oofbey

41 ms·

541.
by oofbey 5y ago
Sounds like it gets the worst of both worlds? The difficult training of a GAN with the slow runtime of a diffusion model.
542.
by oofbey 5y ago
The phrasing "I don't get it" is fairly rude - it implies the post is obvious to the point of not being worth mentioning. However obvious this might seem to somebody, I would point out that turning AI research papers turn in
543.
by oofbey 5y ago
Diffusion models seem like they're poised to completely replace GANs. They obviously work super well, and you don't have this super finicky minimax training problem.
544.
by oofbey 5y ago
One thing I find really interesting about the DALL-E-2 is that the popular blog name ("DALL-E-2") never shows up in either of the research papers that describe it. The paper commonly referred to as DALL-E-2 calls its own algorith
545.
by oofbey 5y ago
But their job is to figure the realistic worst case I.e fastest plausible timeoine. Not the most likely / expected time it will take.
546.
by oofbey 5y ago
Why would somebody use pinecone instead of a high quality open source library like FAISS？
547.
by oofbey 5y ago
You assume everybody who understands crypto thinks it’s awesome? Dead wrong.
548.
by oofbey 5y ago
Yep. The author, as a peddler of recommendations solutions, has an incentive to convince people that this problem is very complicated, and they should hire a consultant. In practice, good old Matrix Factorization works really well. Can yo
549.
by oofbey 5y ago
Those who used to do no evil, but gave up on the idea as not profitable enough.
550.
by oofbey 5y ago
Yeah, Meta will likely stick because Zuckerberg and crew are actively trying to run away from the dumpster fire they lit with Facebook. But Google is still Google, and probably always will be. Just like Youtube is still Google, and Waymo i
551.
by oofbey 5y ago
> They pioneered using the cloud to run at massive scale. Which is to say they were AWS's biggest early customer? Doesn't really seem like Netflix should get the credit for that one.
552.
by oofbey 5y ago
No question they've done some things that have had some impact on others in the industry. But none of them are particularly important. It's all relative. Companies like Twitter, Uber, AirBnb have all released open source projec
553.
by oofbey 5y ago
Not MAGMA?
554.
by oofbey 5y ago
Off-topic, but how did Netflix manage to get itself inserted into the FAANG acronym anyway? Their impact on the tech industry is trivial compared to all the others. Sure, if you just take out the N it's offensive, but we could have s
555.
by oofbey 5y ago
The mantra “it’s not a choice” was a valid counter point to hideous attempts to “cure gayness.” But it never sat well with me. There’s nothing wrong with being gay, so why should there be anything wrong with choosing to be gay?
556.
by oofbey 5y ago
I'd argue that every one of those tasks is better solved with a neural LM. The fact that people still use traditional techniques does not mean those old techniques are better. It just means those tasks haven't caught up to the m
557.
by oofbey 5y ago
You selling buggy whips? Good luck with that.
558.
by oofbey 5y ago
Why is that tragic? Classic NLP tasks are IMHO kinda pointless. Nobody _actually_ cares about parse trees, etc. These things were useful when that was the best we could do with ML, because they allowed us to accomplish genuinely-useful N
559.
by oofbey 5y ago
Evolutionary theory says that it would not be preserved if it wasn't being used. If it was never read or relied on, then random errors would accumulate dramatically faster than in coding DNA. Because in coding DNA mutations generally
560.
by oofbey 5y ago
The term "junk DNA" was coined very early in our understanding of DNA. Even when we had no idea what it was for, very few respectable geneticists actually believed it was "junk" - basic evolutionary theory argues pretty
561.
by oofbey 5y ago
Fundamentally I don't disagree. Except that I see so much wasted effort in our industry trying to evolve just for the sake of enjoying trying new things, which I find fundamentally super inefficient. I know that's how evolution
562.
by oofbey 5y ago
I remember when there were many different reasonable options for source control. I’ve really enjoyed recent years how there’s at least one thing in software development where basically everybody agrees what the right solution to a problem
563.
by oofbey 5y ago
Having a user app like Firefox force autoupdate is not that bad IMHO. Having server infrastructure components update themselves automatically can be disastrous. This caught me by surprise, painfully. Snap for GUI apps, okay, fine. Snap f
564.
by oofbey 5y ago
I have been a huge fan of Ubuntu for a long time. But the way they push snap is really making me question this choice. The more I learn about snap the less I like it.
565.
by oofbey 5y ago
I'm genuinely confused why you take such an extreme position on this issue. You seem like you understand some things about how neural networks operate. So I'd assume you understand their ability to interpolate between examples t
566.
by oofbey 5y ago
The top-end datacenter GPUs have been slowly creeping up from $5k a few generations back to about $15k for the A100's now. So this one will probably continue the trend, probably to $20k or maybe $30k but probably not beyond that.
567.
by oofbey 5y ago
NVIDIA continues to vertically integrate their datacenter offerings. They bought mellanox to get infiniband. They tried to buy ARM - that didn't work. But they're building & bundling CPUs anyway. I guess when you're s
568.
by oofbey 5y ago
I appreciate that you've analyzed GPT3's inability to rhyme rhyme well. And maybe some of that comes from BPE. But saying that "GPT3 _cannot_ rhyme" is a strong and unsupportable statement. What, really, would the dif
569.
by oofbey 5y ago
I love taking these deep tech dives to appreciate how much thought and complexity goes into something seemingly as simple as random numbers. I know they're critically important for security - which truly is high stakes - so I'm n
570.
by oofbey 5y ago
But did they do anything new? Or did they just repeat a bunch of research others have already done into dehumidifying water extractors? I honestly don't know. A lot of the comments here make me think it's the latter.

More ›