swores

19 ms·

271.
by swores 1y ago
Keep in mind that even in the excellent documentary to which you refer, about Dr Hugh "House" Laurie, MD. PhD. McT., the premise is that his setup - a dedicated diagnostics team to focus on difficult to diagnose cases - is a rare
272.
by swores 1y ago
It's an intentional policy shift that has come not only as guidance to medical professionals, but as an awareness campaign (posters in pharmacies etc.) to the general public that most minor infections don't need antibiotics, and s
273.
by swores 1y ago
Could you clarify what you mean by "dressed chicken"? And I'm curious, did all of your family find yourselves allergic or sensitive to all of those things, or is that list a combination that covers everyone while each individ
274.
by swores 1y ago
This is one of those times when I wish HN still displayed comment karma publicly, not only to the author of the comment. Because I'm sure various Microsoft employees read HN, and they should see what I assume will be a large number of
275.
by swores 1y ago
The difference is: what people generally mean by hallucination is "LLM said something wrong as if it was right". And what you are adding to that in your previous comments is the concept of whether or not the LLM knows the right an
276.
by swores 1y ago
But that's not how they actually work. > "If an LLM happens to know the answer to your question, that answer will have the greatest weight " An LLM doesn’t “know” anything in the way you’re imagining. It doesn’t have stor
277.
by swores 1y ago
No need for the apology, and FYI I broadly agree with everything you say (except about 4.5, which I don't actively disagree with I just haven't played with it myself).
278.
by swores 1y ago
That's not correct, and seems to be based on a common misunderstanding of how LLMs work, the rough idea being that when the info the model is being asked for had been in the data used for training, it "looks it up" not unlike
279.
by swores 1y ago
If you could share the actual prompts & info you wanted I would be curious to try and see if it is indeed too complex for it or if prompting differently would work better, because I've had it produce tables with multiple columns pu
280.
by swores 1y ago
I'm just specific in my prompting, rather than letting it decide whether or not to search. These models aren't (yet, at least) clever enough to understand what they do or don't know, so if you're not directly telling the
281.
by swores 1y ago
Ah, yes you're right - I didn't clarify this in my original comment, but my anecdote was indeed the ChatGPT interface and using its ability to browse the web[#], not expecting it to pull URLs out of its original training data. Tha
282.
by swores 1y ago
Which version of GPT? I've found that 4o has actually been quite good at this lately, rarely hallucinating links any more. Just two days ago, I gave it a list of a dozen article titles from a newspaper website (The Guardian), asked it
283.
by swores 1y ago
Last time I gave you the benefit of the doubt that you at least thought your comment was worth posting, now I assume you're just a troll and am blocking so I won't see your comments again.
284.
by swores 1y ago
Thanks for such an enlightening comment in which you were too lazy to write a 13th word naming the country you were thinking of. For anyone else curious as I was: Mexico is the 4th country, and I don't believe there are any others (but
285.
by swores 2y ago
They wrote that they were talking about country code TLDs vs not, not about US vs. other countries. (Which is what I would've said too, it's a more general point than thinking about anything specific to one country.) Ironically th
286.
by swores 2y ago
My apologies for failing to guess that when you complained they used too many fancy words to hide that the reason is to make more money, you actually weren't objecting to their use of a simple phrase acknowledging their goal of making
287.
by swores 2y ago
"Return on investment" isn't using uncommon words to try to be hard to interpret, it's a standard business term in English-speaking countries. The fact that it seemed like its intention was to obfuscate is just a sign th
288.
by swores 2y ago
Reddit's algorithm of what to show you is based in part on what you've chosen to follow and in part on what is most likely to get you to stay on the site and engage with the content. It's not designed to give you an accurate
289.
by swores 2y ago
> "That's a lot of fancy words to say ‘we're doing this because it makes us more money’ lol" You say that as if they're using lots of words to obfuscate that fact, but the quote you pasted has them saying entir
290.
by swores 2y ago
> "[...] the redirected url will still be in the history unless you manually redirect with a bare component. Hence whenever someone used back, the redirect pushed them back to the route they just left." I know this isn't
291.
by swores 2y ago
They aren't nitpicking your language: they asked for clarity on something where they weren't sure they understood what you meant, then when you clarified they responded disagreeing with you, not about your language. Feel free to d
292.
by swores 2y ago
Strongly disagree, both with the claim that the protestors don't care about peace elsewhere, and with the claim that Israel's actions are "defensive".
293.
by swores 2y ago
Oh those pesky people who want other people to behave like decent human beings rather than assholes. Bloody hopelessly woke!
294.
by swores 2y ago
> all sorts of big tech imposed hoops you don't have to jump through Could you kindly share some examples for those of us without big tech experience? I assume you're talking about working practises more than just annoying hi
295.
by swores 2y ago
The chart isn't claiming to be an overview of the best ranking models - it's an evaluation of this particular model, which wouldn't be helped at all by having loads more unrelated models in the chart, even if that would have
296.
by swores 2y ago
See the other HN submission (for the Gemma3 technical report doc) for a more active discussion thread - 50 comments at time of writing this. https://news.ycombinator.com/item?id=43340491
297.
by swores 2y ago
I think it's that it does make a sound, but cleverly directs it so that it bounces of the atmosphere in such a way that it doesn't end up travelling towards the ground.
298.
by swores 2y ago
But AI researchers are one of the resources currently needed! Even though there are lots of them, it's undeniable that if half of the best 1000 AI researchers in either US or China were to move to the other country it would be a massiv
299.
by swores 2y ago
> "I'm not sure how useful it is to keep investigating exactly how it works on a chemical level" It has the potential to be incredibly useful. Not because there's a benefit to being able to tell patients "this
300.
by swores 2y ago
Interesting, thanks for sharing - definitely makes me withdraw my confidence in that prediction, though I still think there's a decent chance they change their mind about that as it seems to me like an even worse naming decision than t

More ›