Also, Meta CFO Susan Li said that they *doubled* the number of GPUs they use to train ad models and that only yielded an increase of 3.5% in FB clickthroughs, 1% in IG clickthroughs -- that's hugely wasteful for such tiny gains. The environmental cost of that doubling is likely horrible
Revealing tidbit from WSJ about Meta's AI revenues (www.wsj.com/tech/ai/meta...).
Basically, Meta is claiming AI is working because it's growing it's advertising business -- ad clickthroughs on FB up 3.5% and $10b in Q4 from AI-generated video ads.
This is AI's business model. Ads.
So now the EPA officially provides resources for data center developers to get around the Clean Air Act. That's nice.
www.epa.gov/stationary-s...
But luckily, the city council rejected the project unanimously.
Take that Kyrsten.
www.politico.com/news/2025/12...
Now, she's apparently working "hand in glove with the Trump administration” on AI, bulldozing local communities concerned about clean air and water so she can make money consulting for Hogan Lovells, a global law firm.
Welcome to the revolving door of US politics.
Here's a fun throwback about Kyrsten. @sunrisemvmt.bsky.social planned to protest her at the 2021 Boston marathon, asking her to be bold & not hinder major climate legislation.
She found out, chickened out, & didn't race.
Then she tanked the Build Back Better plan.
www.boston.com/news/politic...
Just when I thought my days getting riled up about Kyrsten Sinema were over, she comes back into my life.
I was reading about data center fights across the U.S. and found that Sinema, former enemy of real climate policy, has now taken to AI lobbying
zeteo.com/p/kyrsten-si...
We need real power. That means reforming campaign finance so corporations do not control politicians. It means actually governing our workplaces. Stopping media lies. To get those things we need massive social movements, and organization to push the change the actual structure of society
We need to double down on building real power to hold the state accountable. That means being able to strike and withhold labor. That means being able to hold snap elections to remove politicians. To get to the point that we can do these things, we need to reshape society drastically
In a nutshell, the argument is that we have seen so many unaccountable acts of violence from the U.S. government and its corporate sponsors, that we need to take seriously the idea that we do not have control. Protest and speaking truth to power does not work without real power.
When is enough enough? The U.S. violence against Venezuela is an indication that our society is off the rails, totally uncontrolled. I wrote a brief response on my blog discussing how we need to build back the means of controlling our rogue nation.
ricknabb.medium.com/the-attacks-...
Over and over the US inflicts obscene violence against the global South, in flagrant violation of international law, to maintain the conditions for capital accumulation for its decadent billionaire class. It's disgusting and intolerable. We cannot continue to accept this.
This morning from @wsj.com. Trump admin is looking to override states regulating #AI. The contradictions from the party of States' rights are thick.
Tech companies' actions are already deeply undemocratic. This techno-authoritarianism is unacceptable
Bar chart showing that 59% of U.S. teens say they use ChatGPT. Far fewer teens say they use Gemini (23%), Meta AI (20%), Copilot (14%), Character.ai (9%) or Claude (3%). The chart is based on a fall 2025 Pew Research Center survey of teens ages 13 to 17.
NEW from @pewresearch.org: Most popular chatbots among US teens www.pewresearch.org/internet/202...
Had a fun time revisiting Heidegger to comment on AI's potential to create massive alienation through AI-driven production -- making us see the world of human creativity as a means to an end, rather than an end in itself
#AI #tech #philosophy #bigtech
www.instagram.com/p/DR40cuBDot2/
Yesterday, at my university's Academic Senate, we raised critical questions to the administration about the CSU's $17m purchase of ChatGPT-Edu.
This blog post contains the questions as well as lengthy context on AI and education
ricknabb.medium.com/my-universit...
#AI #Education #HigherEducation
Came across this interesting article a bit late: edsource.org/2025/ai-assi...
CSU used AI in a legal filing as it fights against RAs joining the Employees Union & it hallucinated a court case in its argument
Embarrassing for the system pushing to be the "AI leader" in higher ed
#AI #highered
Well this article is a nightmare.
When your technology is so unpopular that you have to force usage of it through firing the unwilling and hiring proponents, you have an issue.
WSJ of course pushing forward the fear of retaliation on behalf of Big Tech
www.wsj.com/tech/ai/ai-w...
#AI
This monster also promoted the invasion of Iraq, which killed hundreds of thousands of people and created mass child malnutrition.
Have others noticed things like this? Would love to hear other examples of this in the wild.
#AI #bias #youtube #creators #genAI
Especially when it comes to views that are critical of AI, we should expect corporate-controlled GenAI systems to be biased towards AI -- why wouldn't they skew their models? It's entirely in their business interests.
There is no way to know if these suggestions are skewed by biased training data (towards pro-AI views) or by intentional corporate manipulation of models. Either way, it's unacceptable. Relying on tools like this will make an effective channel for propaganda, and for creators to spread it
The "outline" it gives talks about "debunking myths of automation's limits and perceived threats" and "how embracing [humanity's unique edge] will unlock new levels of innovation"
The "hooks" it gives are totally framed in a pro-AI lens: Don't think AI will take jobs, here's how it will actually make life better.
An excellent example of AI bias and how it stands to influence thought & media production.
This is a GenAI tool in YT Studio to help creators get ideas for videos based on previous content. It gave me "The Future of Work: Automation and Human Purpose," which I expected to be totally different 🧵
Well, two years later, it seems Blue and Red Wolf have come for us in new form. This type of tech, surveillance, and policing is horrible -- the mark of illegitimate rule. We should be interested in fighting for justice around the world because it always comes back to us.
When I taught my course on surveillance, I told students that Palestine was the best example of modern militarized surveillance in the world. It was a warning for what inhumanity we can justify and build tech for.
ricknabb.medium.com/deconstructi...
Several scholars notes how Palestine is a testing ground for Israeli, US, and allied military technologies. Loewenstein's book is one of the best explanations of this system
antonyloewenstein.com/books/the-pa...
@amnesty.org released a huge report in 2023 detailing Blue and Red Wolf's use for setting up digital apartheid systems in Palestine. Soldiers roamed and took photos of Palestinians. At checkpoints, Palestinian faces were checked against the biometric database
www.amnesty.org/en/documents...
The CBP app allows police to log in and use their phone to scan faces of people they detain. The app tells them to detain and refer them to ICE or not.
This is shockingly similar to Blue and Red Wolf, systems used by Israeli military to surveil and police Palestinians
