DontTakeMySky

On top of everything else people mentioned, it’s so profoundly stupid to me that AI is being pushed to take my summary of a message and turn it into an email, only for AI to then take those emails and spit out a summary again.

At that point just let me ditch the formality and send over the summary in the first place.

But more generally, I don’t have an issue with “AI” just generative AI. And I have a huge issue with it being touted as this Oracle of knowledge when it isn’t. It’s dangerous to view it that way. Right now we’re “okay” at differentiating real information from hallucinations, but so many people aren’t and it will just get worse as people get complacent and AI gets better at hiding.

Part of this is the natural evolution of techology and I’m sure the situation will improve, but it’s being pushed so hard in the meantime and making the problem worse.

The first Chat GPT models were kept private for being too dangerous, and they weren’t even as “good” as the modern ones. I wish we could go back to those days.

I run Debian on most of my systems and run all of my services in docker (with rare exceptions for node_exporter or stable core tools). My base systems get automatic security upgrades, and then I’ll manually check in every few weeks whenever I feel like it.

My services in docker are version locked to a specific major version (when there’s a tag available) so I can usually re-pull to get minor version updates freely without breaking issues. My few more finnickey services get manual upgrades from me every 6 months or so only.

I usually stick to an OS version for as long as I can, and to that aim I stick to LTS versions with long support windows.

4 major versions in 12mo is…a lot. Especially if those include breaking changes for you. Yikes

A Windows VM running Windows terminal, SSH’d back into the host, obviously.

Honestly I stick with whatever the default is and never had a problem that led me to find anything else.

The Lemmy app we are using on our phones needs to download content from Lemmy so it can be displayed to us. Lemmy might just have one big file full of links, but that’s annoying to have to write code to handle. Or it might have a folder full of files where each file is a post, but that’s also a bit annoying to write code to manage.

It (probably) uses a local SQLite database to store all of the cached posts.

Conceptually, a database is just a place to store things, just like a big text file. The database just handles a lot of the grunt work for you and makes it easier to search, organize, and filter the data.

So anywhere there is data, there could be a database.

I see you got your answer, but I’m adding on for anyone else that comes across this.

For me, I learned the most when I had a disposable and replaceable system. When I was dual booted I was too scared to touch anything in case it fucked everything up. Once I started poking round on a Pi, LiveUSB, etc it was a lot easier to learn because I could always restart.

Id start there with something like Mint or Ubuntu. Then set it up in a way where you can easily replace your OS so you can reset it often and fuck around. Then just learn as you go.

It’s probably too expensive for them to record and upload every call, but im sure the transcripts of calls are likely to leak soon after this.

It’s not even limited to smart cars though. Yes used does let you a oid it, but it’s not like this is just people buying the fancy trims either. Shit like this is working it’s way down to the run of the mill standard cars year after year.

And each type of communication needs it’s own switch. Don’t let them pull some BS trying to make you enable all the hardcore tracking via a cell network just because you want to connect to Bluetooth.

Cool. Now all of Google Drive is blocked because one guy hosted a movie there for a few days.

Especially for physical goods.

It’s especially weird when the existing targeting can be so effective for much cheaper.

For tvs for example, they can see what you watch, when, what ads you mute and which you don’t, what you display over HDMI (content ID), the other devices on your network, your location, your accounts for every streaming service, what you search for. Then if you install their companion app they learn the other apps on your phone, your location habits, the media you play on your phone (looking at you Bose connect app…), bluetooth and network devices you are near (connecting you to other profiles they know), and probably a lot more.

Maybe. They might do some processing locally and just upload as text so it might be easy to batch the data, making the upload volume and pattern less obvious.

It also saves them network bandwidth so I’m sure that would motivate them too. Uploading raw mic data from all TVs would be expensive.

This feels most likely to me too, though maybe even for an Edu or “managed” Windows install of some kind where you get more support and more opinionated stable installs or something. Sort of like Windows S mode but maybe with more help.