Frigate's docs has some detection speeds listed: https://docs.frigate.video/frigate/hardware/#openvino .

I believe it also has an advantage of being able to run bigger models like YOLO-NAS. Going off Frigate+ documentation: https://docs.frigate.video/plus/#supported-detector-types

I think the RPi 4 is still competitive. At ~40$, Orange Pi's boards are Allwinner H618 and Rockchip RK3566 based boards. I do have an Orange Pi 3B in use because it has an nvme port, but the RPi 4 is generally faster.

The Raspberry Pi 5 and Orange Pi 5 are just too expensive. I do have some of the Pi 5s from both these companies, but have replaced them with Intel N100 mini-PCs instead. But I'll still use RPi 4's for my 3d printers and other lower-end uses.

And RPi's software support is just better: I've got an Orange Pi 4. Orange Pi hasn't updated their OS for it in years. Last time I tried to get it working in Armbian, HDMI output was (is?) broken: https://forum.armbian.com/topic/26818-opi-4-lts-no-hdmi-outp...

Orange Pi also has Rockchip RK3588 based boards.

RK35XX only worked well because volunteer maintainers like Joshua Riek were doing hard work, the support of those boards falls when people like him burn out.

For me personally, much of the value of the RPI products is knowing they will have long term support.

RockChip really only appears to care about embedded auto apps etc...thus their support horizon is more about sustainment and not enhancement, and they do little to support the community.

Where can I download the documentation for the BCM2712 in the RPi 5?

I can get this for the RK3588 though there are problems with my Orange Pi board that uses it.

https://pon.wiki/ has a few bypasses for AT&T and other fiber ISPs. Unfortunately a lot of the info isn't on the website, but in their Discord server instead.

I think he means self-hosted on your own hardware, rather than requiring their Unifi Protect appliances. You used to be able to with their old Unifi Video .

Unofficially (and therefore not supported), you can just take their binaries and run it on your own hardware. But since they only support their own hardware and it's arm64 hardware, you only get arm64 binaries: https://github.com/dciancu/unifi-protect-unvr-docker-arm64

With them adding generic onvif camera support, I might try it again because the other options aren't amazing either.

It's also funny that the Lightning to HDMI adapters had to add hardware to decode a compressed video stream from the Lightning port.

It's basically a wired AirPlay adapter. That's why they cost so much.

Lightning was meant to be a software-defined serial data interface, I suspect because of the confusing breaking mess that was the 30 pin adapter over its lifetime. The digital logic to work with hardware was pushed outside the phone into the cable.

I think it's also notable that the iPad Pro never supported usb 3 client either. The iPad Pro had a separate USB3 host controller, and so only supported that camera adapter. There was never a USB-A (or C) 3.0 to Lightning cable.

I use mergerfs too. I ended up with all the drives under mergerfs running on bcache with a single SSD as the cache.

Mergerfs's documentations brings up ways to do tiered caching: https://trapexit.github.io/mergerfs/preview/usage_patterns/ . Though it doesn't really do a read cache. It only keeps the newest files on the SSDs.

I hope bcachefs (not bcache) gets better. It supports combining multiple different slow drives like I do under mergerfs, and also a file level cache on SSDs. And even supports keeping metadata on the SSDs completely.

Erasure coding isn't done yet either, so that's no replacement for SnapRAID.

There is Box86/box64/box32. A bit confusing but it's box86 for x86 > ARM32. Box64 for x86_64 > ARM64. And Box32 for x86 > ARM64.

I think they mention games because a lot of other software for Linux is generally open source. So a lot of times it's pretty easy to get an ARM build.

It also does a neat performance trick where it intercepts library calls and redirects them to native versions of the same library.

http://github.com/ptitSeb/box64

In a similar use, a few of my hobby projects are hosted on static web servers. So instead of rendering out everything into individual pages, which would've been tens of thousands of pages, I have a JSON file that gets rendered by the client in a SPA. I have even used Github Pages for this

I'm playing around with a newer version that uses a sqlite database instead. Sqlite officially has wasm builds, and the database file is already built to be separated into pages. With HTTP Range Requests, I can grab only the pages I need to fulfill any queries.

Sqlite full text search even works! Though I'm hesitant to call that a success because you do end up grabbing the entire FTS table for shorter searches. Might be better to download the entire database and build the FTS table locally.

It is not monopoly laws, it is anti-competition laws. It's not anti-competitive to have a monopoly from a superior product:

https://www.ftc.gov/advice-guidance/competition-guidance/gui...

> Obtaining a monopoly by superior products, innovation, or business acumen is legal