th0ma5

10 ms·

1. 151.▲
   by th0ma5 1y ago

   For sure when people don't understand the fundamentals (or in the case of LLMs they are unknowable) then all you have is superstition.
2. 152.▲
   by th0ma5 1y ago

   Perhaps you should add a privacy policy or just release the source rather than assume people will trust your site. Why do you do these demos if you aren't upfront about all the things the LLMs didn't do?
3. 153.▲
   by th0ma5 1y ago

   There's to me a sole unimpeachable use case... Two SMEs of adjacent fields can work with an LLM SME to find overlaps in research. This is a manual slog of evaluating hints at patterns and holistically trying to discover something and t
4. 154.▲
   by th0ma5 1y ago

   We have no idea how they work. We know the training making a model and generating things from it, but the thing we want from it being an Oracle or something, we have no idea how that works at the level of the specific knowledge it generates
5. 155.▲
   by th0ma5 1y ago

   I think about this occasionally trying to rationalize it. I see similar patterns in other things like R and Julia where they design something in the environment to seem like a composable tool, and maybe it is but only within two or three sp
6. 156.▲
   by th0ma5 1y ago

   I don't have dissonance. The language has the features that were designed or there is a bug in what they intended and I don't know why you think adding two empty objects together should make sense anyway. That it doesn't work
7. 157.▲
   by th0ma5 1y ago

   For me the difference is that you wouldn't then feel compelled to cynically bemoan it for retweets but this schtick went on for years where people would do something out of spec and then complain it was really bizarre to me! Anyway ...
8. 158.▲
   by th0ma5 1y ago

   In the context of the time it was created it was fine to mess with and having an enlightened view from the future can't negate that even though I understand the complaint.
9. 159.▲
   by th0ma5 1y ago

   I've read your comment several times and I still don't get how I'm supposed to be frustrated by this?
10. 160.▲
    by th0ma5 1y ago

    Thanks for posting this. People still don't seem to get that there's no mystery out to make their life difficult and that unfortunately everything can probably ultimately be understood in a deterministic system.
11. 161.▲
    by th0ma5 1y ago

    This guy single handedly caused a wave of unproductive cynicism with this stunt. It's like the joke about all the things an everything bagel doesn't have except essentially shitting on the hard work of the entire industry by makin
12. 162.▲
    by th0ma5 1y ago

    Reminding me of the near decade of WAT snark where people thought undefined behaviour was a complete proof of the futility of technology when it was actually just people mistaking what technology is. Like it isn't funny that you can&#x
13. 163.▲
    by th0ma5 1y ago

    If it is all original, you'll want to be careful with the power supply. It's known to fail in a way that over voltages. Replacement ones can be made or bought relatively inexpensively.
14. 164.▲
    by th0ma5 1y ago

    Since when do people like the fuzziness of outputs? I think you make an interesting point but it also seems to imply that benchmarking will never truly be possible, which I think is true unless we can also make them observable which also as
15. 165.▲
    by th0ma5 1y ago

    Pipes and services in general are reliable but the issues were social and economic. Getting everyone to agree was seen as a great way to poach users and give up control, plus the usual problems with open world vs. closed world assumptions.
16. 166.▲
    by th0ma5 1y ago

    Simon's opinion is unsurprisingly that people need to read his blog and spam on every story on HN lest we be left behind.
17. 167.▲
    by th0ma5 1y ago

    I think you to as he is human, but I respect your desire to question it!
18. 168.▲
    by th0ma5 1y ago

    Yup
19. 169.▲
    by th0ma5 1y ago

    People should be demanding consistency and traceability from the model vendors checked by some tool perhaps like this. This may tell you when the vendor changed something but there is otherwise no recourse?
20. 170.▲
    by th0ma5 1y ago

    Maybe you should try something other than demos? Have you tried creating a reliable system?
21. 171.▲
    by th0ma5 1y ago

    I have a lot of respect for organizations that get a lot done with Microsoft technologies. I think your perspective could be thought of as the benefits of vertical integration and vendor lock in. These do help people get things done! In the
22. 172.▲
    by th0ma5 1y ago

    I thought this was a very good read about the many of the issues that are faced without having any ground truth to reason against. It is interesting how many different ways people have developed to work around missing information, and the m
23. 173.▲
    by th0ma5 1y ago

    This also assumes that non human written code will be of any use to humans and no one has shown that to be possible, it is all humans patching it up so far.
24. 174.▲
    by th0ma5 1y ago

    This has been the problem with higher level natural language programming for years. I really wonder what people are doing if they don't see this core issue that precludes their use.
25. 175.▲
    by th0ma5 1y ago

    A lot of people try to hedge this kind of sober insight along with their personal economic goals to say all manner of unfalsifiable statements of adequate application in some context, but it is refreshing to try to deal with the issues sepa
26. 176.▲
    by th0ma5 1y ago

    Anecdotes unfortunately are not data :/
27. 177.▲
    by th0ma5 1y ago

    Anecdotes unfortunately are not data :/
28. 178.▲
    by th0ma5 1y ago

    That you don't know when it will make a mistake and that it is getting harder to find them are not exactly encouraging signs to me.
29. 179.▲
    by th0ma5 1y ago

    For sure, just the gains of LLMs in the mix cannot be measured and most still recommend human in the loop as always.
30. 180.▲
    by th0ma5 1y ago

    Yup! I've done this and it sucks!