Ukv

37 ms·

451.
by Ukv 2y ago
> blames software that he told to remove all his files when the files were removed There's definitely a lot of anger in the report, but underlying that I think there's a valid UX issue in that "discard changes" does n
452.
by Ukv 2y ago
True that memoization makes factorial trivial, but I think that's erring on the side of skipping the computation we're meant to be benchmarking; a pre-computed array may be even faster. Numba's JIT is applicable in a wide ran
453.
by Ukv 2y ago
That you were able to use Python to define and train the network in so few lines of code, with no prior knowledge and a 2-day deadline, sounds fairly successful to me. I think a hackathon will inevitably favor quick copy-pasting, gluing API
454.
by Ukv 2y ago
I'd recommend checking out numba ( https://pypi.org/project/numba/ ) - where relevant, it can give huge speedups for minimal effort. Just adding @numba.njit above the factorial function gives me a 36X speedup,
455.
by Ukv 2y ago
I agree - but a persistent identity (which is what the comment I'm replying to is advocating) would be. Need to be careful not to use it as an excuse to create a digital ID system.
456.
by Ukv 2y ago
> All those solutions offer a way for a service-provider to just ask for a service-provider-specific persistent identity Even if working solely as advertised, and somehow implemented flawlessly without compromising a user's right to
457.
by Ukv 2y ago
> the fact that I have no idea (as a non-AI-researcher) what benchmark this is The figure labels it as as "average score [on] 6 multimodal reasoning benchmarks", and the caption notes that the full results are in table 7 - whic
458.
by Ukv 2y ago
The old CSS and HTML logos had identical shape aside from the text. The new CSS and HTML logos have different shape (albeit subtle), larger text, and a greater difference in lightness. Comparison, in monochrome at small size: https:/&
459.
by Ukv 2y ago
If machine learning is found to be fair use, the license you choose does not matter - in the same way Google Books can scan books and make them searchable without a specific license to do so. If machine learning is not found to be fair us
460.
by Ukv 2y ago
> Poems are not just about rhymes and beauty The poems were rated on 14 factors, grouped into emotional quality, structual quality, atmosphere, and creativity. > I've generated a couple of poems by AI, and it never hit a quality
461.
by Ukv 2y ago
> Dear companies, please stop using the term “Copilot” for everything I don't see much harm in "copilot" becoming a generic term like "assistant", but I feel the article doesn't even really do a good job at
462.
by Ukv 2y ago
In some places it looks almost as though they've taken the old tiles and fed them individually through some neural upscale/img2img diffusion. The rocky coast[0] seems to change style slightly from tile-to-tile without much regard
463.
by Ukv 2y ago
> Meta’s “FungiFriend” chatbot showed up in a popular Facebook mushroom foraging group, and soon enough, it advised users to cook a toxic mushroom in butter. Yes, you read that right. To my understanding, sarcosphaera coronaria is edible
464.
by Ukv 2y ago
> if you’re storing arbitrary text from untrusted sources you will always need to clean it first Reversible escaping of characters is pretty common (though not always ; length-before-text formats don't require it). But to "cle
465.
by Ukv 2y ago
> yes, you would still need to clean your inputs before randomly adding it to your table. Lossy is fine in some cases, but in many cases you do actually need the specific text you're trying to store - not just something similar to i
466.
by Ukv 2y ago
> CSV does not have anything like that. If you're using a fallible CSV dialect, your application does need to handle that case in some way (or in some cases it may be fine just to let it crash). Something like length-before-text is
467.
by Ukv 2y ago
> But there is never any reason to use these characters literally, they are just delimiters. I can put []* in my comment (maybe because I'm demonstrating the format, referencing the characters, or just being capricious), and now
468.
by Ukv 2y ago
> March 2022. Cognitive scientist Gary Marcus publishes a scathing article: “Deep Learning Is Hitting a Wall.” [...] In March 2022, no less! Was he out of his mind? That’s what the experts said back then—amid laughter and barely conceale
469.
by Ukv 2y ago
> but these specific characters are not text. they exist solely to be delimiters. Even if people used them only for their intended purpose, someone could use them as delimiters within the text you want to store (e.g: list of tags in file
470.
by Ukv 2y ago
For machine-readable formats, I'd argue length-before-text is simpler to parse than splitting by separators - even before having to add extra application logic to handle this method being fallible (e.g: if you want to use it to store f
471.
by Ukv 2y ago
If you need a machine-readable format, why not go with escaping like most other formats, or length-before-text, to include all characters - instead of a format that fails on some (albeit rare) characters?
472.
by Ukv 2y ago
> Not if you just say those characters are invalid data The author's claim is "with no restrictions on the text". It's easy if you can forbid certain characters, but then you can't store arbitrary text (e.g: file
473.
by Ukv 2y ago
> with no restrictions on the text in fields or the need to try and escape characters. Maybe I'm missing something, but wouldn't it still need escaping for those ASCII separator characters (or alternatively, a restriction for t
474.
by Ukv 2y ago
> Not necessarily. You claimed asking for clarification requires understanding, and then agreed my example showed it asking for clarification. That does logically necessitate understanding (if ¬U ⇒ ¬C, and C, then U). > It would be mu
475.
by Ukv 2y ago
I'm relatively optimistic, on the grounds that: 1. Diminishing returns on a model's capability with respect to scale means that, even if big tech datacenters grow at a faster rate than infrastructure available to individual resear
476.
by Ukv 2y ago
> typically only occurs in very simple cases Would you agree then that it has understanding of at least these simple cases? Personally I think the fact ChatGPT/etc. don't ask for clarification more often is largely just down to
477.
by Ukv 2y ago
> > Do you believe that asking for clarification on ambigious questions requires understanding? > Absolutely! Recognizing ambiguity requires a deeper level of cognition and analysis [...] Then if asking for clarification requires
478.
by Ukv 2y ago
> It is applying symbolic precedence rules to prose language. These rules are not universal and certainly not when writing in prose form. You have to apply some form of precedence in order to evaluate it at all. Question is whether to us
479.
by Ukv 2y ago
> Yes. It shows how easily an LLM can misinterpret subtlety and ambiguity in human language. Without parentheses or context, -42 is a perfectly reasonable answer to "one plus one times forty five minus eighty eight", to the ext
480.
by Ukv 2y ago
> Reason? Clearly no. It doesn't understand what you're asking well enough to "reason". If you ask something a question ("Could a car fit in a bread-box?"), it maps from words into concepts in its internal s

More ›