gallerdude

Neural networks feel like a type of strange magic, like a hack that shouldn’t work but definitely does. This helps reduce that feeling, in the best way possible.

I think is a completely reasonable take, and I'm surprised you're being downvoted for it.

What the heck does this use to search so fast? We use Elastic at work, and for 100K entries, it crawls compares to this.

I thought I heard it applied even with proportionally equal bitrate, but that could be incorrect.

I've heard before that watching 4K video on a 1080p display is still better than watching the same video in 1080p - is that because there's a kind of supersampling going on there?

10 years ago, you could have said that the existing auto industry was better positioned to electrify their existing vehicles than Tesla is to bootstrap their own electric cars out of nothing - but here we are.

I'm curious what they've figured out that generations of scientists have missed. More power (ha) to them, but I'd be surprised if the solution to this infamously hard problem is just 3 years away. It's like civilization leaving a few trillion dollar bills on the ground.

I just started a new blog [1].

I can't explain why, but the whole thing is kind of terrifying. I don't want to become one of those blogs you find that have 300 articles since 2008. The kind of blogs only have 1 or 2 comments on each post, screaming into the void what's going on with their life.

Because of this fear, the only way I ever did write something was basically to force myself to have something posted by the end of the weekend. But now I have the same fear again, so I'll probably have to do the same thing.

[1] https://www.museum-on-the-coast.com/

I don’t know how the UX works, but it seems like it de-activated itself when it got closer to the pedestrians? Or was that a manual override?

Thanks!

I love metaculus! I also probably check PredictIt [0] daily. In 10 years, since I have all of these listed as probabilities, I can check my Brier Score [1]

[0] https://www.predictit.org/

[1] https://en.wikipedia.org/wiki/Brier_score?wprov=sfti1

I hope both my probability is right and you’re probabilities are correct, leaving us with a >70% of all of above

Not a lawyer, but couldn’t you make people sign a waiver before activating the feature?

> You could even be an unemployed alcoholic billionaire with a PHd who just got divorced from Zendaya

My initial draft had a line really similar to that, great minds… Also, all things considered, I think that’d be a better than average outcome. Alcoholism and heartbreak can be overcome.

Thanks! And I equally wish you luck on whatever improbable yet possible dreams you may have.

I don't even see how this could scale to a whole University's worth of students.

The current population of U of Oregon is 23,600. If you have say, 10 administrators dedicated to monitoring social media, that means each have to sift through ~2000 students posts each and every day. Yes, some students don't have any social media, but a lot of them are on multiple social networks.

I wonder if the beaver was ok afterwards.

I’m aware, that’s why I said this will remove the bottleneck :)

This is huge. From what I’ve read, a lot of neuroscience is bottlenecked by having a hard time reading neurons through the skull. This will remove the bottleneck in whole new types of brain/mind/consciousness research.

  ...would be like bringing mathematics to Mesopotamia.

Can you expound on this? What sort of breakthroughs are bottlenecked by developments in quantum computing?

I’ve always thought it’s wild how we can apply one concept to so many different types of things. For example, if I say something is “soft,” you probably think of the opposite of firmness. But at the same time, I can describe a person as “soft,” and the same descriptor can say something meaningful about their character.

Seeing the Spider-Man neuron work on multiple types (pictures, drawings, text), makes it seem like we can teach AI to learn these same type connections.

And if we scale up the network size enough, what if we could see these types through the equivalent of a being with 1000IQ? What connection types are the most effective for a being like that? Can we even understand them? Maybe they would be deep, and archetypical in the way that Odysseus and Harry Potter are the same, despite the fact that one is an ancient Greek king, and the other is a modern British wizard. Even more interestingly, maybe the connections would be completely inexplicable to us, with no apparent rhyme or reason perceptible to humans.

Wait - the island of stability can be for more than a year? That's mindblowing. I wonder what the properties of these huge elements are.

What is the name of the paper?

I don't understand why the other comments are being so negative, this is awesome. There's a Jony Ive quote along the lines of "when you build something, you get the actual thing, but even more valuable is how you've grown and changed from building that thing.

Best of luck on your future projects, great on you to make awesome stuff and grow as a person!

You don't need to be so negative.

Off-topic, is there any mid-level 3D API that I can plug vertices into directly? Everything that I see is either super-low level (Vulcan, OpenGL) or super-high level (Unity).

If you haven't seen any of his posts before, gwern is a treasure. He delves so, so deeply into topics that he finds interesting. A modern Library of Alexandria.

Another problem is that movies with high average rating can also be really generic. A lot of my favorites are controversial, you either love them or hate them: Lost in Translation, The Lighthouse, even Frances Ha to some extent.

By sorting on average you miss the most unique.

Oh interesting, I didn't know the Nvidia encoder was regarded as trash. What tools can one I use to evaluate visual quality of a video file? My proxy for quality has always been bitrate, but I know that bitrate is just chasing visuak quality anyways...

You're basically arguing a techno-religious Pascal's Wager, so my contra-argument is the same: if the nonzero probability is a 0.1% chance, would you really risk your one and only life dedicated to something that has a 99.9% chance of not happening, whether it be real or techno-heaven?