Profile avatar
brantaaudensis.bsky.social
an injury to one is an injury to all
709 posts 327 followers 518 following
Regular Contributor
Active Commenter

San Francisco’s Donald Trump Connection A closer look at how powerful SF figures and institutions have backed Trump — and how they’re shaping local politics.

I wonder if people born in the last 20 years know the TV meltdown trope, where our hero flips through channels and they’re all showing the same thing

WATCH THIS. He gets it right Nottoway Plantation: the so-called “White Castle”, burned down. With it, a monument to whitewashed brutality. We don’t need romanticized nostalgia for slavery. We need truth. Let this be the end of pretending these places were anything but what they were: crime scenes.

Hot sauce after 6pm?

💡 💡 ⚡ 🧊 !! Inspired by this great new piece on how much water "AI" uses, I have a start-up idea that just zapped me on the head like a lightning strike. It blends modern AI with pretty ingenious historic techniques. 🧵 1/n www.bloomberg.com/graphics/202...

My god I love train travel

California can't afford to *not* build High Speed Rail. CA High Speed Rail will save the state $100 billion on airport and highway investments - not including the steep economic costs of air pollution and climate change. #FundHSR

AI’s only general purpose use is to control labor. Thank you.

Having already acquired the Republican Party, the Crypto Fascists are now buying Democrats a la carte. Beware the Crypto Democrat — a wolf in wolf’s clothing. Crypto is for crime and Tech Fascism. Period.

Across the country, a May Day Strong coalition has planned over 1,000 mobilizations between May 1-3 under the banner “For the Workers, Not the Billionaires." Follow for updates on actions. ⤵️

A Zoox (Amazon’s self driving car subsidiary) car allegedly hit an immigrant food delivery worker on a moped last week in San Francisco. Under Prop 22, the worker neither gets the min wage nor workers’ compensation. What a dystopia!!!

LLMs are nothing more than models of the distribution of the word forms in their training data, with weights modified by post-training to produce somewhat different distributions. Unless your use case requires a model of a distribution of word forms in text, indeed, they suck and aren't useful.