Live data from Hacker News

Viewing profile — richdougherty

richdougherty

HN member
Joined
Fri, Aug 12, 2011, 11:27 PM UTC
HN karma
1,796
Public activity
403 items

About richdougherty

https://rd.nz

Recent public activity

  1. comment
    Comment #47125431

    If you'll excuse the self-post, here's a blog post on goto with delimited continuations. https://rd.nz/2009/03/goto-in-scala.html It uses an experimental compiler plugin for the Sc…

  2. comment
    Comment #44989273

    Seems likely because the article says "Meta’s deal with Google is mainly around artificial intelligence infrastructure, said one of the people".

  3. comment
    Comment #42604842

    "The cost-effective nature of AI makes it highly plausible we're moving towards an agent vs agent future." Sounds right. I assume we will all have AI agents triaging our emails try…

  4. comment
    Comment #41855210

    Kudos to the dev for coming up with the eye position fixing solution. Building further on this idea, I wonder if instead of changing the image to look at the camera, we could chang…

  5. comment
    Comment #41394966

    So it's also vulnerable to a Helicopter Injection Attack?

  6. comment
    Comment #40817821

    If you have control of the tokenizer you could make sure it doesn't produce these tokens on user input. I.e. instead of the special " " token, produce something like " " - whatever…

  7. comment
    Comment #40746413

    Source with more information: https://www.tuxedocomputers.com/en/TUXEDO-on-ARM-is-coming.t...

  8. comment
    Comment #40418813

    A few details here: "Recall leverages your personal semantic index, built and stored entirely on your device. Your snapshots are yours; they stay locally on your PC. You can delete…

  9. comment
    Comment #40371610

    I can't track down the citation (either Google or DeepMind I think), but I remember reading research from a year or two ago how adding extra languages (French, German) improved Eng…

  10. comment
    Comment #40330303

    Related? "Let's Think Dot by Dot: Hidden Computation in Transformer Language Models" https://arxiv.org/abs/2404.15758 > Chain-of-thought responses from language models improve perf…

  11. comment
    Comment #39958050

    Anyone who enjoyed this, might also like "All the Horses of Iceland". It's a fictional work, but written by a medieval scholar. It's set in the 9th century where a Norse trader tra…

  12. comment
    Comment #39657462

    http://ymeskhout.substack.com/p/eleven-magic-words-unlocked

  13. comment
    Comment #38999481

    Agree, nesting files with cat >Dockerfile and having a basic bash script seems way nicer than putting all the shell logic on the #! line.

  14. comment
    Comment #38417814

    https://twitter.com/ylecun/status/1728126868342145481 "Please ignore the deluge of complete nonsense about Q . One of the main challenges to improve LLM reliability is to replace A…

  15. comment
    Comment #38286307

    Site not loading for me, but probably this: https://nothing.tech/pages/nothing-chats

  16. comment
    Comment #38232863

    Just speculating, but perhaps share prices would drop until the returns per dollar invested recovers to the desired level. So it might be like a one off shock to investors who hold…

  17. comment
    Comment #38186923

    Supported in Java 10 (and backported to Java 8) since 2018. Not sure about .NET. - "The JVM has been modified to be aware that it is running in a Docker container and will extract …

  18. comment
    Comment #37940173

    https://math.stackexchange.com/users/97378/cleo

  19. comment
    Comment #37823554

    https://en.m.wikipedia.org/wiki/Generalized_second-price_auc...

  20. comment
    Comment #37788868

    Agreed, it needs to be Android based - the other option would be to implement iOS APIs, but Android would be a lot easier. An analogy would be if you make a browser, to use standar…

  21. comment
    Comment #37788812

    https://arxiv.org/abs/2310.02207

  22. comment
    Comment #37788786

    A really interesting article! https://thegradient.pub/othello/ "How do these models achieve this kind of performance? Do they merely memorize training data and reread it out loud, …

  23. comment
    Comment #37698152

    I'd love them to fund an Android variant or replacement for Android services. Mobile is the other main platform apart from web.

  24. comment
    Comment #37425724

    You can run them either for general purpose inference. You can also fine-tune them and get improved performance for specific use cases. It's safe to assume they're worse at every t…

  25. comment
    Comment #37266827

    I haven't read the article, but "The Winner's Curse" is an economics concept I'm familiar with. Basically, The Big Insight is that every bidder has an estimate of the value of some…