I haven't used this tool specifically, but anything involving Youtube is always going to be an uphill battle. In the instance of a video downloader like Channeltube, you frequently need to feed it fresh cookies to make sure Youtube doesn't flag it as a bot. I assume this is a similar situation.

The requirements also list ''2gb if you restart frequently'' so there might be a memory leak issue, or the way it's setup means the RAM usage will increase in perpetuity.

I would just script a task to restart the docker container every hour. Shouldn't be too much a hassle, you're unlikely to ever notice since it shouldn't take more than a few seconds. I would track its RAM usage over a few days too. See if there's something fishy happening. Something like Beszel will let you track your docker socket, so usage per container. Preferably left on read only imo.

Distros are tools for a job. I stick with the ones that have the most helpful communities and/or the best documentation. I avoid communities that tell you off for asking help or documentation that is obtuse.

Still successfully troubleshooting stuff through 10 year old Ubuntu forum threads. Meanwhile learning the most through the FreeBSD handbook and it's not even Linux. lol

Hell yeah! I was using Ubuntu 10-14 in college and now back on the train with 24, 25 and 26. Everything just works every time and all the professional apps I need are first class citizens and well supported.

Ubuntu Server, FreeBSD and OpenBSD are my server distros of choice as well. FreeBSD for a NAS. Gotta pick the right tools for the job!

Any plans to support FreeBSD in the future? Or mainly focusing on a docker solution?

So far the only real vulnerabilities that Ai have spun up seem to be really obscure edge case privilege escalation bugs that have been around for ages but nobody discovered. They then get patched in days if not less. I would expect them to run out of low hanging fruit pretty soon.

Most of them can even be mitigated by preventing local access to a system.

There's tools to audit your network and systems. You can setup a SIEM like Wazuh. It'll tell you the same kind of stuff that have been good practice for a while and mitigate most of the ai assisted vulnerabilities.

Prevent password logins where ever possible too, especially SSH. Use ed25519 keys. Perform server actions with least privilege users, prevent root logins. Immutable systems and containers are a helpful tool too.

I also just host my public facing stuff on a VPS. DDOS protection and just otherwise isolated. 2FA on the VPS too.

I have a directory in my home folder for all my different docker containers. I back that up like my other critical files on my NAS, Backup NAS, cloud storage and external drive. Its all scripted to happen periodically.

The docker compose files I also upload to my local Git so I can version them. Make the git ignore skip everything that isn't a .yml

If you would want to spin them up again, yeah you could just copy them in a new place, make sure the docker-compose.yml is at the top and docker-compose pull the image, then docker-compose up. If your yml specifies that your configs are ./ , it'll know to look for them in the same folder as your yml and begin its file paths there. Anything outside your ./ in theory should be read only or be some kind of data off a NAS that is independent from the service. Its just not best practice in my eyes for your docker application to write outside of the folders beneath its docker-compose.yml location.

Just organize your volumes correctly and you can move the docker container anywhere you want with a copy paste and a docker-compose up. Imo docker-compose is WAY easier and faster than any GUI instance of Docker.

sudo apt install kate it is then!

What made you switch from Notepadqq to Kate? I'm currently still stuck on my Notepad++ bias too.

Oh, I've been there! lol Yeah, you sometimes don't expect to have to update an entire stack, not just one thing. Possible the companion app image was just released later and it wasn't on you either. :P

But yeah, give the cron job a try. I'd be curious to try it out myself! In theory your webinstance shouldn't cut out unless you click on something the exact moment the container is resetting.

Fuck yeah! This slaps. Brb while I get my headphones to listen to it again.

Exactly! It's not even that revolutionary in my eyes... heck that's effectively what it does most of the time.

But nevertheless, it can write a ton of code and it will for the most part achieve what you want. It'll still need to be reviewed, trimmed down and optimized. To me it's just like a college essay. I can write 2000 words in a 20 minute blitz as I am sure any decent dev can write a ton of lines in the same amount of time. Is it going to be any good? Is it going to be acceptable for the level you are writing? A college paper of 2000 words I'll rewrite 4-5 times and keep maybe as much as a paragraph's worth of text from the original draft.

Like what I just wrote above, I could probably condense into a few lines if I was trying to make it a ''professional'' statement rather than casual writing. There's these degrees of competence that I think LLMs haven't reached yet.

We should celebrate these kinds of changes. Like its great that they've already achieved so much and not just that, but left behind a group of folks just as committed in the future of the project.

I always feel that posterity of your work beyond your own contribution is the real achievement and a sign of humility.

What kind of issues have you had? How did you resolve them?

Well, they do share one thing in common and that they are all massive users of or suppliers for datacenters. This makes sense since Linux is used in all kinds of servers and appliances. The possible bias of these companies is going to more towards the Cloud and SAAS rather than anything else. Likewise, generative ai is marketed nearly solely as a cloud technology, when it can be hosted locally. We are like a year or two off the average person's web ai usage being possible to replace by ewaste gpus or entry level consumer hardware.

Nevertheless, I don't see how the Linux foundation can really be steered towards that bias in a way that is concrete. Its just the kernal. All the userspace development is done by other parties.

You always have to wonder if they even have the skills to debug the thing in the first place. Having tested generated code or code suggestions, its just like the age old practice of copying code off a google search. You need the skills to at least understand what the code does and why its the right implementation for what you want to achieve.

I wished more people tried their own machine learning experiments in video games, rather than only using existing models. Would be a really interesting way to develop procedural map generation or NPC/Enemy behaviours trained on your own work in progress builds. Probably difficult to tweak but I'm still curious to see someone try it out. I just see this persistent lack of genuine curiosity from those most interested in Ai technologies.

Otherwise, using generative Ai in general is just bad for business. I don't see the controversy and lost sales outweighing the benefits. Heck, even in business relationships: working with publishers, outsourcing etc... A number of partners will just flat out not work with you if you use Ai.

Throw in the morale effect on your team who is most likely spanning from apathy to open hostility. Its just not a good way to retain your best contributors,

Tokens are and will moreso get really expensive, so its just not a sustainable long term strategy. Creating a dependence on the highest end models will backfire financially and likely cost more than labour very soon. We are already seeing this with some of the largest tech companies busting their Ai costs.

Generative ai creates the environment where, I believe, the market will be dying to see genuine artistic creations and expressions. I think games, and most creative works, will soon be permitted to be of ''humbler'' qualities, as long as they are genuine. People want to interact with things people made. There's an opportunity here.

Compare it to game playing Ais, we have been able to beat all humans at chess for years, yet people still prefer playing other humans at chess. Chess is probably as popular now as it has ever been. I don't see people stop playing chess or making art just because some machine can do it ''better''. Better isn't even the point of those things and it misunderstands the human condition and experience.