"People say, It’s just glorified autocomplete ... Now, let’s analyze that. Suppose you want to be really good at predicting the next word. If you want to be really good, you have to understand what’s being said. That’s the only way. So by training something to be really good at predicting the next word, you’re actually forcing it to understand. Yes, it’s ‘autocomplete’ — but you didn’t think through what it means to have a really good autocomplete." - Geoff Hinton
He could possibly stay in a pure Chief Scientist role while abdicating his board seat. But if I were a CEO, I'd have a hard time trusting a C-level role to someone whose vision is diametrically opposed to my own.
FB employees used to be able to help with these types of requests, but unfortunately some time ago they changed their policy and an employee may only help their relatives or close friends, and must personally vouch on their behalf (this was true a couple of years back, not sure if anything changed). Basically, that loophole is closed for all intents and purposes :(
#3 is actually unlikely to be effective due to the well-studied bystander effect (everyone stands around hoping someone else will help). Asking a concrete person is much more likely to get it done.
And yet the Metaverse is the punching bag of any discussion about Meta, despite Zuck throwing the weight of an almost-trillion-dollar company behind this vision. Perhaps a whim is not enough? And if the execution is excellent and the vision takes off, sounds like it does solve a problem, so what's the issue? If the choice is between spending one's billions on luxury items or pushing the envelope in some fantastical direction, I'll always vote for the latter (though I am a fan of Gates' approach to philanthropy as well).
Zuck's Metaverse may not be solving a problem you face, but it is solving one for Zuck (trying to, anyways) - controlling a distribution platform where potential competitors like Apple or Google cannot dictate terms. SpaceX and Starlink are solving plenty of real-world problems even if "get to Mars" is the long-term vision (and one I'm frankly thankful for, as little as I like Musk). LLMs are clearly a distinct phase in AI evolution, even if it's too early to talk of AGI - real-world applicability here is a bit more tenuous, but I certainly wouldn't throw it in the same bucket as, say, crypto. I think this idea that these explorations at the edge are a waste is a form of Luddism. Countless innovations that _do_ benefit humanity (including allowing us to have this conversation) have sprung out of them.
Science fiction: a genre of speculative fiction, which typically deals with imaginative and futuristic concepts.
Given sci-fi authors regularly speculate about all sorts of potential futures (and presents), is it really that hard to imagine that some of these futures might come true in some shape or form? It's really rich to trace modern implementations to decades-old ideas and then claim you were the one to create them. Get to pat oneself on the back while wagging a finger at the folks actually doing the building. Zuck's Metaverse would've likely existed without Snow Crash, though probably with a different name. And Motorola Razr would have likely existed without Star Trek communicators. I am not saying that all modern innovations are good ones or that some inspiration didn't take place, I just find the argument that their existence is entirely rooted in science fiction a weak one and full of survivorship bias.
I just discovered him yesterday and ended up watching something like 20 videos last night. This was also the first thought that came to my mind when I was reading the announcement.
A system of morals has nothing to do with religion, no matter how much the religious folk like to claim that they are inextricably linked. So no, absence of belief is not religion-lite.
I'm terrible with design, but what I went with was Tailwind + commercial TailwindUI library, which was more than worth the money. Got many compliments on my app's UI, which is funny considering how non-intuitive I find design to be.
I mean, it's not like that sentiment is totally baseless. I've used PHP years ago and both language and core libraries were not well-designed and came with many footguns. I've heard that modern PHP with frameworks like Laravel fixed a bunch of this, so presumably it's quite capable these days. But PHP certainly wouldn't be anywhere near the top of my go-to list at this point.
Doh, you're absolutely correct. I started with .NET back in 2003, so there's still quite a bit of baggage mixed in with the .NET team's terrible naming history.
I would personally recommend .NET for backend work, as MS track record of UI frameworks is rather suspect. For the main project I currently support, using .NET 6 API and a React frontend (+EF Core to access a Postgres database). Works beautifully.
Fantastic post-mortem that I regularly share out with others. A great cautionary tale on the importance of backup testing procedures (aka why "set and forget" with backups is not sufficient).
This matches my experience as well. When I was making the choice for switching my team from SVN to a DVCS ten or so years ago, it was not at all obvious which option was better. Went with Mercurial because Windows support was better.