daenz

63 ms·

571.
by daenz 4y ago
I can see that perspective, but I don't think that will happen. Companies jump at any chance to steal each others customers, so the necessity threshold has to be very high for something like this to be adopted across the board.
572.
by daenz 4y ago
Reminds me of the AI that tattoos "Not Sure" on the protagonist in Idiocracy. I'm glad people are highlighting how inadequate these models currently are in playing any kind of non-trivial role our lives.
573.
by daenz 4y ago
Doesn't hurt to ask for written approval from your manager to write about it! I'd love to read it as well. Getting to use optimization algorithms at your job is super rare (unless you're in specific industries).
574.
by daenz 4y ago
I wonder if 40mph is a magic number that is a good tradeoff between speed and survivability (and injury-related out-of-court settlements)
575.
by daenz 4y ago
Are there any downsides to messing with your gut bacteria? I seem to see a strong bias of positive studies, but if our gut bacteria is as important as it seems to be, shouldn't there also be horrible outcomes to these kinds of experime
576.
by daenz 4y ago
I waste time on the internet when I am "compiling" My thoughts. Being in knowledge work, like many others here, means that most of the work is performed by thinking about a problem. After the problem has a good solution or set of
577.
by daenz 4y ago
The article is in full damage control about how the data was anonymized. But they seem to be pretty squirrely about whether or not an "anonymous" device id showed specific locations and routes. Even if an anonymized report for a d
578.
by daenz 4y ago
I'm sure this book is an excellent resource (upon skimming the online version, it looks very thorough and well thought out). I think there is some confusion about who "the masses" are though. I don't think a subchapter t
579.
by daenz 4y ago
I've enabled this "grab" feature on every Linux window manager that has supported it. Hold alt or the windows key, click, move. Simple usability improvements make a world of difference, but for some reason, they are uncommon.
580.
by daenz 4y ago
I got 2 actually, one for me and my partner, to see if there was any big differences. This is mine https://www.amazon.com/gp/product/B07V7F8QYK It's larger. No real complaints other than you can only tip it u
581.
by daenz 4y ago
It would be nice for Amazon to commit as well. AWS has support for only a single Yubikey, which is mostly useless, unless you don't care about being locked out of your account if you lose that one key.
582.
by daenz 4y ago
Can you say more about why you are seeing the process as hacky?
583.
by daenz 4y ago
It's been awhile since I've done game engine work, but is this impressive? The first thing that comes to mind is they're using instanced rendering. This allows the CPU to only deal with 1 sprite, while telling the GPU to rend
584.
by daenz 4y ago
So...it only works for 99% of the population?
585.
by daenz 4y ago
I got a portable treadmill in June 2021. I put it under my standing desk and walk at a slow/medium most of the day while I work. I'm at 723 miles so far.
586.
by daenz 4y ago
You would be surprised. I worked with a fellow that wouldn't acknowledge that there is a height difference in the average man's height and the average woman's height, because "everyone is different and there's no su
587.
by daenz 4y ago
I can think of 4 things that would make a science-based discussion around those sensitive topics more palatable: 1) All data is in distributions, and so no predictions can be made about an individual based on distributions. 2) "Better&
588.
by daenz 4y ago
>Pits can bite and shake because they have ridiculous jaws and a powerful build. If a miniature poodle tried, it wouldn't have much effect. They got that way because humans designed them that way intentionally with selective breedin
589.
by daenz 4y ago
What you describe goes against the narrative that any living thing in a species is just as X, Y, and Z as any other living thing in that species. It's part of the "equality of outcome" push because it denies the presence of i
590.
by daenz 4y ago
The point is that it's possible and we're improving on it every day.
591.
by daenz 4y ago
Can you identify GPT text versus authentic text? If so, then there are features in that text that give it away. It stands to reason that there exist other features in the text, based on the training data the model was fed, and other charact
592.
by daenz 4y ago
>That being said, I just don’t see where the ROI would be in something like that. Seems like a lot of expense for no payoff. I consider these types of models as information weapons, so I wouldn't be surprised if they have some contr
593.
by daenz 4y ago
To be clear, I wasn't intending to come across as attacking voz, only pointing out that I don't think anyone "in the know" at Meta/Facebook would admit to it even if they were doing it, so hearing "This is nons
594.
by daenz 4y ago
I agree it is not exactly the same as a human, but the content it produces is based on its specific training data, how it was fed the training data, how long it was trained, the size and shape of the network, etc. These are unique character
595.
by daenz 4y ago
In the same way we can train models that can identify people from their choice of words, phrasing, grammar, etc, we can train models that identify other models.
596.
by daenz 4y ago
So in the entire field of machine learning, we can't train a model that can identify another model from its output? Just can't be done? And there's absolutely no value in having tools that can identify deep fakes, or content
597.
by daenz 4y ago
>You cannot fingerprint models like this A GAN can absolutely be trained to discriminate between text generated from this model or another model. >that's hilarious What's hilarious about it?
598.
by daenz 4y ago
Would an AI @ FB employee admit it if it was true?
599.
by daenz 4y ago
By training a GAN. A trained GAN will be able to accurately guess whether a block of text was produced by this GPT model, some other GPT model, or is authentic.
600.
by daenz 4y ago
It's pretty simple. GPT models are essentially information weapons. People are going to get their hands on them, so might as well give them a model where you can identify content generated with them, so you can know who is using them f

More ›