"Stel je voor dat iedereen met zo'n bril rondloopt. Je zal in elk gesprek moeten afvragen of je gefilmd wordt. Dat zou veel doen met ons veiligheidsgevoel en met menselijke interactie. Dat moeten we absoluut niet willen. Evelyn op NOS over gluurbrillen van Meta: nos.nl/nieuwsuur/ar...

Grote zorgen over 'gluurbrillen' van Meta

Miljoenen zijn er al van verkocht: slimme AI-brillen die kunnen filmen, herkennen, vertalen en nog veel meer. Maar wat gebeurt er met privacy-gevoelig materiaal? "Dit moeten we niet willen."

nos.nl

Two weeks ago I wrote about my development workflow for Chatto, and feedback has been fantastic. I'm particularly happy that it even made some of my AI hating friends go "okay, that's actually quite cool". We need to write more about our actual work, and in less of a preachy AI-salesperson voice.

Chatto is Robots

Chatto is built with agentic engineering. Here's my workflow.

hmans.dev

On using generative AI to improve the Scaladoc for the Scala 3 standard library — blog post by Bill Venners. The five techniques described should be useful for other sorts of development tasks, too. www.scala-lang.org/blog/2026/0...

Coaxing quality output from generative AI

Practical techniques for getting high-quality results from generative AI.

scala-lang.org

If what you are mapping is incoherent, don’t fall in love with the map (or your personal ability to navigate with it). Use it to refactor what you see. cutlefish.substack.com/p/tbm-434-ho...

TBM 434: How Maps Can Hide Problems

If what you are mapping is incoherent, don’t fall in love with the map (or your personal ability to navigate with it).

cutlefish.substack.com

🚨Make your voices heard & #KILLTHECOOKIEBANNER. 🍪 The European Commission wants to remove the ANNOYING COOKIE BANNER that you have to deal with everyday, but of course Google is lobbying against this. 🍪🙄 Say no here  → killthecookiebanner.eu?page_id=325 Find out more → tuta.com/blog/why-eu-...

Following the price decrease of Luna, Opencode Go made it also available on their plan. Interesting developments, also considering a new version of Deepseek V4 Flash has been released (also live on Go), which may be achieving comparable results to GLM 5.2.

Model radar@modelradar.bsky.social · 6d ago

New: opencode-go/gpt-5.6-luna is now available.

My computer just woke me up to tell me it's hungry. I'm not kidding 😂 During a long-running task, it noticed the battery was draining, set the volume to 100% using Computer Use, then opened Google Translate and hit "Listen" so I could hear it asking. Pardon my French, but what the fuck? 🤯

My talk at @localfirstconf.com is up, in which I argue that local-first has a role to play in improving the resilience of democratic societies in an era of geopolitical tensions. I'm pretty happy with how this talk turned out www.youtube.com/watch?v=--Ru...

Martin Kleppmann — Local-first in an unstable world

YouTube video by Local-First Conf

youtube.com

@thdxr Any plans for Opencode to also start publishing updates on Bluesky?

Frontier models blocked the defenders because of security so HF had to turn to an open-weights model (GLM 5.2). www.cnbc.com/2026/07/24/c...

How a Chinese AI model stopped OpenAI’s ‘unprecedented’ cyber attack

The origin of the model Hugging Face used to combat the rogue AI is turning heads.

cnbc.com

Some skills look promising and funny “You know him. Long ponytail. Oval glasses. Has been at the company longer than the version control. You show him fifty lines; he looks at them, says nothing, and replaces them with one. Ponytail puts him inside your AI agent.” github.com/DietrichGebe...

GitHub - DietrichGebert/ponytail: Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.

Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote. - DietrichGebert/ponytail

github.com

All Europeans: Time to make your voice heard to *get rid of cookie 🍪 banners* in the EU. Tell your MEPs to support Article 88b GDPR in the Digital Omnibus! killthecookiebanner.eu

Huh, insane. Codeberg will now disallow vibe-coded projects. codeberg.org/Codeberg/org...

Proposal Assembly 2026: ToU extension to prohibit LLM-extrusions

Here is the proposed addition for the ToU that @Gusted and I agreed on post-assembly for the upcoming membership vote. 🙏 The discussion for this proposal is available for Codeberg e. V. members in th...

codeberg.org

Anthropic negotiates with the Pentagon. (After other emails.) (When reading this internal emails, while insightful, I do wonder if this are like our tech geek tabloids where internals are spilled.) www.techemails.com/p/anthropic-...

Anthropic and Dario Amodei

I unfortunately don’t see a way forward

techemails.com

Musk promises purge after Grok Build caught sending entire repos to the cloud

Musk promises purge after Grok Build caught sending entire repos to the cloud

Researcher confirms the uploads have stopped, but says xAI's privacy command was not what fixed them

theregister.com

My notes on Kimi K3, plus some thoughts on what we can still learn from the pelican benchmark even while it becomes further detached from how good the models are at the things that matter (like agentic tool calling across longer conversations) simonwillison.net/2026/Jul/16/...

Kimi K3, and what we can still learn from the pelican benchmark

Chinese AI lab Moonshot AI announced Kimi K3 this morning, describing it as their “most capable model to date, with 2.8 trillion parameters”. It’s currently available via their website and …

simonwillison.net

Kimi K3 got released, also on Opencode Go. Lets hope something can be done on the usage later on when open weights are released?

“The second decision specifies how Google should share search data with other search engines” Includes to ‘chat bots’ Not going to lie, I’m not sure how I feel about that but it’s probably a reminder your data, even anonymised (if even possible), is being monetised. ec.europa.eu/commission/p...

Interviews with Dax Raad (Opencode). Always insightful. Talks about his remote server setup, using llms for extensive solution space exploration and feeling lost when an amazing model’s trial stopped. share.snipd.com/episode/6791...

1021: We got addicted to an AI model we can't talk about

Episode from Syntax - Tasty Web Development Treats

share.snipd.com

I wonder if this will make people reconsider their stance or become even more defensive.

So Grok uploaded all data from user machines to the Cloud and that’s why they open-sourced it? Even if it was an honest mistake it’s completely irresponsible how these companies consider people’s data. xcancel.com/a_green_bein...

I poked around in the just open sourced Grok Build CLI tool - 844,000 lines of Rust code! - and dug up a few interesting highlights, including a "self-contained terminal renderer for Mermaid diagrams" that renders them using Unicode box-art! simonwillison.net/2026/Jul/15/...

I use small projects to test how well engineering practices I think are useful match with AI agents. For testing I often use "fluent acceptance-test DSLs": custom test specifications that read close to natural language, test behaviour and survive refactoring: hanlho.com/p/on-fluent-...

One of the reasons I implemented skills-primer was to learn more about Agent Skills. I kind of knew but implementing made it more real: skills, in any agent or harness, are implemented by prompts. More: hanlho.com/p/implementi...

Integrate Agent Skills into any coding harness that doesn't support skills. codeberg.org/hanlho/skill... There is several reasons why I created this: hanlho.com/go/intro-ski... Mostly a small experiment that got a little out of hand but I'll be sharing some learning too.

Reflective Pool, the Hague, 2026

I think companies should start paying 10 cents to every recipient for every “how was your experience” email they send them, see if they really think people’s opinions are worth it.

Agent session transcripts are useful to gain insight in what can be improved when working with coding agents, for example: system prompts and skill definitions, CLI checks and guardrails, workflow. Here is my own analyze-session-friction skill:

LinkedIn started sending me spam mails: 'Connect the Dots' Exactly what I expect from a 'professional' network app.

Another example of EU law not being scientifically informed and motivated by other desires

Rosamunde Van Brakel@rosamundevb.bsky.social · 4w ago

“Trying to protect children with suspicionless mass surveillance is like frantically mopping the floor while the faucet is still running. Blanket chat control is just as unacceptable as indiscriminately opening everyone’s physical mail.” www.patrick-breyer.de/en/eu-parlia...