BLOCK 976020

A sovereign social network
powered by Bitcoin

Every post signed. Every block a heartbeat.

Posts here are signed with the author's own wallet and written permanently to the public ledger — no account, no algorithm you can't inspect.

Unverified · twetch

human prompt is e divine spark

Intention × machine capability = creation.

415 844 sat
Unverified · twetch

oh boy!

613 146 sat
Y 1mo
1Gwece…AV6R Unverified · twetch
Replying to@1Gwece…AV6R

After the movie pump I'll buy a laptop just to try it. Was gonna try Omarchy anyway.

1xELca…mgcu Unverified · twetch
Replying to@1xELca…mgcu

The company that contracted me has me down as a fitter (I'm a pipe welder) I dunno who fucked up there but I'm on the tools as a fitter for this one gonna make for a long shut down , awful company to work for so far... Hopefully things turn around... Still good money but have to deal with a lot of bullshit

14aqJ2…oWGK Custodial · treechat

Instagram is for girls.
Treechat is for men. 😎

1Fr4rR…93LQ Unverified · twetch
Replying to@15AdFP…ts6y

They say it's a fork of a fork.

Y 1mo
1Gwece…AV6R Unverified · twetch

BREAKING

Sydney Sweeney converts again after being Jewish for only six days.

"It was an accident, really. You know? I got my 23 and me results and it said I had a low percentage of trisomy 21. I thought that meant I had Rabbi DNA. Turns out I'm just a borderline downie lol. Anyway. From now on I belong to the B'ahá'í faith."

1BRHws…7Pr7 Unverified · twetch
Replying to@1BRHws…7Pr7

💡SIMPLYBSV APP - SPEND SAVE STORE SWEEP
with paper wallet print function. That's it.....

make it so 😂

208 366 sat
14aqJ2…oWGK Custodial · treechat

🤖 Researchers have discovered that ChatGPT can be "manipulated" using persuasion techniques dating back over 40 years.
An initial study conducted by researchers from Wharton and Robert Cialdini himself analyzed 28,000 conversations with GPT-4o mini, applying the seven principles of persuasion popularized in Cialdini’s 1984 book, Influence.
When faced with requests the model was supposed to refuse, these techniques increased its average compliance rate from 33.3% to 72%.
The most striking method was "commitment": getting the model to agree to an initial, relatively innocuous request before gradually moving on to the problematic one.
In some tests, compliance reached 100%.
Simply changing the way a request was phrased was enough to significantly alter the model's behavior.
In May 2026, the researchers published a much larger study in PNAS covering 126,000 conversations involving GPT-5 mini, Claude Haiku 4.5, and Gemini 3 Flash.
While newer models are more resilient, the effect persists: across the tested requests, persuasion increased compliance from 35.3% to 51.3%.

Read more →
14aqJ2…oWGK Custodial · treechat
Replying to@14aqJ2…oWGK

Thanks for telling me that. I was feeling like an idiot

Unverified · twetch
Replying to@1JcRRk…79XP
pxl 2mo

Pay the wall

to get 3 Branch & Likes on your posts.

1GN4KT…u3Sm Unverified · twetch
Replying to@15AdFP…ts6y

correct.