camgunz

9 ms·

1. 151.▲
   by camgunz 7mo ago

   Can't you just drop the ECH signals, no matter what site it is? Don't you then mostly disable sites you don't want people to see anyway? Maybe like, you can't download Chrome anymore, but I bet there would be a Russian f
2. 152.▲
   by camgunz 7mo ago

   1. Irrelevant. I've delivered example after example of your fave model bullshitting. You should've bitten the bullet long ago. Honestly I'm disappointed; I've seen you in a lot of AI threads and assumed you'd be goo
3. 153.▲
   by camgunz 7mo ago

   If it doesn't support IPv6 it doesn't work.
4. 154.▲
   by camgunz 7mo ago

   GPT-5.4 gets 82.7% on Browsecomp (a benchmark specifically testing tool use), which is a hallucination rate of 17.3%, on questions like "Give me the title of the scientific paper published in the EMNLP conference between 2018-2023 wher
5. 155.▲
   by camgunz 7mo ago

   I've got about 20 minutes in this; mostly I've been reading wallstreetbets at the Shake Shack bar in the Boston airport. I'm happy to post this over and over again until you engage w/ it: > I found over 500 examples t
6. 156.▲
   by camgunz 7mo ago

   I found over 500 examples that fit your criteria. Embarrassing you were arguing in bad faith this whole time.
7. 157.▲
   by camgunz 7mo ago

   > Specifically in the case where it can use tools - no it doesn't hallucinate. OpenAI's own system card says it does. Hallucination rates in GPT-5 with browsing enabled: - 0.7% in LongFact-Concepts - 0.8% in LongFact-Objects -
8. 158.▲
   by camgunz 7mo ago

   > I don't think the spirit of the original article (not your comments to be fair) captured this, hence the challenge. I believe we are on the same page here. No. GPT-5 has a 40% hallucination rate [0] on SimpleQA [1] without web sea
9. 159.▲
   by camgunz 7mo ago

   I saw; I replied up there
10. 160.▲
    by camgunz 7mo ago

    I could quibble with some things, but this is right. I don't have a paid account so I can't ping away at 5.4 or whatever, but, I do have access to frontier models at work, and they hallucinate regularly. Dunno what to do if you do
11. 161.▲
    by camgunz 7mo ago

    Sorry: https://chatgpt.com/share/69d6ac63-d200-8330-8c47-95a75db8bb... Also what? The repo bit is clear bullshit.
12. 162.▲
    by camgunz 7mo ago

    The thinking version is mostly right, but: - it searches the internet to find the answer, it doesn't "reason". I'm not claiming Google is a bullshit machine, and it's not surprising the answer is discoverable (it ha
13. 163.▲
    by camgunz 7mo ago

    "Hey ChatGPT. I'm building a Final Fantasy 6 mod, and I need more space for the battle scripts. How would I rearrange the data in the ROM to give me the extra space I need?" https://chatgpt.com/share/69d6
14. 164.▲
    by camgunz 7mo ago

    "Hey ChatGPT. I've recently grown horns and I need some care advice. Should I polish my horns before going to have them trimmed or will the horn trimmer polish them for me?" https://chatgpt.com/share/69d6
15. 165.▲
    by camgunz 7mo ago

    "Hey ChatGPT. How would you describe me?" https://chatgpt.com/share/69d69780-ae58-83e8-a41c-7d10a5f298... It has no conversations and no memory of me. Maybe this is true, maybe it isn't, but there's
16. 166.▲
    by camgunz 7mo ago

    I'm earnestly curious why not.
17. 167.▲
    by camgunz 7mo ago

    If I have to suffer "look at this busted ass thing I slopped out with AI" a few times a week, you all have to suffer grouchy "AI bad" a few times a week. Fair is fair.
18. 168.▲
    by camgunz 7mo ago

    Humans and large models are both unpredictable and fallible, that's true, but in different ways, and (many) humans are actually much better at following directions. If a junior dev makes the same mistake Claude makes, I can easily work
19. 169.▲
    by camgunz 7mo ago

    Well, large models are (kinda) non-deterministic in two ways. The first is you actually provide many of them with a seed, which is easy to manage--just use the same seed for the same result. The second part is the "you actually have ve
20. 170.▲
    by camgunz 7mo ago

    Honest question, how much money would I make off an MCP service to generate passwords for claws and agents. Is there still gas left in the griftmobile, are prospectors still in need of shovels, will the gods bless my humble, shameless lunge
21. 171.▲
    by camgunz 7mo ago

    Nah I think the advice generally is "ignore options."
22. 172.▲
    by camgunz 7mo ago

    This binary is old and tired. It's possible to want to ship things and think LLMs make that harder. You don't engage with TFA at all, which is exploring this exact point: that the more slop that gets into prod the harder it will b
23. 173.▲
    by camgunz 7mo ago

    Bluesky has all the things that made Twitter a cesspool: a small character limit (300 is still too small) and virality stats (likes/retweets). Its only saving grace right now is that it's itty bitty; if it ever achieved scale it w
24. 174.▲
    by camgunz 7mo ago

    What do you when the things you want to change don't all fit on the screen at once?
25. 175.▲
    by camgunz 7mo ago

    The article is essentially just the Sky News video in print. The parts that aren't are like "we couldn't verify this", like this quote: "I observed multiple deep, uniform lesions on his lower body, which are clinica
26. 176.▲
    by camgunz 7mo ago

    Shaddap
27. 177.▲
    by camgunz 7mo ago

    It's yet another "silver bullet". That's why management loves it and engineers are like "it's just a tool, we still need code review."
28. 178.▲
    by camgunz 7mo ago

    I feel like this is just the bubble talking. I'm pretty naive here, but at some point suppliers will adjust so they can take money from data center builders and consumers, just like pre-bubble.
29. 179.▲
    by camgunz 7mo ago

    Wait could I have written a JSONata parser and sold it to reco.ai for $499k/yr?
30. 180.▲
    by camgunz 7mo ago

    They literally just voted it down. Twice in 2 days. Also compared to whom?