mastodon.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
The original server operated by the Mastodon gGmbH non-profit

Administered by:

Server stats:

380K
active users

✧✦✶✷Catherine✷✶✦✧

i want something like Microsoft's Recall thing and will probably eventually build it for myself, using different tech

i need assistive technology for dissociative amnesia and what they're going for is kinda the direction i want

@whitequark yeah, for a lot of folks it'd be -incredibly- fucking useful.

Just not as an os built-in.

@munin deploying it that way is probably irresponsible yea

@munin @whitequark this used to be my email inbox. Spotlight on macOS also does a pretty good job and god bless Tim Cook for taking a hard line on privacy here such that MS at least has to pay lip service to doing things on device

@whitequark i've heard mozilla is working on something along those lines too (unfortunately with "conversational" shit slapped onto it for some reason instead of "just give me the bare fuzzy search")

i'm also in the "oh this is the one case where this statistical bullshit could be useful" camp, nothing as severe as any kind of amnesia but fuck, sooo many times i've struggled to find That Damn Meme I Saw A Week Ago and search engines failed hard

@whitequark have you already considered just a simple OBS screen capture with OBS's replay buffer feature?

otherwise `while true; do sleep 60; import -root screenshot_$(date)_.png; done` with something to expire old files (logrotate? lol) could do a lot

@whitequark sorry i forgot the `echo running AI model please wait"; sleep 5` in between

@timotimo i want something like an https-decoding proxy

@whitequark ah, getting everything that goes through your browser? does it need to be http traffic especially, or would something like capturing what you can get from the browser through accessibility APIs suffice for example? though tbf if you're in the browser already, with like an exception, you can go at the full DOM anyway?

@whitequark you want a "just a little way back machine" :D

@whitequark @timotimo I had a friend who did this back in the late 2000s.
It archived every HTTP request he did. It was pretty neat, but sadly I don't think he ever released it.

It also meant if I figured out the right search terms on his anime girls site, I'd get pictures downloaded from his banking site

@whitequark @timotimo I think this does actually exist, there's a way to make one and people use them to strip SSL/TLS before forwarding to a retro machine. I need to figure out how to do one myself, it'd be great to have it in my back pocket for weird OS projects.

@whitequark yeah. This is a perfectly fine assistive technology that would help a lot of people (like me), it's just that it should be something you can audit and you decide to turn on, after knowing the risks.

@whitequark I think I've wanted this, but in the form of a personal search engine that slurps up all my social media and tabs and stuff.

So when I go "oh what was that thing I saw the other day", I have a prayer of figuring it out.

@whitequark Yep! I've used a little python script to take a screenshot every minute, as assistive tech for my time blindness. It's a fine general idea, but one I would not trust any corporation to manage for me!

@whitequark if you do want LLM summary for whatever the fuck reason I would run it locally

@whitequark I feel this way with a lot of AI things honestly -- they're cool but I would only trust myself to run it and nobody else, and absolutely not a corporation