Philpax

You have a device that can present a virtual reality that you can share with people all around the world and you want to use it to visit an office and block out the people there? That seems a little b…

I'm keen for the Vision Pro, but the iPhone and iPod touch had already reached incredible market penetration by the iPhone 3GS. The reason the iPhone 4 leak was such a huge deal is because the iP…

The PlayStation VR was not very good. Modern headsets, including the AVP, are much better at presenting a consistent low-persistence view - you should feel much better with it.

Ah, I see - my apologies, I think there's an unfortunate misalignment over the word "backend." When Pocketbase is talking about a backend, they're talking about something like Fire…

Your initial comment presents Redbean as a comparable solution, which it's not; they serve very different use cases. Redbean requires you to write server code; Pocketbase does not. Redbean does n…

You'll never believe what this article is about :-)

Live map of Tokyo trains: https://minitokyo3d.com/

Nominations aren't awards...?

This is quite a balanced take on the matter. Appreciated!

Er, no, not really? I live in Sweden and a large slice of society take the trains. Ditto everywhere else I've been in Western Europe. Is it possible that your train network specifically is medioc…

That's interesting. I wonder how `Rc<str>` would fare - it should be able to avoid the locking. (Of course, you lose the ability to send it across threads... but that might not matter)

Yes, I suspect so; there's been some work on this already (from a cursory arXiV search: [0] [1] [2], but there are more). I'm also curious to see if graph neural networks can be used for str…

Does he write his own books, though? [0] [1] [2] It's his name on the front of the book, but he is not necessarily the primary author. That sounds a little unethical to me. [0] https://…

This is great! I’m always on the lookout for advancements in decompilation. Looking forward to Part 2 :)

Isn’t this the GamerGate guy?

That’s not the same thing. Perplexity is using an already-trained LLM to read those sources and synthesise a new result from them. This allows them to cite the sources used for generation. LLM trainin…

> I guess the Quest headsets are included? Yes, they've shipped tens of millions of headsets across the world. > It’s hard to figure out what Meta spent all that money on. R&D on every …

Yeah, I think so. The technology will become cheaper, better, more accessible, and they'll be well-poised to take advantage of it.

LLMs are almost always grammatically correct (it's one of the first things they learn); it's their semantic correctness/grounding that's lacking. You would be very hard-pressed t…

“Everything is smoke and mirrors in computer graphics - especially the smoke and mirrors!”

Why do those things not apply to someone using AI? I hate making this kind of argument, but: The choice made by the AI artist is made based on a myriad of factors and intuitions. The result is a snaps…

> Putting some prompts into an image generator and choosing which one you like is hardly the same thing. At a base level, "putting some prompts [..] and choosing" sounds like intent to me…

Not necessarily - their comparison continues to mix in observations of the night sky, and similarly we’d do the same (continue mixing in organic data). That’s not the exciting bit, though - if you hav…

Try Fuyu-8B or CogVLM. Failing that, there’s always the GPT-4V API. There are also other non-LLM solutions for image captioning / tagging (e.g. what the smartphone vendors do), but I’m less famil…

To be clear, this is not using QuIP#, and is thus not particularly relevant to the discussion at hand.

Yes, that's more or less it - there's no guarantee that the chosen expert will still be used for the next token, so you'll need to have all of them on hand at any given moment.

Excited for the next generation of Personas for the Vision Pro, etc :-)