Apple is adding video support to its podcasting apps — that’s great. But the way it’s doing so might jeopardize the openness that’s kept podcasting from getting enshittified like every other form of media. It’s not too late for them to fix this. https://anildash.com/2026/02/28/apple-video-podcast-power/
Everybody is (understandably) very excited that Anthropic isn't selling their AI for Hegseth to use to automate his war crimes. But I would gently suggest that we not accept setting the bar quite so low. https://www.anildash.com/2026/02/27/a-cookie-for-dario/
One of the things that's most frustrating about the hype cycle around OpenClaw is that so much of what people are excited about is just... software? You can automate doing lots of cool stuff with your inbox with deterministic software that's been around for decades, and it won't delete your inbox! And much of it is easier to install than OpenClaw is!
Okay, so I've been talking about the ways that Big AI directly harms kids. But what can we *do* about it? Here are some things you can do, at work and and your kid's school, to take action today, complete with scripts to follow: https://www.anildash.com/2026/02/23/taking-action-ai-harms/
@pattykimura@knowprose@Katiemorag certainly anti-Blackness was a poisonous and immoral path for them to choose in their pursuit of citizenship; only solidarity can lead to freedom. I did a talk with this as a central pillar a decade ago: https://youtu.be/y1SrFctEYY4
It is a vile scandal that 2 of the biggest AI companies in the world attack children by inciting them to self-harm, and by generating sexually exploitative imagery of kids for profit. I wanted to document this so that people in tech will stop saying they don't know: https://www.anildash.com/2026/02/18/threatening-kids-with-ai/
The follow-up will be about how we take direct action to hold companies like OpenAI & Grok accountable, but this first record of their actions is designed to be clear enough for your colleagues or acquaintances who may not be familiar with the threats that these platforms have leveled towards kids.
@elithebearded he didn’t have a current will at the time he died, because I don’t think he expected his prescription medicine to ne poisoned with fentanyl.
@carnage4life the idea that we’re going to return to the pre-Trustworthy Computing era, but with *intentional* remote code execution built into the system is unfathomable. I get why young devs who grew up in A16Z’s bubble don’t know better, but Omar is getting lost in hype here — if he wanted a pluggable natural language platform with “Sonos API” on one side and “iMessage” as an interface on the other, the additional risks of openclaw would be an unacceptable set of tradeoffs to get that.
@carnage4life so I’ve been playing with doing this stuff without OpenClaw because I don’t trust its trust model (and building from smaller building blocks is appealing at this stage, even if you lose network effects of reusing others’ work). A huge thing that folks have missed is how much functionality *should* be happening with deterministic code, even if that’s speaking to an MCP endpoint. Something like n8n + MCP with a thin final layer of very-locked-down LLM would be so much safer).