Absolutely WILD that the development of sofrware proceeded with so little heed to record-keeping that 52 year old source code is a major archaeological trove that requires special tools to gain access to, like it's a treasure room in a sealed royal tomb from 3500 years ago.

Anyway. Off to make millions of tiny difficult incremental decisions which will be invisible to outsiders when consumed at speed but felt as a distinct texture and voice within the work. Byeeee!

Or maybe you can just push a fucking button.

Who knows. Maybe the difficulty of the process was where artists honed their craft and sharpened their communication skills, and that the millions of decisions in any piece of work, made over time by an individual, were where the richness of human artistic expression was hidden.

Gen AI requires minimal creative or craft skills, works at enormous speed, and is touted as a way of crashing through limitations and gatekeepers, freeing your creative vision.

So shouldn’t we have had some AI art that isn’t total shit by now?

Or was speed and ease not the problem after all?

Guillermo del Toro on wealth: "a wealthy man is a man who has enough, not a man that needs more. If you have enough to invite someone for a beer? You're rich. If you have a yacht, planes, islands, and you still need more? You're not rich."

Vou criar aqui também a "Thread das Threads" com os conteúdos mais importantes publicados pelo perfil.

Para iniciar com o nosso fio sobre as mentiras da Brasil Paralelo quando falam da história do Brasil.

bsky.app/profile/brpa...

I spent the week at the police surveillance convention and let me tell you my biggest observation: The name of the game now is consolidating as much information as humanely possible from surveillance devices, the internet, other governmental data, and literally a million other places. 🧵

I am calling it: the great enshittification of AI has begun, and the Subprime AI Crisis is brewing. Cursor is jacking up prices and adding rate limits, seeing reports of similar things happening with Replit, very nasty stuff.

When a chatbot gets something wrong, it’s not because it made an error. It’s because on that roll of the dice, it happened to string together a group of words that, when read by a human, represents something false. But it was working entirely as designed. It was supposed to make a sentence & it did.

Chatbots — LLMs — do not know facts and are not designed to be able to accurately answer factual questions. They are designed to find and mimic patterns of words, probabilistically. When they’re “right” it’s because correct things are often written down, so those patterns are frequent. That’s all.

This research is coordinated by @pinheiromachado.bsky.social with contributions from @jessicamatheus.bsky.social, @marinafrid.bsky.social Wagner Guilherme Alves da Silva, Priscila Cardia Petra, and Janaína Penalva.

Trem do desenho animado antigo

Desenho de 1982 que foi animado pelo estúdio que viria a se transformar no Estúdio Ghibli

HAPPY ZARDOZ DAY EVERYONE!