Macrochips

Not sure why you are downvoted. Reolinks work without internet and can stream locally using rtsp. I have a doorbell cam from them and it works fine. If you block it from the internet you only get video and basic doorbell functionality though, which is fine.

You can also stream Reolink video to a Frigate box (free software) and access that from wherever.

No cloud involved if you run your own VPN.

Or even better, _just a little_ cloud involved if you expose through Cloudflare tunnel for just you and whoever else needs to access it.

Given that cloudflare tunnels require ssl termination, this technically isnt much different to a cloud server that promises not to record. Of course, cloudflare are less likely to be inclined to analyse it

The difference if you can film out on the street (random people walking by) or just your property (my own fault if I approach your door).

Some countries have laws for this, as in you can only point the camera so that you don't catch everyone. This can have downsides, e.g. if you have no (even short) front yard and your (organization's) door is directly on the curb - but I completely agree that this is just tough luck, the privacy of random people walking on public property past private property should not be filmed.

Wouldn't you deal with spatial reasoning by giving it access to a tool that structures the space in a way it can understand or just is a sub-model that can do spatial reasoning? These "general" models would serve as the frontal cortex while other models do specialized work. What is missing?

That's a bit like saying just give blind people cameras so they can see.

I mean, no not really. These models can see, you're giving them eyes to connect to that part of their brain.

They should train more on sports commentary, perhaps that could give spatial reasoning a boost.

A few things I can’t stand about Youtube’s desktop website:

1) Spacebar sometimes skips to the next video when playing a playlist. Just why?

2) You never know if the small buttons like play next on the thumbnails will work or just play that video right away.

2) when on the homepage, you open a few videos in new tabs and close the homepage only to find out that you just open bunch of “this video contains paid promotion” disclaimer pages. Re-open the homepage to actually open the videos and they are all gone, the page shows a grid of different videos.

so yes, I agree that the web interface went downhill.

I am using it less because I can't find any videos as the search is completely broken (no, I won't enable history tracking). Many days when I am in the mood to watch videos, I just give up. It doesn't help now that most searches return AI-generated video (that has millions of views). Who watches these things?

glad it’s not just me. we used to have multi-playlist saves with a modal that showed more than three pixels of your library. now it’s slow, cramped, and forces you to hit SAVE over and over even though the backend supports it. feels like a regression dressed up as bad UX.

Not a fan of asking to log in to continue after the user has entered data. It's a dark pattern that indicates you are trying to inflate sign up numbers rather than just making your site sticky enough that people want to sign up organically.

In our case, in order to run the queries we need to provision a sandbox and connect to your data sources to give meaningful answers, unlike a general purpose chat. We need to have some authentication to prevent abuse here, but the product itself has a free tier so you can use it without paying or needing a card.

Fair enough but the examples shown could surely just display some pre-cooked examples to give a demo of how rhe product works with no real cost to you or barrier to potential users.

True! Tried to do that with the video + the example prompts but definitely could use improvement. Thanks for the feedback

Even after creating an account, it didn't even finish a single request before it ran out of credits. I'm not going to pay to try it out.

We include $5 in credits for every new account, after which you can pay-as-you-go for credits. I can drop some more credits in your account so you get to try it out fully, let me know what's the email (I'm a at livedocs)

"Discord alternatives aren't needed because I'm different from the majority of millions of people that use it differently than me."

is this demo running fully in the browser?

No, it's server-side.

Model is around 7.5 GB - once they get above 4 GB running them in a browser gets quite difficult I believe.

Because it's a 4gb download?

I think that web browsers only allow up to 4GB of memory per tab.

Headscale mostly works pretty well but its pretty finicky to get set up in a way where the tailscale clients on linux and android aren't always complaining with warnings or having route or DNS issues. I'm considering investigating one of these non commericial solutions where the entire stack was built to work together.

> So yes, currently playing PS2 games on PC via emulator is still absolutely fantastic, but native ports would be the holy grail of game preservation.

I would think that emulation of the original game as closely as possible would be the gold standard of preservation, and native ports would be a cool alternative. As described in the article, native ports are typically not faithful reproductions but enhanced to use the latest hardware.

Indeed, the focus for preservation would be to increase the accuracy of emulators.

pcsx2 is pretty good today in terms of running games (there is a single digit list of games it does not run), but it's far from accurate to the hardware.

Porting to current systems via recompilation is cool, but it has very little to do with preservation.