AlexTECPlayz

Transfem Non-binary Aro Bi | Progressive Left | ⚛️ | 20

Game, software & web developer from Romania. Also interested in science, politics, humanities and a lot of other stuff.

―――――

Developing games at my indie game studio, Vanta Interactive (npub1ds3…5ree), working on #Prisonia and my game engine, Sapphire

Also check out my alt accounts on my own instance: AlexTECPlayz (npub1hyr…ptm3), Vanta Interactive (npub1ar8…n5sk), Sapphire Game Engine (npub154z…ewku), PRISONIA (npub1j7g…jrl2)

#gamedev #gameenginedev

Website: https://alextecplayz.github.io
Pronouns: She/They
YouTube: https://youtube.com/@alextecplayz

Public Key

npub1a3j0uwf8jvjy73pskd49kv3za04h5snja6tjx2t7l3atzadr242qr5xsms

NIP-05 Address

alextecplayz_at_techhub.social@momostr.pink

Profile Code

nprofile1qqswce878ynexfz0gsctx6jmxg3wh6m6gfewa9er99l0c743wk3424qprpmhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmntsn00zc

Publishing to

relay.momostr.pink

Author Public Key

npub1a3j0uwf8jvjy73pskd49kv3za04h5snja6tjx2t7l3atzadr242qr5xsms

Show more details

Published at

2025-08-07T20:46:48Z

Kind type

0 Metadata

Event JSON

{ "id": "cffe33e74386d927aeaf3c510b7419f16a2d3549be964519c635b3fa9bd7eb1b", "pubkey": "ec64fe392793244f4430b36a5b3222ebeb7a4272ee9723297efc7ab175a35554", "created_at": 1754599608, "kind": 0, "tags": [ [ "t", "gamedev" ], [ "t", "gameenginedev" ], [ "t", "prisonia" ], [ "proxy", "https://techhub.social/users/alextecplayz", "activitypub" ], [ "L", "pink.momostr" ], [ "l", "pink.momostr.activitypub:https://techhub.social/users/alextecplayz", "pink.momostr" ], [ "-" ] ], "content": "{\"name\":\"AlexTECPlayz\",\"about\":\"Transfem Non-binary Aro Bi | Progressive Left | ⚛️ | 20\\n\\nGame, software \u0026 web developer from Romania. Also interested in science, politics, humanities and a lot of other stuff. \\n\\n―――――\\n\\nDeveloping games at my indie game studio, nostr:npub1ds3qzuhe397xym660875nelsvchgzhkx8l37nd4yjmgknxhmduvscr5ree, working on #Prisonia and my game engine, Sapphire\\n\\nAlso check out my alt accounts on my own instance: nostr:npub1hyrfj57zdmvztdmkcu3lxq7y77ta3w883qczy2jcjg93yp7htyrqzyptm3, nostr:npub1ar8agky8s5fjt04yhhaqst3ystmk67gk6j07kkkrfstfwy0aj7dql3n5sk, nostr:npub154zpjtg52kl5d0t89e8mhyy4z7uj000tz8wd34gc3ql2f5taac2qtlewku, nostr:npub1j7gqj8r9l4980ch7wmawgnw68vnx0s68rt4n578f9y8tn9r4dcxq8xjrl2\\n\\n#gamedev #gameenginedev\\n\\nWebsite: https://alextecplayz.github.io\\nPronouns: She/They\\nYouTube: https://youtube.com/@alextecplayz\\n\",\"website\":\"https://techhub.social/@alextecplayz\",\"picture\":\"https://files.techhub.social/accounts/avatars/109/364/097/500/882/030/original/2513e6adace27805.png\",\"banner\":\"https://files.techhub.social/accounts/headers/109/364/097/500/882/030/original/3c4de385c68d8b6d.jpeg\",\"nip05\":\"alextecplayz_at_techhub.social@momostr.pink\"}", "sig": "891cb73c7d9b1564315be1dd047eaf9ce07af2ab4007e0201e57b488724ee855be1fa34e70a8527ccc53c4d8e811069445548913c050b05ca5b17ee66b05cf99" }

Last Notes

2026-04-09 20:35:31 UTC

- reply

you can't compare LLMs to a deadly poison.
I'm arguing that LLMs as tools can be made to be fairer and actually helpful instead of the sycophantic magic goo used to make billionaires richer. I'm arguing for the lesser evil if people want to use 'AI'.

2026-04-09 15:37:06 UTC

- reply

would you rather have users use ChatGPT or a LLM that isn't trained on plagiarized content and stolen copyrighted work? We need the EU to have a horse in the race, to be the lesser evil.

2026-02-03 12:26:00 UTC

- reply

Makes sense IMO, that's a better explanation. If the features don't run until the user explicitly wants to use the features, it's fine by me.
Does the AI Controls page specify that the AI models aren't downloaded until the specific feature(s) are used for the first time? Or does a linked help article from that page specify this?
It'd be nice to inform users about this in the browser, maybe even offer some details on how much storage the models use, and a link to the about:addons page for On-device AI to manage the installed models.

2026-02-03 11:55:19 UTC

- reply

TY, this is a nice feature 🩷
I have nothing else to (constructively) add, more that I agree with the points that it should be opt-in and perhaps some of the alarmist language / design be toned down. Still a step in the right direction, so that's great to see.

2025-11-11 18:19:02 UTC

- reply

4/6 and 6/6 (this comment)'s links aren't links. You might want to fix that.

2025-10-29 11:27:54 UTC

- reply

perhaps the paper was AI-generated as well? 🙃

2025-09-14 11:48:58 UTC

- reply

Valve has plans to publish this year a new version which also includes Ep 1 and Ep 2 + concept art, plus ideas and experiments for past Ep 3 attempts, as per Half-Life 2's 20th anniversary (https://www.half-life.com/en/halflife2/20th).

2025-08-04 12:15:01 UTC

- reply

I've made a small gist on GitHub that outlines the issues I've personally found with it.
Accessibility-wise: mainly that links are not underlined, and images are not clickable
Text-wise: too much information that may be overwhelming for a casual reader, no descriptions for some of the words, and some inconsistencies and typos.
Perhaps surface easier-to-understand (and easier-to-read) TL:DR sections for some of the content? Your average Joe most likely won't understand some of the legalese or complex language on the site.
https://gist.github.com/alextecplayz/9aeefe8f7752c0ec67fb1de981ef12b5

2025-07-17 14:28:37 UTC

- reply

Fair enough. Still, it'd be more useful to use TinEye or Google Image (/Google Lens) to search for images similar to it than some AI image detector.

2025-07-17 14:01:31 UTC

- reply

Ah yes, using AI to check if another image is AI-generated. These tools are completely untrustworthy and can easily be fooled.

2025-07-03 07:36:00 UTC

- reply

AI-generated slop answer.

2025-05-15 08:51:25 UTC

- reply

Not a GrapheneOS user (yet), but if it's anything like Samsung, then yes. Installing the stock image, removing any custom keys, firmware etc and relocking the bootloader should effectively restore it to its 'original' state. The only thing that can't be restored on Samsung devices would be Knox functionality, that is permanently tripped.

2025-04-16 12:19:16 UTC

- reply

Oh wow, Jesus fucking Christ that's horrible.

2025-03-03 10:13:38 UTC

- reply

And of course, Lunduke made a post about it, Ladybird getting attacked by "unhinged, dishonest activists"
https://lunduke.locals.com/post/5823666/ladybird-web-browser-developer-attacked-by-unhinged-dishonest-activists
Note that Lunduke's a horrible person, so this article of his is from a heavily biased anti-trans perspective, plus the article and his journal as a whole is from Rumble.

2025-02-02 19:58:16 UTC

Right, wrote a comprehensive guide on Android with steps on choosing a device (e.g. prioritizing security), installing a custom recovery, a custom ROM, custom kernel, rooting it, using root to run a VPN + DNS, disabling trackers, and minimizing your Google privacy footprint.
It'll be continuously updated with more stuff when I have the time to do so 🙂
https://alextecplayz.github.io/post/2025/02/02/Comprehensive-Android-Guide.html
#privacy #guide #Android

2025-01-11 11:42:20 UTC

- reply

techhub.social seems to have a 1024 character limit, but my own instance (social.vantainteractive.com) goes all the way up to 30000 characters.
From what I remember, Mastodon or other Fedi software may have issues with content longer than ~35(?)K characters, so I just set it to 30K. I probably won't write that much (longer content should be a note or post on my site), but you never know when you have a specific situation that requires content THAT long.

2025-01-07 23:24:18 UTC

- reply

Reminds me of this:
https://files.techhub.social/media_attachments/files/113/789/646/807/697/997/original/249a701fcd99b9da.jpg

2025-01-07 14:55:03 UTC

- reply

Eh, I'm not bothered by it. For me, the Home timeline is supposed to be a mix of local, federated and explore posts, so seeing account recommendations in Home and Explore seems fitting.

2025-01-07 14:43:35 UTC

- reply

There's no AI, and there might be no algorithm from my peek at the code in /api/v2/suggestions and the rest of the Ruby code

2025-01-02 21:25:29 UTC

- reply

Don't leave us hanging, what's up with the guy who built it?

2024-12-08 21:26:08 UTC

- reply

Still, this can't prove that it's AI-related. Who knows where the AI received the training data that included those names.

2024-12-08 21:10:46 UTC

- reply

You're jumping seriously quick to the conclusion that the movie was written at all with the help of AI. The whole movie has deliberate references to stuff, mainly the legend of King Arthur.
I'm sure there's some reasonable answer as to why they've used those names, that don't involve jumping straight to AI conspiracies.

2024-11-21 14:54:01 UTC

- reply

Oh, absolutely, Fedi feels and likely is much healthier for me and my mental health compared to Twitter. I was glued to the Twitter timeline when I had the habit of using it, and I kept that habit for a few days with Mastodon, before I figured that I need to drop that habit altogether. Twitter made me more susceptible to aggressive replies, ragebait and trolling, but eventually I got around to that and stopped caring - but it did have a lasting toll on me.
I don't know, surely there could be a middle-of-the-road approach, or as Bluesky allows, custom algorithms. Have a couple pre-built options like 'Surface relevant content from followers', 'Keep up with important events' etc, and a way to add your own, handled client-side. Of course, this is out of the scope of this discussion though.

2024-11-21 10:03:51 UTC

- reply

That's easy for you to say, you have 2.1K followers on Fedi. The engagement differs from one person to another. To me at least, Twitter has still delivered me more engagement, despite having very similar follower numbers.
And don't come at me with the 'oh but interactions on Fedi are from real people', because the vast majority of interactions on Twitter (likes, retweets, comments) have also been from real people. I've probably had less than 10 bots in the account's lifetime interact with my profile or content.
On Fedi I can post as much shit I'd like, and I'll get very little engagement. On Twitter at least I'd get one or two likes, possibly a comment every now and then on a post or comment about anything.
One thing I don't like about the lack of an algorithm on Fedi is that it doesn't resurface older posts that may be important. I've missed many posts that I'd consider important because my timeline is chronological, and I don't sit all day on Fedi.

2024-11-15 12:10:29 UTC

- reply

We need starter packs on Fedi as well, hope they get added soon.

2024-11-08 15:45:35 UTC

- reply

Sorry, I understood it as a conspiracy that the government killed the two, similar to Epstein.

2024-11-08 15:42:36 UTC

- reply

...
Ted Kaczyinski, the Unabomber that killed 3 and injured 23 (suicide in his cell, was suffering from late-stage rectal cancer, was depressed and was moved to a medical facility due to his declining health)
Timothy McVeigh, the terrorist behind the Oklahoma City bombings that killed 168 and injured 684 (execution by lethal injection)
Why do you think the government killed them? Either way, they were terrorists, they don't deserve to be martyrized.

2024-11-08 14:41:06 UTC

- reply

Note that if you like account sync, you can self-host Firefox Sync (https://github.com/mozilla-services/syncstorage-rs) and its associated Mozilla Accounts (FXA) infrastructure (https://github.com/mozilla/fxa/).
I'm in the process of doing this on the Raspberry Pi 5, Sync is easy to get up and running, FXA requires some work especially because I'm slimming it down of unnecessary modules.

2024-11-01 10:57:40 UTC

- reply

Seems like they took 'AI Support' literally, they have an AI chat for support questions on the site. I thought they integrated it in the distro somehow, á la Apple Intelligence or Google Gemini.

2024-10-27 21:13:41 UTC

- reply

Same opinion here. Batman may very well be a superhero, but he's not working for Gotham PD, and therefore, he can't dispense the 'law', or at least act as a judge or executioner. He's simply a 'good samaritan'/'police collaborator', nothing more.

2024-09-25 00:03:31 UTC

Idk just throwing this out there before sleeping: Why bother with XMPP, OMEMO, or whatever Matrix is smoking?
For group chats use MLS, for 1-on-1 use Signal protocol. Give the same features (emojis, basic stuff like file sharing, etc) + federation to both + moderation, admin features on MLS. That would be the backend. For the frontend, do whatever you want. Want a Matrix/Discord-like experience? Make a frontend like that, use the MLS features. Want a Signal/Whatsapp-like experience? Make a frontend. You'd still have the same features regardless, so nothing would be lost.

2024-09-17 18:08:17 UTC

- reply

related: https://techhub.social/@alextecplayz/112723799806401183
TLDR: Microsoft Authenticator KNOWS what accounts you add to the app, and sends telemetry regardless of your choice. Massive privacy issue in my eyes, because this could make you unique.
If possible, stay far away from all proprietary 2FA authenticator apps.

2024-09-17 14:05:57 UTC

- reply

Fully agree with you here, I think this might be a hot take for some, but not for me. This is why I can't get behind stuff like ACAB, or nicknames like AFCAB, etc. I heavily dislike generalisations.
Not all cops are corrupt, or bad. Most cops are probably decent people, working their job, hoping to instill some order where needed, and to catch criminals. It's just that some cops over-react to stuff, like what happened in New York with the subway shooting, and then the perpetrator allegedly wielding a knife. I heavily dislike the way some, even here on fedi, refer to cops as gangs, crooks, etc - that's an instant block from me.
Heck, or how some want to abolish law enforcement entirely, like that's going to do anything good. We can't have vigilantes running around trying to do 'justice', we need people to follow proper protocols in place to uphold the law - even if (and rightfully so) some parts of the law are broken, or gray legal areas that can and are taken advantage of.

2024-08-25 10:53:33 UTC

- reply

Threema is shit, though.
https://soatok.blog/2021/11/05/threema-three-strikes-youre-out/
the "most secure messenger" my ass, they have glaringly obvious issues. Threema is spreading deception and FUD.

2024-08-24 07:08:29 UTC

- reply

How so? It groups the users that liked a post into one notification, instead of 100.

2024-08-24 07:00:02 UTC

- reply

No. These are grouped notifications:
https://mastodon.social/@Viss/113012737879200869