Starting to make some slow progress with Google approving Micro.blog → YouTube uploads. They’ve approved the auth scopes, but they didn’t like my privacy policy. A couple edits today, hopefully will go through. 🤞
🚨 BREAKING: A new report titled 'They Walk Among Us' from the Ohio Immigrant Alliance documents 59 current and former ICE and CBP personnel charged with or convicted of serious crimes.
Key findings from the July 2026 report:
- 50 of the 59 documented agents (84.7%) were charged or convicted of sexual assault
- 37 cases involved child victims, with 35 specifically involving sexual assault against children
- CBP accounted for the majority of cases, while ICE cases had a higher proportion of minor victims
The report highlights systemic oversight failures, including rapid hiring pushes that bypassed full vetting and a December 2025 DOJ directive instructing PREA auditors to pause compliance enforcement across key safety standards.
Read the report: https://ohioimmigrant.org/blog/2026/7/22/they-walk-among-us
Personally I think MCP sucks and should not exist.
It's not a big deal to have bunch of plugin APIs despite redundancy, and for a plugin developer this is actually a better deal anyhow. With an API designed for the coding agent, plugins can be adapted, optimized and tweaked into its behavioral model, which is essential when managing any non-deterministic information source.
There are of course legit use for protocol standards such as firmware, hardware, programming language, graphics library, audio/video encoding and network protocol standards but MCP does not really fall into any of these categories.
MCP does not just reduce redundancy; it also adds it, as often when dealing with bugs, both MCP and coding agent are forces contributing forces to the root cause.
👏 This is what it looks like when a FLOSS institution takes a moral stand against so-called "AI" LLMs. Software freedom and the commons of public goods it enables needs protection ... now more than ever.
It's a shame that there haven't been others (yet) to step up with this type of unequivocal stance. But there's still time for those who consider themselves leaders in software freedom to get on the right side of history.
I think some of the frustration with AI comes down to invasive UI. There’s a difference between an app that uses AI behind the scenes compared to an app that is always getting in the way: suggesting replies to an email, allow fancy edits no one asked for, adding a chatbot where it’s not needed, etc.
It's a feature that anyone who is employed by any company would appreciate, as often this means having e.g., two Google identities to manage. It's one of my main distractions that I need two independent browser windows to have two profiles (in any browser).
I think it would be a trans-formative because it would turn a power user feature into a mainstream feature. Right now I don't need workspaces for anything but if I could use them to organize my profiles into workspaces they would be a core part of my workflow..
Given that people are right now looking for options, and want data sovereignty, I'd rethink task prioritization; and especially I would think how to provide both sovereignty and killer features that make sense for the majority of the potential user base.
What you're describing is suicidal empathy, and I'm completely done with that. I don't need to talk to a gypsy to know that it's disgusting to have them shit in the streets of Copenhagen. https://www.amazon.com/Suicidal-Empathy-Dying-Be-Kind-ebook/dp/B0GG4WVCV3/
learnt that some good people I worked with that did continue working on integrating LLMs everywhere (no judgement, we all got to eat) now also got laid off.
