Impressive marketing spin on “our product and deployment strategies are wildly insecure.”

Leaders said there was a “robust process” to ensure the contracts align with Google’s AI principles.

Ah, I think you’ve identified their “robust process” and what the key “principles” are.

Yeah, got headhunted for 10 replace-doctors-with-AI startup for every 1 ed-tech company that even looked at my resume, and the company I’m at now, though good on paper, is squeezing AI into every nook and cranny as fast as they can while sidelining security concerns.

For me, it’s more that it’s a vivid image. I have felt that “immiseration” so it immediately resonates; I don’t need a metaphor. But when people who don’t know technology are gushing about the latest agentic process, I wonder if having a somewhat grotesque, embellished counter will be useful.

I clearly need to up my adblock game. But do y’all also use PeerTube, Nebula, Curiosity Stream? Happy to vote with dollars if there’s a good candidate.

Hadn’t heard “precaritized” either. Brings to mind some penultimate additions to pillow forts, though.

We may start to see people realize that “have the AI generate slop, humans will catch the mistakes” actually is different from “have humans generate robust code.”

And how much power does it use?

Since the headline doesn’t say:

According to data from San Diego Association of Governments (SANDAG) and a city-installed bike counter, annual cycling trips along 30th Street hovered around 50,000 before the lanes went in. In the first full year after installation, that figure jumped to 114,682 rides — a staggering 120 per cent increase. Ridership hasn’t plateaued since. Counts rose another 1.7 per cent in 2023, 5.7 per cent in 2024 and 6.9 per cent in 2025, marking four straight years of growth.

From the article:

At least seven states with a total of about 35 million registered voters have publicly reported the results of running their voter rolls through the system. Those searches have identified roughly 4,200 people — about 0.01% of registered voters — as noncitizens. This aligns with previous findings that noncitizens rarely register to vote.

Maybe they did? It still would take critical time. But I agree it would be totally appropriate.

And Ars published a piece about it — with AI hallucinated quotes attributed to the human maintainer. They have since retracted it.

I was having a discussion related to this with my team at work: some of them are letting through poorly-reviewed AI code, and I find myself trying to figure out which code has had real human consideration, and which is straight from the agents net. Everyone said they closely review and own all the agentic code, but I don’t really believe it.

I wouldn’t have thought of the judicial/policing ramifications.

Plot twist: the server giving worldwide access to send people electrical stimulation was also implemented by Claude.

Cool use of AI for spelunking, though.

Sounds great. I can’t figure out what the status is. Working prototype? Manufacturing?

You can print on standard sheets or paper rolls and choose between black or color cartridges, refillable at your convenience.

Looks like it’s more like NiMH than LiPo, but higher power than NiMH (which I guess lines up with their claims of charging super fast).

Chatbots are terrible at anything but casual chatter, humanity finds.

The one thing the article doesn’t cover is picking filters. Is there just a number, like N95?

Sounds like the mask might be used for paint/varnish fumes while out of lawless Federal officers’ throwing range.

Illegal in the US sense of “for protesters only but fine for LEOs”?