rf15

28 ms·

511.
by rf15 3y ago
Pride coupled with politeness I guess: you don't want to look like a cheap imitator or impostor, so you write everything yourself.
512.
by rf15 3y ago
Care to elaborate?
513.
by rf15 3y ago
The company name alone already tipped me off. It's all very ridiculous, and yet posted here.
514.
by rf15 3y ago
The sad truth is that all of this was very predictable. It is also very convenient for Netanyahu, whose family[0] and government coalition has some interesting views about the area as a whole[1], including who Jordan belongs to. One could a
515.
by rf15 3y ago
I have been using github copilot for a while now, but when I first tested these, together with the code AI offering from Amazon, they were... kinda useless. Only after a few updates did it get to a point that was a net positive in my daily
516.
by rf15 3y ago
Yes, but if you're looking for code completion at the quality level of Codex/Github Copilot, that's still a bit off.
517.
by rf15 3y ago
Give it another year. In theory it's not going to be difficult, but sampling speed/power consumption is prohibitive for normal hardware at the moment.
518.
by rf15 3y ago
In my experience the vastly oversimplified-beyond-actual-usability calculation is that you take the parameter number (34) and count that as the amount in GB you'd need for an 8 bit model. with a 4-bit quantization that's halved, s
519.
by rf15 3y ago
I think that's covered by "[...] and has become a fixture in the lower-left corner of laptop keyboards ever since." ? But yes, it seems fairly Mac-centric.
520.
by rf15 3y ago
Sadly the current normal, and has been for some time. It's sad to see an educational institution deal so wholeheartedly in intentional and active disinformation.
521.
by rf15 3y ago
Turbo Pacal! It's been ages, but it's the first language I learned. The Borland IDE was amazing for its time and allowed me to learn the language in no time. My dad learned it too, and even developed some small games with it. Even
522.
by rf15 3y ago
I work in the field. It's just not how text-token-based autoregressive models can ever work. I can't talk about my work of course, but even a quick glance on Wikipedia can tell you they'd need to be at least a symbolic hybrid
523.
by rf15 3y ago
Yeah this strikes me as half generated, especially since it's three different poses. This is the baby teeth thing all over again and it must suck knowing that your Wedding photos are just an AI's reinterpretation of the moment.
524.
by rf15 3y ago
they can't reason though, sadly - the premise does not hold.
525.
by rf15 3y ago
What kind of approach do you see here? Yes you can create an infinite amount of samples to test on, but you still have to deal with the millions of copies in the wild.
526.
by rf15 3y ago
I have to use SQL like 1-3 times a month and can confirm that the syntactic inconsistencies of the language make me go to the cheat sheet regularly.
527.
by rf15 3y ago
Thanks for pointing out the Spring connection... I wonder how that will impact their development.
528.
by rf15 3y ago
Our Brother printer and sewing machines serve us well. After years of Canon, Samsung and others everyone in my family settled on brother products for their reliability and because they forego the usual anti-consumer antics of the other manu
529.
by rf15 3y ago
As someone who doesn't take any form of caffeine: Considering how widespread it is, it's probably just another of those trade-off scenarios for your body.
530.
by rf15 3y ago
Yeah, basically. Just with a lot of fancy math behind it. Making it all the funnier that it's so prompt/chat focused, since you can just write matter-of-factly to steer large language models.
531.
by rf15 3y ago
Even by the standards of blind corp greed, that's a bad move - you want people to know your platform when they're done studying, so they can advocate for it in their jobs. Why would you destroy this revenue stream? Unless they hav
532.
by rf15 3y ago
Why would someone want this? It seems a bit of an odd target.
533.
by rf15 3y ago
But what are you going to do if your ambitions cannot be met with the money and resources they require, especially in the long term? They look like they're tripping over hard limits.
534.
by rf15 3y ago
This entire show will damage Altman and what remains of OpenAI equally - OpenAI because of its unprofessional way of dealing with internal problems, and Altman, well, there's apparently a reason they were unhappy with him.
535.
by rf15 3y ago
I do interviews for my team, and I'm very aware that framing questions the wrong way quickly leads people into the trappings of their own obsessions. Be that a need for optimising the wrong things (e.g. optimising minute but easy to op
536.
by rf15 3y ago
I kind of agree, the track record of people obsessed with BigO is not good in my book. The last person that valued it would still get the smallest element in a collection by... sorting it and then selecting the first element.
537.
by rf15 3y ago
> I don’t like hard or tricky questions. ... > I’ve actually expressed the problem in an ambiguous way. So when other people do it, hard and tricky questions are bad, but when you deliberately set your candidate up for failure by with
538.
by rf15 3y ago
I'd assume that you have a lot more leverage over a random small startup than OpenAI who serves half the world. This leverage can be used to ensure better privacy/etc..
539.
by rf15 3y ago
This and Apache Commons is the jQuery of Java - once a useful tool to get cutting edge features and entire paradigms into older versions, now a security risk and surprisingly unperformant compared to what the standard library has to offer.
540.
by rf15 3y ago
On the other hand there's an entire industry around selling placebos of various shapes and sizes; and even the meds that are functional often ruin your kidneys and other organs over time. Medication, while useful in many cases, is ofte

More ›