Amp Code has 'optimized agents' (Oracle, Librarian, etc) each with what the Amp Code developers think is the most suitable model for the task. I have not tried Amp, there is only so much time, but yesterday I learned about Oh My Opencode with its ‘curated agents’ and more: github.com/code-yeongyu...

Opencode is the first CLI agent that, IMO, competes with Claude Code on features. It lacked Skills until recently, but it had LSP support, customizable shortcuts/themes, and custom agents before Claude Code. Plus, you can use almost any model and use provider plan (like Claude Code's).

Blogged: My Best Technical Decision of 2025: Moving to Proton for Peace of Mind hanlho.com/p/my-best-te...

In the past days, roll-out has started in The Netherlands: users of Instagram and Facebook get the option to only see messages of the people they follow. May the rest of Europe and other platforms follow soon.

Yeah, I'd be pretty furious if I got spam email from some "AI agent" thanking me for my contributions too I dug into what happened here, turns out it's an experiment called "AI Village" which unleashes all sorts of other junk emails on the world: simonwillison.net/2025/Dec/26/...

Opencode now also supports skills. Even Claude Code compatible.

Mistral Vibe CLI just added agent skills support as well: github.com/mistralai/mi... They are moving fast. (Earlier I opened a small feature request and it got implemented the next day.)

A new site on ‘skills’. Includes specification and indicating support across a lot of agents. agentskills.io PS: I wrote ‘indicating’ because at the moment of writing skills in Codex are expected but not in yet. Opencode supports it via plugins (not tested yet).

Wrote a short experience report on using Zed, Opencode and Claude models to implement a missing feature in a personal project of mine. Key elements of success: always plan, structured approach to testing using DSLs. hanlho.com/p/implementi...

I am really enjoying having some time between contracts at the moment... I just wrote an article about the test desiderata - this has been on my mind to write about for several years at this point! coding-is-like-cooking.info/2025/12/test...

Opencode `/share` command copies your session online and give a public URL to share. Actually useful when writing a blog post and you want to share how you prompted. Maybe too much details (watch out for sharing sensitive information) and not as structured as I would like it to be but still.

Just published: Composable Claude Code System Prompts Currently exploring the simplest and most effective approach for composing prompts and other types of content so I can focus on the work itself. I don't know if this is the best way, but it works. nicktune.medium.com/composable-c...

can't stop watching this clip of a tesla Optimus teleoperator taking his headset off before properly logging out the robot

I made a little prioritization tool johnpcutler.github.io/prioritizer/...

This is an increasing problem that needs regulation. The big tech co’s have become so big, so powerful, but also so careless. All while we’ve trusted them with more and more of our work and lives. Which they can take away in an instant, “without recourse.”

Teams collaborate constantly, but not always for the right reasons. That’s why coordination feels heavy. Interaction modes help teams choose the right way to work: 🔹 Collaboration for discovery 🔹 X-as-a-Service 🔹 Facilitation When intent matches interaction, friction falls.

Where has this been for all my CLI life: github.com/charmbracele...

“Linux Foundation Announces the Formation of the Agentic AI Foundation (AAIF), Anchored by New Project Contributions Including Model Context Protocol (MCP), goose and AGENTS.md” Includes Anthropic so we may be getting an Agents.md in Claude Code…aaif.io/press/linux-...d/

Giving the Mistral Vibe CLI a spin ...

Put together some notes on Mistral Vibe, Mistral's new CLI coding agent (think Claude Code / OpenAI Codex) which is Apache 2 licensed, written in Python+Pydantic+Textual and has a neat set of system and tool definition prompts that are fun to read simonwillison.net/2025/Dec/9/m...

Decided to give Opencode a more serious try. One nice find: if you have subscriptions to both Claude Code Pro/Max and ChatGPT Plus/Pro you use Anthropic and Codex models in Opencode benefitting from their daily and weekly limit plans.

These architects are attempting to maintain system integrity across the organisation while trying not to block the teams' development flow. The challenge lies in shifting the approach from dictating to facilitating decision-making. ...

Missed Signal’s President @meredithmeredith.bsky.social on #Buitenhof? About #Signal, the power of Big Tech, and why privacy matters. Watch it back here🍿: www.youtube.com/watch?v=FTD5... Not using Signal yet? Try it out for free: signal.org/install #signalapp #signalmessenger #privacy #whatsapp

I've been experimenting with AI-assisted development where most of the code is actually written. However, describing this as 'vibe coding' did not sit well with me. CHOP, "Chat-oriented programming": LLM-assisted programming that is almost the polar opposite of "Vibe Coding".

While I was reading I kept thinking, isn’t this like Rust’s borrow checker and indeed: “Given the foundation provided by the capture checker, the Scala team is now working on using it to implement separation checking. This will allow tracking the use of mutable values and the aliasing of references”

Also, Zed has been supporting rainbow _indents_ for a long time. Below is how you configure those and what it looks like:

Easy to get caught up in optimising flow and tool usage of the agentic coding tools out there (ask me how I know). Here is a refreshing no-fluff read from the trenches: steipete.me/posts/just-t...

Google has now admitted that they will tap into users' personal data within Gmail and Drive in order for their upcoming AI search to "be more helpful". For years, Google has advertised their services as "encrypted" and "secure", which begs the question: are they really? 1/5 🧵

Ready to use ChatGPT Atlas? Please don't‼️ ➡️ It can be hacked by reading websites ➡️ It reads everything you're logged into: your email, your CRM, your bank account ➡️ "Delete" doesn't mean deleted ➡️ "Incognito" mode isn't private ➡️ GDPR/compliance nightmare tuta.com/blog/dont-in...

Anthropic currently provides limited time, until Nov 18th, $250 on Pro, $1000 on Max plan on their Claude Code web at claude.ai/code or from the app. Fun and worth to give it a try. Works well. But it feels too easy to add new features. Found my self vibing a little too much to my own liking.

New on the blog: Think for Yourself "You're about to commit a chunk of LLM-generated code into your product's codebase. Before you do, however, pause to consider and act on these questions." kevlinhenney.medium.com/think-for-yo...

We're so relieved to see Germany reaffirm its opposition to the dangerous Chat Control proposal--the one that would mandate mass scanning of communications. Germany's long been a solid champion of privacy, and the news that it was considering backing mass surveillance was alarming. 1/

For me, DORA’s annual research is always worth paying attention to, most importantly because of its scientific grounding and its practical insights that keep on evolving with our field. DORA Report 2025 Infographic: dora.dev/research/202... Full report: dora.dev/research/202...

After wrapping my head around the constant changes in LLM subscriptions and performance, here’s my new coding stack for October: - Warp Pro: my go-to agentic CLI, supports most top end LLMs. - Zed Pro (using $20 trial) - OpenCode for agentic CLI dev (local + cloud models, easy swapping, no lock-in)

📣 Germany's close to reversing its opposition to mass surveillance & private message scanning, & backing the Chat Control bill. This could end private comms-& Signal-in the EU. Time's short and they're counting on obscurity: please let German politicians know how horrifying their reversal would be.

We are alarmed by reports that Germany is on the verge of a catastrophic about-face, reversing its longstanding and principled opposition to the EU’s Chat Control proposal which, if passed, could spell the end of the right to privacy in Europe. signal.org/blog/pdfs/ge...

Claude Code 2.0.0: `/rewind` killer feature Anthropic's models are among the best although, as in the past weeks, wonky at times. But with their CLI they currently keep on having an edge on the competition. (Disclaimer: I have not tried all of them but I know Codex and Gemini lag.)

📣 NEW -- In The Economist, discussing the privacy perils of AI agents and what AI companies and operating systems need to do--NOW--to protect Signal and much else! www.economist.com/by-invitatio...

When it comes to exploring different decision recording formats quick experiments are easy to do with Claude Code. For example, I've always found it difficult to write a Y-statement as a shorter alternative to the typical decision records in markdown.

Been experimenting with having Claude Code help me write decision records in ADR format after our technical discussions and use them as basis for further implementation. hans.lhoest.eu/grounded-dec...

ccstatusline: very helpful to get insight in Claude Code's context statistics. Most useful for now: how much time is left in the 5h window (Block Time)? github.com/sirmalloc/cc...

Wrote an article: hans.lhoest.eu/how-to-use-j... #jj-vcs #claude-code

TIL: you can add 'logic' directly to Claude Code's system prompt. E.g. to detect which VCS: ``` Before ANY commit-related commands, you MUST: 1. Run `bash: if jj root >/dev/null 2>&1; then echo "jj"; elif git rev-parse --git-dir >/dev/null 2>&1; then echo "git"; else echo "none"; fi` ``` ...

Software architecture humblebundle www.humblebundle.com/books/softwa...

Claude Code has git pretty much baked into its system prompt. Tried quite a few things to make Jujutsu ('jj') work with Claude Code and this seems to work:

Claude Code keeps on referencing itself in commit messages no matter where I tried to instruct Claude to stop doing that automatically. Turns out it is part of the system prompt so it pretty much baked in: - Create the commit with a message ending with: 🤖 Generated with [Claude Code](...)

Was asked: What are the pros and cons of an "Agile" approach? Here's my list, but bear in mind that I am NOT talking about Scrum or SAFe or Kanban. I don't see those frameworks as even slightly Agile, at least not inherently. 1/7

Git and #jj-vcs: in miniature lottia.net/notes/0013-g...

Context engineering shows interesting potential to ground documentation to actual code, or as how I sometimes refer to it: reality.