You know what's great to learn at 38, right before bedtime one night, after a whole life time of struggle...

That your mother traumatized you with ADHD to the point where actually pushing myself against ADHD is a trigger... making it so no amount of coping skills can help me push myself for any real length of time...

I'd try and do something that requires a push, even a little one... and be crying... and I thought it was just the feeling of hitting my dopamine... but nooooo... now I'm unpacking that it was me getting triggered at the feeling of pushing my reserves at all...

That explains so damn much and makes me feel so damn hurt and angry...

I was forced to push myself so far so often as a kid and she didn't ever relent when I was critically over-extended on dopamine... usually around cleaning. I'm remembering so many times crying and sobbing on the floor because she demanded I clean to an extreme standard and I wasn't allowed to do anything else until I met her approval...

And I've been running my whole life fucking kneecapped by this... I thought I just had it worse than most (with ADHD) on my ability to push myself on tasks... but no... it's because I realize now I can't fucking push myself at all because my fucking brain just jumps straight to that extreme pain and trauma right away...

Now I'm fucking crying when I should be trying to sleep...

#ADHD #Neurodivergent #Trauma
@adhd group @actuallyadhd group

I still want an eggbug plushie
maybe i'll get it second hand

#cohost

Why did we ever fall away from BBCode? As far as I can tell it can do way more than Markdown and it doesn't have 90 million different implementations.

#Markdown #BBCode

Lemmy <-> Friendica federation is a bit jank. I wonder if there's anything I'm able to do on my end.

#lemmy #friendica #apMeta

So, fellow butterfles, remember when I mentioned Friendica yesterday? Instances running the latest version let you post to ActivityPub, the in-house DFRN protocol, and ATProto. You'll have to give it access to a preëxisting BSky account but there you go.

#atproto #friendica #ActivitiyPub

I shared this a few years ago but a lot has changed since then...

Not a perfect graph regardless, but I like it for communicating.

Mozilla announces on-device (as in not using the cloud) AI image recognition to generate alt text for users who need screen readers...

And people are flipping out over AI still in the comments

Mozilla

2024-06-03 09:35:50

Firefox 130 is bringing a game-changing feature: automatic alt-text generation for images using a fully private on-device AI model! 🙌🏾

Initially available in the built-in PDF editor, our aim is to extend this to general browsing for screen reader users. hacks.mozilla.org/2024/05/expe…

Experimenting with local alt text generation in Firefox Nightly - Mozilla Hacks - the Web developer blog

Firefox 130 will feature an on-device AI model that automatically generates alt-text for images, integrated into its built-in PDF editor.

^{Tarek Ziadé (Mozilla Hacks - the Web developer blog)}

After one of the owners died, work has been struggling with not having admin access to some of our systems, namely Google Workspace.

We finally got that access... so now I've sent the email to my boss about my name change. I verified with my boss a month or two ago that he's not going to fire me for being trans (it's actually a cute story...), that way I could safely start the legal name change process... but I've been holding off on otherwise being "out" at work because of this account thing...

So finally sent an email to my boss about the new name... it's the last place where I wasn't out!

#lgbtqia #trans #ComingOut

I made something to scratch an itch that's been bugging me for years... will save me a lot of time and trouble on some of my work (and possibly personal) projects...

In other words, I made something: pypi.org/project/pyasdb/

(For anyone wondering my current use case, I have a tool I wrote for work that's sitting on top of peewee and sqlite and it's been a massive pain to expand it out... this isn't something super special, but for the 20k-40k or so entries this should do just fine searching/querying/etc, even before adding indexing... and it'll make rebuilding my toolkit faster as well as easier to maintain since there's no firm structure to the resulting database)

pyasdb

A simplified NoSQL offline database built on top of shelve/pickle

^PyPI

Want to learn real quick that "leftist" does not automatically mean "good person" or "free from bigotry"?

Call them out on body shaming or misgendering monsters.

It's also a painful reminder that so much "acceptance" is just performative.

If you think I'm valid for being fat, but then insult a monster for being fat, it tells me that you never actually saw me as valid.

If you think my gender transition is valid, but then purposefully misgender an awful trans person, it tells me that you never saw my transition as valid.

Because if these things are true and valid, then being an awful person wouldn't matter!

#lgbtqia #leftist #bigotry

OMG I'M WHEEZING!

#GoodOmens #GoodOmens2 #NeilGaiman

... now, to cry in relief...

Server News

2024-05-23 03:53:25

<Insert Profanities>

So the temporary system had some sort of failure, I'm not even 100% sure what caused it to be honest. It went down sometime yesterday and some of the virtual drives got corrupted, which caught the database and the virtual gateway device.

I was able to restore the system... most of the way. Thankfully there are backups of the database, but some of them were also flawed as well, the most recent intact one was from 5/16, so 5 days were lost.

To be clear, this problem was exacerbated by the fact that there's not as much redundancy in the temporary setup (sadly it looks like it'll be a few more months before I have a place of my own and can spin up my own hardware again). But I'm going to still look at how I might get those in better shape.

As far as how long it took: I had a busy day yesterday and didn't see that the server was down until I was too exhausted to do anything about it, so it had to wait until I got off work today... each attempt at restoring the database takes around an hour, so that took *a while* to get restored.

... ugh...

youtu.be/DlFkfOqtgR8?si=2e-AF2…

#ADHD #neurodivegent @actuallyadhd group @adhd group

What Don't You Understand? - A Short Film about ADHD

ADHD is a very complicated thing, and it is very hard to explain, especially to those without it. This film takes you through common experiences that come wi...

^YouTube

The price of love is pain*...

What do you do when your capacity for pain is getting used up faster than your capacity for love?

(* for those unfamiliar with the saying: loving someone inevitably means pain at their loss, pain at their suffering, and general vulnerability. In my case, mostly loved ones struggling and being traumatized by various systemic injustices...)

The raw surrealness that hits you from time to time when transitioning...

I thought I was a guy... now I'm thinking of myself as a woman.

I used to look in the mirror and always feel blah, and unmemorable unattractive face... now I see a face I genuinely recognize and remember, a face that I can feel attractive wearing (even if I'm not my own type and will never really see it myself)

My hair which was always just an uninteresting mess... turns out it likes to naturally grow into a cute bob (though I'd still like to try styling it more).

Oh... and boobs, just boobs.

#trans #transfemme #SecondPuberty #lgbtqia

Rant about AI:

Sadly there's no reasonable way to differentiate AI content from "real" content. And regardless of your opinions on AI there's no "stopping" it (it's a "cat's out of the bag" situation, you can run these things on your home computer with open source software... there's no way short of an apocalypse to stop development from here).

What we do have is a lot of fighting and little effort to work on solutions of living with this. And I think worse yet many taking the anti-AI stance, especially the loudest of them, are basically making things worse because real solutions are anathema to them (ie. anything short of an outright ban on the technology is unacceptable, which means they tend to push back against even efforts to rein in AI or talk over those who want to push those efforts).

On top of that you have the borderline predatory push of "AI Detection Tools" and "AI Poisoning". The detection tools are a question of "How many real lives are you okay with ruining to catch a handful of bad uses cases in AI because there is zero way to have any certainty on the accuracy of these tools?" while poisoning tools are a security blanket that leads to people dropping their defenses because they don't stop AI, just slightly delay it's access to your content (even the creators of those tools acknowledge that AI will quickly bypass them, at which point there's no difference in whether or not you used that tool), worse yet as AI gets further incorporated in search tools it can make it harder to get visibility and exposure over AI generated content.

What we really need to be focusing on to address the problems with AI:

• Learning how copyright works (in my experience artists tend to have a woefully bad understanding of what is or isn't covered) and making sure corporations don't lobby the government into allowing copyright on AI works (under current law they are public domain, aka. no copyright, but there's already been one case of pushing that they can copyright "arrangements" of AI works). This means if they want to actually have a copyright on art, they've got to pay a human artist
• We need to push for reporting requirements/standards. One of the most toxic elements is how much AI floods spaces and bumps out human artists, especially when they attack the prompt containing the artist's name (meaning searching for that artist can turn up more AI work than their actual work)... there needs to be a requirement that AI art be labeled. This also works with the previous point as it is similar to being able to search for something released Creative Commons.
• Push for copyright responsibility in outputs rather than training data inputs. This sounds like something that is already one of the loudest arguments, but really isn't. Most arguments I hear try to go after AI tools for copyright content in their training data... but if you actually learn copyright you realize that a victory here largely means that major companies get more of an advantage because copyright only applies when content is copied (ie. when the training data is made available for smaller companies to run their own) vs when content is transformed (despite popular opinion, the vast majority of AI output does not violate copyright and qualifies as a transformative work... see again learning copyright law, plus a dash of learning how these tools actually work). Responsibility in outputs means that an AI can violate copyright (if I ask an AI tool to give me the first chapter of a copyrighted book and it does so... that is a violation and they need to genuinely be responsible for taking measures to prevent this from happening, but there should also be leeway for "forced violations", ie. when you bend over backwards to make it break copyright vs just saying "give me the first chapter of...")
• Work on learning and developing responsible usage. Again despite popular artist opinion, there genuinely is a lot of responsible use cases for all these AI technologies, from using LLMs to help debug code, summarize text, prioritize lists to voice duplicators used, with the license of the original VA, being used for dynamic speech (ie. voice assistants or actually speaking a player's name in a video game in the middle of otherwise pre-recorded output). And that's not to ignore image generators which can be used for enhancing/repairing old photos, or just used for general visual effects on your own art (ie. the filters everyone uses on instagram or the like... much of them are the exact same tech as AI Image Generators)
• And as always... fighting capitalism because the real threat of AI is the same as any other technology advancement: if CEOs can replace you with a machine, they will, and we live in a society where no employment means risk of death.

#AI #ResponsibleAI #Rant

• Sensitive content
• Filtered word: nsfw

Click to open/close

youtube.com/shorts/kenDWpNu6Go…

#ADHD #AuDHD

Brennan just learned the word "Deltohedron"

If you need to distract Brennan, talk to him about dice names #shorts #dimension20 #fantasyhighFor select free seasons of Dimension 20, clips, and more, visi...

^YouTube

Test from Friendica to:

Mastodon
Friendica
Hubzilla
Lemmy
Kbin
Pixelfed
Gotosocial
Iceshrimp
FunkWhale
Peertube