nabakin

19 ms·

1. 211.▲
   by nabakin 3y ago

   Don't rely too much on automated benchmarks for LLMs. They are often gamed, made to overfit, and result in worse performance in the general case. Human evaluation is the gold standard and the Llama 2 paper gave significant evidence tha
2. 212.▲
   by nabakin 3y ago

   Yeah, well, if iOS didn't force everyone to use the Safari engine, maybe the situation would be better
3. 213.▲
   by nabakin 3y ago

   Interesting that HN was able to DDOS a Google service
4. 214.▲
   by nabakin 3y ago

   Thanks for this. I'm glad people have good, evidence-based responses to my comment. This gives us a great idea of how likely a Crimean who considers themselves Russian would actually vote between the two and that while the correlation
5. 215.▲
   by nabakin 3y ago

   You make some good points. I agree, any census done after Russia took Ukraine in 2014 can't be used and I don't doubt people who once considered themselves Russian started to consider themselves Ukrainian after Russia attacked Ukr
6. 216.▲
   by nabakin 3y ago

   While this is a provocative response and there is no excuse for the Russian invasion of Ukraine, the 2001 Ukrainian census[1] states 60.4% of the Crimean population considered themselves Russian and 24% of the Crimean population considered
7. 217.▲
   by nabakin 3y ago

   Aaaand now it's back
8. 218.▲
   by nabakin 3y ago

   > Microsoft made him walk that one back real quick - he probably wasn't really aware that he is now one of their employees Source?
9. 219.▲
   by nabakin 3y ago

   Edit: they added the line back again -_- They just made another edit and removed the line. Here's the edit history going all the way back to March: - 4/1 https://www.diffchecker.com/dCuVSMnp/ - 7/1 ht
10. 220.▲
    by nabakin 3y ago

    It's gone. They removed this part. https://www.diffchecker.com/RLiqgAaA/
11. 221.▲
    by nabakin 3y ago

    I'd go so far as to say this entire post is grossly misleading and should be flagged.
12. 222.▲
    by nabakin 3y ago

    Not sure. It's working for me
13. 223.▲
    by nabakin 3y ago

    It sounds like the problem I was having on FF. Switch to Chrome and see if this problem persists
14. 224.▲
    by nabakin 3y ago

    I figured out a way to get a list of videos all matching the time you want. Should help. https://news.ycombinator.com/item?id=36499194
15. 225.▲
    by nabakin 3y ago

    Should help https://news.ycombinator.com/item?id=36499194
16. 226.▲
    by nabakin 3y ago

    I'll copy a comment I made earlier: "00:00 / <minutes>:<seconds>" youtube Google that, including the quotes, replace the minutes and seconds with your given time, then look through the results and find a URL
17. 227.▲
    by nabakin 3y ago

    "00:00 / <minutes>:<seconds>" youtube Google that, including the quotes, replace the minutes and seconds with your given time, then look through the results and find a URL which fits the password criteria. Every r
18. 228.▲
    by nabakin 3y ago

    Reached 28 and forgot to feed Paul. That Rule is so frustrating
19. 229.▲
    by nabakin 3y ago

    tl;dr K = King Q = Queen B = Bishop R = Rook N = Knight Pawns have no letter . You append a "+" to the end when it is a check move You append a "#" to the end when it's a checkmate move You add an "x" when
20. 230.▲
    by nabakin 3y ago

    I think the chess one is bugged. I've entered every possible move for the board I received and none of them worked. Edit: figured it out
21. 231.▲
    by nabakin 3y ago

    If anyone is interested in how many of the top 250 subreddits are participating, I created a website for it here https://save3rdpartyapps.com/
22. 232.▲
    by nabakin 3y ago

    Not total activity, but I think half of the top 250 subreddits participating in the protest gives a good idea https://save3rdpartyapps.com/
23. 233.▲
    by nabakin 3y ago

    Around half of the top 250 are participating atm https://save3rdpartyapps.com/ which I think gives a better idea of how big it is. Despite that, I think it won't be enough to change anything. Maybe if we got all top 25
24. 234.▲
    by nabakin 3y ago

    Around half of the top 250 aren't participating atm https://save3rdpartyapps.com/
25. 235.▲
    Save3rdPartyApps.com: See how many of the top subreddits are going dark(save3rdpartyapps.com)

    2 points by nabakin 3y ago|0 comments
26. 236.▲
    by nabakin 3y ago

    The Reddit API will still have a free tier after the changes on July 1st which can be used and Shreddit uses the Reddit API too so it will have the same problem.
27. 237.▲
    by nabakin 3y ago

    I wrote this comment a couple years ago, but unless something significant has changed, it still stands. Be careful. Shreddit does not get rid of all reddit comments, only the ones on your profile. Reddit stores a list of the last 1000 comme
28. 238.▲
    by nabakin 3y ago

    I would differentiate between those subs but I don't have a list of them. I've created a wiki page[1] if people want to start compiling that info [1] https://github.com/xNul/save-3rd-party-apps/wiki/
29. 239.▲
    by nabakin 3y ago

    Green means they are participating. Thanks, I'll put something for that. I wondered if it would be clear enough
30. 240.▲
    by nabakin 3y ago

    I made a website for the protest. Could you folks critique it for me? https://save3rdpartyapps.com