it comes to a rest at the apex in finite time. Thus, due to time symmetry, it&#x2
11. 371.▲
    by munchler 2y ago

    Sorry, how does the snow factor into this tale?
12. 372.▲
    by munchler 2y ago

    I did a quick search for references to mt32-pi on that forum and found nothing unpleasant. Maybe I'm missing something. https://www.vogons.org/search.php?st=0&sk=t&sd=d&sr=posts&ke...
13. 373.▲
    by munchler 2y ago

    Thank you for the suggestion. I will give it a try.
14. 374.▲
    by munchler 2y ago

    Scientific American used to be a great magazine and should get back to what made it so valuable: Covering important ideas in science without watering them down. Their website now describes it as "the essential guide to the most awe-ins
15. 375.▲
    by munchler 2y ago

    I grew up reading SciAm in the 70's and 80's, and I don't remember a single article about politics.
16. 376.▲
    by munchler 2y ago

    As an FP fan, I agree. This is disappointing.
17. 377.▲
    by munchler 2y ago

    Downloading and running software is not speech. (Just trying to steel-man the opposing argument.)
18. 378.▲
    by munchler 2y ago

    Trying to meet strange management requirements is normal for just about any professional developer. I don’t understand why you think game developers deserve a special exemption.
19. 379.▲
    by munchler 2y ago

    YES. Did you read the part where the game devs use RCE with admin privileges to run patches? Any developer who does that should be aware of the security risks they’re taking.
20. 380.▲
    by munchler 2y ago

    That's why they keep the evaluation set private: "Submit a solution which scores 85% on the ARC-AGI private evaluation set and win $600K." [0] https://arcprize.org/guide
21. 381.▲
    by munchler 2y ago

    That's why I put "AGI" in quotes. The point is that six months ago, no one expected an LLM to score this well.
22. 382.▲
    by munchler 2y ago

    Take a look at the ARC Prize, which is a test for achieving "AGI" created in 2019 by François Chollet. Scroll down halfway on the home page and ponder the steep yellow line on the graph. That's what OpenAI o3 recently achieve
23. 383.▲
    by munchler 2y ago

    I never said things are improving in the US.
24. 384.▲
    by munchler 2y ago

    I understand that you're unhappy with the state of things in the US, but setting up a false equivalence with China doesn't make your case. The simple fact that we can have this discussion without fear of imprisonment is strong evi
25. 385.▲
    by munchler 2y ago

    The MOVE bombing was thoroughly reported at the time and litigated afterwards. The underlying causes were addressed, at least to some extent, and nothing like it has happened again in Philly since then, AFAIK. That’s why it isn’t well known
26. 386.▲
    by munchler 2y ago

    Soon you’ll be freed of the tedious task of upvoting at all.
27. 387.▲
    by munchler 2y ago

    https://marketoonist.com/2023/03/ai-written-ai-read.html
28. 388.▲
    by munchler 2y ago

    They were first, which is a huge accomplishment. History will remember the ChatGPT launch no matter what happens next. However, first is not necessarily best.
29. 389.▲
    by munchler 2y ago

    Dang, I was hoping this might be something cool about graphs in the Lean language/theorem prover.
30. 390.▲
    by munchler 2y ago

    > Food manufacturers use the SRMs to calibrate their test methods and equipment. So we trust them to self-report their nutrition labels, rather than having an unbiased lab perform the test? That seems problematic.