jdiff

23 ms·

541.
by jdiff 1y ago
Even limiting its scope can be risky, if it's a complex problem in a niche that's not well-represented in the training data. Been learning Gleam lately, and yesterday when wrapping my head around recursion was starting to give me
542.
by jdiff 1y ago
At least currently, the only use pattern that can withstand complex codebases is as advanced speech-to-text, but without the speech if that makes sense. The main issue with that is phrasing things in English is often far more verbose, so wi
543.
by jdiff 1y ago
One of the very first examples published by GPT-2, the model that started the "too dangerous to be open anymore" trend, was a fake news article about the discovery of unicorns in the Scottish mountains. It's been part of Open
544.
by jdiff 1y ago
I believe it'd be pretty wonky coloring, or at least it could be, since it'd be capturing snapshots of individual frequency responses. If something is visibly green, reflecting across most of the greenish areas of spectrum, but ha
545.
by jdiff 1y ago
None of that gives me any useful spatial sense of where . "Payment terminal." Okay. Where is it? Left? Left where? How much left? How far? The only truly useful bits I see in your stream of text is, again, "Cream of Mushroom
546.
by jdiff 1y ago
It might be useful for telling Cream of Chicken from Cream of Mushroom, but for locomotion I can't see this adding anything over existing strategies people use to get around sans sight. "There's a tree. There's a tree. T
547.
by jdiff 1y ago
I'd be very interested in that discussion for one of my own projects, any chance you still have a link handy?
548.
by jdiff 1y ago
How about the fortified rice? The salt? The flour? Bread? Milk? Pasta?
549.
by jdiff 1y ago
Where are the studies supporting this claim? A random person on the internet saying "prob higher IQ" is not a convincing argument.
550.
by jdiff 1y ago
There are clear factual errors in the underlying assumptions of what was stated about water fluoridation. Those are simply table stakes for having a discussion about anything at all. If one thinks that water fluoridation is useful "jus
551.
by jdiff 1y ago
The questions you posed are not questioning fluoride, they're asking what the basic premise even is. If you don't understand that, you are far from the position needed to be evaluating and analyzing the necessity or benefits of it
552.
by jdiff 1y ago
"For the few seconds people brush their teeth"? That's not how fluoridated drinking water works. Fluoridated water works all of the time, not just when brushing teeth, and it's not a vital chemical that the body craves.
553.
by jdiff 1y ago
Both can be true. There can be nothing physically or metaphysically "special" about human cognition, and at the same time, we can also be very, very far away from creating even a holistic facsimile. We've got echoes of it i
554.
by jdiff 1y ago
It is a severe reach to say that AI can provide human connection. At best, it can provide the illusion, for some. And for those people, I'm not going to say that that's not valuable or legitimate for them. But it's pushing it
555.
by jdiff 1y ago
There is absolutely no reason to jump immediately to conspiracy here.
556.
by jdiff 1y ago
I sincerely wish that had been the case. No, I built the meat of the code. The most common benefit is helping to reducing repetitive typing, letting me skip writing 12 minor variations of `x1 = sin(r1) - cos(r2)`. Similar to that, in this p
557.
by jdiff 1y ago
My own experience writing a web-based, SVG-based 3D modeler. No traditional back end, but when working on the underlying 3D engine it shits the bed from all the broken assumptions and uncommon conventions used there. And in the UI, the case
558.
by jdiff 1y ago
I've had to yank tokens out of the mouths of too many thinking models stuck in loops of (internally, within their own chain of thought) rephrasing the same broken function over and over again, realizing each time that it doesn't m
559.
by jdiff 1y ago
Not even front end, unless it literally is a dumb thin wrapper around a back end. If you are processing anything on that front end, AI is likely to fall flat as quickly as it would on the backend.
560.
by jdiff 1y ago
I'm not sure that's a convincing argument given that crypto heads haven't just been enthusiastically chatting about the possibilities in the abstract. They do an awful lot of that, see Web3, but they have been using crypto.
561.
by jdiff 1y ago
Capturing HTML as scalable SVGs is huge, how do you manage condensing all of CSS and its quirks into an SVG? Do you only support a subset of styling properties and rely on the browser to calculate layouts for you?
562.
by jdiff 1y ago
As someone currently engaged in teaching the Adobe suite to high school students, that doesn't track with what I see. When my students are getting stuck and frustrated, I look at the problem, remind them of the constraints and assumpti
563.
by jdiff 1y ago
Working on extending the [Zdog]( https://zzz.dog ) library, adding some new types and tooling, patching bugs I run into on the way. All the quirks inherit from it being based on (and rendering to) SVG. SVG is Y-down, Zdog only add
564.
by jdiff 1y ago
My guess is that those questions are very typical and follow very normal patterns and use well established processes. Give it something weird and it'll continuously trip over itself. My current project is nothing too bizarre, it's
565.
by jdiff 1y ago
Definitely matches my experience as well. I've been working away on a very quirky, non-idiomatic 3D codebase, and LLMs are a mixed bag there. Y is down, there's no perspective distortion or Z buffer, there are no meshes, it's
566.
by jdiff 1y ago
That's been said since at least 2021 (the release date for GitHub Copilot). I think you're overestimating the pace.
567.
by jdiff 1y ago
While I have absolutely no idea what the politics of Super Hexagon could be framed as, with regards to the paintings, yes. Having a tree, having a backyard, having a pet (or a cat specifically), these aren't universal across cultures a
568.
by jdiff 1y ago
Tesla's valuation is based on the heady tales that Musk spins. It's not based on the fact that reality falls far short of them.
569.
by jdiff 1y ago
That quote itself seems to be disingenuous, conflating "one successful drug" which costs $1B with "drug candidates in clinical trials." Even more so when taken from context, as the next sentence in the article is > Wh
570.
by jdiff 2y ago
Not the original, but a lossy facsimile that's Good Enough for almost anything. And as the short history of LLMs and other nets has shown us, they're often not even all that lossy.

More ›