Tarcizio Silva
banner
tarciziosilva.bsky.social
Tarcizio Silva
@tarciziosilva.bsky.social
Tech, AI and human rights advisor. https://tarciziosilva.com
Pinned
Racismo Algorítmico e Regulação de Inteligência Artificial: o contrato racial na produção do PL 2338/2023 tarciziosilva.com.br/blog/racismo...
Reposted by Tarcizio Silva
"It is against the TOS to be talked into suicide by our LLM" is possibly the most evil tech bro thing I've ever read
Additionally, OpenAI argues its not liable because Raine, by using ChatGPT for self-harm, broke its terms of service
November 26, 2025 at 12:29 AM
Reposted by Tarcizio Silva
Google at its peak was basically the best information retrieval system in human history and they and every competitor decided going from there to “you didn’t want answers you wanted half-assed auto-complete 80%-wrong hallucinations” in a few years was the right idea
November 25, 2025 at 1:57 AM
Reposted by Tarcizio Silva
😵😵😵

🔗 jnascim.info/tCyDZn
November 24, 2025 at 8:22 PM
Reposted by Tarcizio Silva
Poxa, 3 dias só galera... Faltou força aí para apoiar e defender o seu idolatrado MITÔ[mano]

brasil.elpais.com/brasil/2019/...
November 24, 2025 at 1:31 PM
Reposted by Tarcizio Silva
Orgulho de ter participado junto com Marcos Hecksher desse estudo sobre Bets no Brasil, ainda inédito, do Instituto de Estudos para Políticas de Saúde (IEPS). Os números são alarmantes e merecem ser vistos pela sociedade assim que publicados. Saiu hoje na coluna de @laurojardim.bsky.social
November 23, 2025 at 1:46 PM
Reposted by Tarcizio Silva
“Meta shut down internal research into the mental health effects of Facebook after finding causal evidence that its products harmed users’ mental health, according to unredacted filings in a lawsuit by U.S. school districts against Meta and other social media platforms.”
Meta buried 'causal' evidence of social media harm, US court filings allege - The Economic Times
Meta reportedly halted internal research into the mental health impacts of Facebook and Instagram after finding causal evidence of harm. Internal documents revealed users reported lower depression and...
m.economictimes.com
November 23, 2025 at 4:46 PM
Reposted by Tarcizio Silva
They love to say it's a way to help underserved groups of children too. Extra scary.

Forbes, S. H. & Guest, O. (2025). To Improve Literacy, Improve Equality in Education, Not Large Language Models. Cognitive Science. doi.org/10.1111/cogs...

PDF: philpapers.org/archive/FORT...
To Improve Literacy, Improve Equality in Education, Not Large Language Models
Huettig and Christiansen in an earlier issue argue that large language models (LLMs) are beneficial to address declining cognitive skills, such as literacy, through combating imbalances in educationa....
doi.org
November 23, 2025 at 3:05 PM
Reposted by Tarcizio Silva
lembrando que ela foi punida por isso uma vez e não baixa a cabeça
Carol Solberg, eu te amo!!!
November 23, 2025 at 12:53 PM
Reposted by Tarcizio Silva
Se ele dissesse que o que ele está fazendo não é tão diferente do que se faz nos gêneros massificados eu poderia deixar passar como hipérbole, mas, mesmo assim, só pros 90% que são cópias esquecíveis, pq até aí tem uns 10% que tem alguma fagulha criativa. Enfim, que merda, bicho.
November 21, 2025 at 5:36 PM
Reposted by Tarcizio Silva
e SEMPRE usamos tecnologias para criar, cultura é linguagem e tecnologia, mas a comparação que ele faz é ridícula, sobretudo aplicada a "IAs" que são apenas reconhecedoras e propositoras de padrões estatísticos. fora o fato de que isso é feito roubando trabalho, com custo ambiental etc etc etc
November 21, 2025 at 4:21 PM
Reposted by Tarcizio Silva
Surreal como esses caras se DESUMANIZAM de boas para justificar usar IA – ironicamente, a mesma operação que a BigTech faz para chamar LLM de "inteligente". Só que o mecanismo mais básico da criatividade é: tentar fazer igual e fazer um igual DIFERENTE. A evolução cognitiva humana foi isso. PQP!
November 21, 2025 at 4:08 PM
Reposted by Tarcizio Silva
"O que a IA está fazendo é o que sempre fizemos quando dizíamos 'vamos compor um rock tipo Led Zeppelin? Vamos fazer um blues tipo Chicago?' O Ira! está cheio de referências do The Who?"

Sim, cara, referência e recombinação estatística feita por uma máquina que nem sequer OUVE são a mesma coisa 🤡
November 21, 2025 at 4:04 PM
Reposted by Tarcizio Silva
Foi preso soluçando alegrando o sábado
November 22, 2025 at 10:37 AM
Reposted by Tarcizio Silva
Aprender alguma coisa pelo ChatGPT te deixa com um conhecimento mais raso de determinado campo do que simplesmente procurar no Google e ler sobre o assunto, de acordo com uma pesquisa que comparou o que mais de 10 mil pessoas sabiam depois de usarem um método ou o outro.
Relying on ChatGPT to teach you about a topic leaves you with shallower knowledge than Googling and reading about it, according to new research that compared what more than 10,000 people knew after using one method or the other.

Shared by @gizmodo.com: buff.ly/yAAHtHq
November 21, 2025 at 7:42 PM
Reposted by Tarcizio Silva
'In a blog post on Thursday, LinkedIn acknowledged the trend, but said it did not consider “demographic information” in deciding who gets attention. Instead, it said, “hundreds of signals” factor into how a given post performs.' 3/3
November 22, 2025 at 8:21 AM
Reposted by Tarcizio Silva
'The uptick in engagement has led some to speculate that an in-built sexism in LinkedIn’s algorithm means that men who speak in online business jargon are more visible on its platform.' 2/3
November 22, 2025 at 8:21 AM
Reposted by Tarcizio Silva
'Dozens of women joined a collective LinkedIn experiment this week after a series of viral posts suggested that, for some, changing their gender to “male” boosted their visibility on the network.' 1/3
Bro boost: women find LinkedIn traffic ‘drives’ if they pretend to be men
Collective experiment found switching profile to ‘male’ and ‘bro-coding’ text led to big increase in reach, though site denies favouring posts by men
www.theguardian.com
November 22, 2025 at 8:21 AM
Reposted by Tarcizio Silva
PODE JÁ!!
September 11, 2025 at 9:02 PM
Absurdle 6/∞ wordle-unlimited.io/game/absurdle/

⬜⬜⬜⬜⬜
⬜⬜⬜⬜🟨
⬜🟨🟨🟨⬜
⬜🟨🟨🟨⬜
🟨🟩⬜🟩⬜
🟩🟩🟩🟩🟩
Absurdle
An adversarial version of Wordle
wordle-unlimited.io
November 21, 2025 at 3:59 PM
Reposted by Tarcizio Silva
If It Doesn't Exist, MAKE IT.

Aka. The motivational bnuy.
#originalart
November 20, 2025 at 9:52 PM
20 de Novembro. Entendam o recado.
November 20, 2025 at 5:13 PM
Isso é excesso de podcast
Só pra garantir que todos já ouviram esse áudio
November 19, 2025 at 10:16 PM