FileBot Thanks, looks promising. Only problem closed source. I do not mind to pay for license, but do not like running unknown code on my server.

NFS / Shares - over complicated. Just syncthing it around (and snapshot on backup server), for nearly static data as music it is perfect.

albums one by one with --timid flag. You are using Navidrome , than you music should be well tagged already. Should be fast and easy, just repeat it 2000 time. You can also disable lockups and just import “as is” if you trust your taggig

Sorry, not mbuff but mbuffer. It just adding a memory buffer between STDIN and STDOUT. Useful to smooth data flow, so reading from disk do not need to wait for SSH to transmit and other way around. In you case it cold be something like that. ( I’m writing command from memory, do not use it unless checked)

dd if=/dev/sda bs=16M | mbuffer -s 128k -m 2G -o - | ssh remoteuser@remote.host.com dd of=/storage/backup.img

nfs over the internet is usually bad idea. Network delays is killer for NFS. You can use ssh instead. And add some mbuff for better results.

you are mistaken about it. There is a lots of other use for ipfs.

Ollama has long history of exploits. PLease do not feed anything which come from outside to it.

I got some Matrox G400 on PCIx1 for this from ebay. (Yes they still making them) Really cheap. native linux support, very low power consumption.

With AI search engines hosting public repo is very expensive.

Probably. It just not as fast as 9070 XT. I’m using 9070 XT myself and limitation for running LLMs is memory, not speed. If model fit in memory it will runs fast enough to be practical.

If you happy with 16g , nothing can beat in speed/cost of AMD RX 9070 XT.

Some standalone WAPs for WiFi and PC based router. Depends on what you are getting you can get it dirt cheap. WAP also need firmware upgrades, but it is less a problem.

Then read my post again. Contributing and writing opens source is no longer about how much time one willing to spend on it, it is about how much money someone willing to spend on LLMs which will write code. And all these money will go to AI overlords.

Yes, but in each joke there is bit of truth. Open Source have to change. Open Source code written by LLMs is still open source, but it drastically different from current one.
Instead of spending time to “scratch the itch and help others in the process” - now people should give money to corps to use LLM to to do same.

This is one good article. I guess humans are now mostly redundant in open source. Bots can do everything themself, write code, submit PR, merge them and even blog about it. Time to book a place for myself in a graveyard.

for start private keys should never leave the system which uses them. Wildcards are even worse, as if one host got compromised, all others can be spoofed.

Just do not use wildcard, very bad security practice. Getting individual cert for each service is easy these days.

you can’t going wrong with squid. It was around forever, and still in development.

“Maybe because I run Frigate in a virtual machine in Proxmox, so the Coral has to be passed through to the VM? Not sure.” FreeBSD + linux VM under Bhyve . Never had problems.

Why? Why anyone needs online version?