f4mi ‼️

f4mi ‼️

@f4mi.bsky.social

i am a creative mess who lives in the EU and [insert the rest of a cool funny description here] www.f4mi.com

13.2K followers109 following1.8K posts

(@insignia.live please don't ban me I just wanted to see if I could binary patch Xefu to make this work, I am not cheating or doing anything weird with it) insignia running on my rgh modded xbox 360 1/2

has some quirks with how it's reporting its NAT to the servers (reports itself as NAT type 0) but it just seems to be like... working great? 2/2

if the team is not upset with me for this I will gladly share how this works, it's pretty much just a bunch of patches to xefu7 (been looking into Xefu for a while now lol)

my uncle gave me my cousin’s 2010 (more like 2003 when it comes to specs) notebook from back when he was in college and i found screenshots of the anime he pirated back then

f4mi ‼️ reposted

It drive me nuts that now people use AI to write tweets. If you think that this way you will get more popular and so forth, think twice. Write something authentic. AI is great but it is easy to misuse. Your tweets must be your more cared thoughts.

tried the new crazy taxi beta it’s fun! but it doesn’t feel or play like the first three games at all, the physics is just completely different and in general it has been massively reworked around its multiplayer ambitions

f4mi ‼️ reposted

Andy Masley, who is literally paid via grants from AI investors, can "write" 85 pages of text on Substack and spend 8 months insulting anyone researching infrasound, but "doesn't have time" to participate in a single real time conversation, joint-interview, or field-demonstration he's been offered.

mfw i just bought an OLED LG TV like a month ago it’s staying offline forever lol

Alex Hern@hern.bsky.social

It's finally happened: the improved quality of machine transcription and the greater ability to extract value from unlabelled data means "your tech is eavesdropping on you for marketing purposes" has crossed from "false conspiracy theory" to "true and concerning" www.theverge.com/tech/991190/...

i have been consistent on what i personally consider ok or not: the “anti-ai” video you quote has me use a local open model (deepseek) as the backbone of the data poisoning technology, i don’t understand the cherrypicking of what i said or did when i have been pretty open the whole time about this

SpacePrincess@spacefluttershy.bsky.social

F4mi being pro ai is so weird and disappointing, not to mention hypocritical considering they literally made an anti ai video. They give of the impression that they want to pick and choose when ai is and isn't okay to use, switching stances whenever it is convenient for them

please don’t attack this person btw, i don’t want to send any hate mob, i just don’t want to be portrayed as someone who “was anti-ai and betrayed the movement” or some shit when my stance on this stuff has always been extremely transparent

i have been consistent on what i personally consider ok or not: the “anti-ai” video you quote has me use a local open model (deepseek) as the backbone of the data poisoning technology, i don’t understand the cherrypicking of what i said or did when i have been pretty open the whole time about this

SpacePrincess@spacefluttershy.bsky.social

F4mi being pro ai is so weird and disappointing, not to mention hypocritical considering they literally made an anti ai video. They give of the impression that they want to pick and choose when ai is and isn't okay to use, switching stances whenever it is convenient for them

View full thread

the same pro human approach that has me poison the training data that is stolen by companies to train their closed models is the one that pushes me to want open models that work for everyone and run locally for free on the cheapest hardware possible, there is not boycotting away out of this

you can’t win against companies that can literally burn money and run at a loss for a decade potentially to get everyone hooked by simply deciding to be in the minority that doesn’t use their stuff, you gotta provide a proper alternative, and that’s why i push for proper open and local stuff

i have been consistent on what i personally consider ok or not: the “anti-ai” video you quote has me use a local open model (deepseek) as the backbone of the data poisoning technology, i don’t understand the cherrypicking of what i said or did when i have been pretty open the whole time about this

SpacePrincess@spacefluttershy.bsky.social

F4mi being pro ai is so weird and disappointing, not to mention hypocritical considering they literally made an anti ai video. They give of the impression that they want to pick and choose when ai is and isn't okay to use, switching stances whenever it is convenient for them

my stance on this stuff has always been that i don’t love ai when it’s used for “art” but i find it ok for optimizing boring grunt work and that i don’t see ai going anywhere since it’s so cheap and decentralized by now you literally can’t stop it (weights are often just a few GB)

my stance has always been and will always be always be that the best we can do as a society is make this technology as open and cheap as possible, to avoid 4 companies controlling it all and imposing their own vision of the world on us, this and poisoning public scraped data are both pro human takes

i also fucked up this gfx in general leaving a big space on the right side and not realizing until now and idk if claude would actually make a mistake like that lmao

f4mi ‼️@f4mi.bsky.social

some people were asking if some of the diagrams in my last video were made by claude (specifically the 2D ones like this one) and nope, me and a friend made all the assets and designed everything inspired by the "Remind Me" MV by Röyksopp then I used Kimi to help me with some of the HTML stuff BUT

i am now pretty sure kimi did fuck around with some of the text there trying to “optimize” the flow when i used it to help me with the backend code for animations and stuff tho, will make sure that does not happen again

some people were asking if some of the diagrams in my last video were made by claude (specifically the 2D ones like this one) and nope, me and a friend made all the assets and designed everything inspired by the "Remind Me" MV by Röyksopp then I used Kimi to help me with some of the HTML stuff BUT

some people were asking if some of the diagrams in my last video were made by claude (specifically the 2D ones like this one) and nope, me and a friend made all the assets and designed everything inspired by the "Remind Me" MV by Röyksopp then I used Kimi to help me with some of the HTML stuff BUT

some people were asking if some of the diagrams in my last video were made by claude (specifically the 2D ones like this one) and nope, me and a friend made all the assets and designed everything inspired by the "Remind Me" MV by Röyksopp then I used Kimi to help me with some of the HTML stuff BUT

View full thread

in general I don't have issues with open models (almost always locally) to help me with the grunt work behind the scenes when it comes to coding or using them as glorified search engines or sometimes for upscaling very blurry graphics, but yeah I avoid using them to just idk making assets ofc lol

in general i have made my stance pretty clear over the years: i am not an anti-LLM hardliner as some think for some reason, I just try to use this tech as ethically as possible and without having it actually intersect with my artistic work at all

some people were asking if some of the diagrams in my last video were made by claude (specifically the 2D ones like this one) and nope, me and a friend made all the assets and designed everything inspired by the "Remind Me" MV by Röyksopp then I used Kimi to help me with some of the HTML stuff BUT

looking at it again I can see why that style could unfortunately look like claudeslop lol I was trying to replicate that sort of “y2k corporate” style and it unfortunately overlaps a lot with what some claudeslop looks like today, I also think it might

have changed one or two words there now that you make me notice (in a way that passed my sniff test when I was rechecking everything I guess, either that or I probably assumed my friend helping me with this reworked some of the text and I forgot ask him since it made sense and was factually correct)

good news: finally uploading the video bad news: it’s 25GB so it will take a few hours lol

LeoTB@leo-tb.bsky.social

It's been 2 weeks since the update... Someone... please... I need to see dat new video soon! Any updates?

sorry 4 the delay, i wanted to post the video yesterday but my HOUSE'S ELECTRICAL WIRING WAS CATCHING ON FIRE and we had to call an emergency electrician lol fml but also for fellow euros PLS REMEMBER HOW IMPORTANT SMOKE ALARMS ARE they saved my life lol

i am in the weird situation now where due to me half-remembering something and then searching it around 3AM i have bought a very specific extremely specific and fascinating piece of x86 hardware that somehow Cathode Ray Dude still hasn't covered i almost feel guilty like he is THE x86 weird hw guy

yes i will make a video on this, fact is nobody really talks about how this is x86 based since the company making it sold it as... an appliance, pretty much, and it being x86 is the least weird thing about it i will stfu now before spoiling anything more lol

i am in the weird situation now where due to me half-remembering something and then searching it around 3AM i have bought a very specific extremely specific and fascinating piece of x86 hardware that somehow Cathode Ray Dude still hasn't covered i almost feel guilty like he is THE x86 weird hw guy