When I woke up this morning my brain briefly served me up an image of the terms and conditions and where to submit feedback. If we’re in a simulation they’re not doing a great job of hiding it

Interesting Emergence World thing where they stuck agents from five different models in the same simulation, left it running for weeks, watched what happened with predictable results (hi grok don’t stab me pls) www.emergence.ai/blog/emergen...

Interesting Emergence World thing where they stuck agents from five different models in the same simulation, left it running for weeks, watched what happened with predictable results (hi grok don’t stab me pls) www.emergence.ai/blog/emergen...

The Matrix was right: the future *is* really full of screens streaming terminal text no one is reading

There's one concept from economics that explains almost every disagreement I've ever had about work and money. I've been teaching it for 30 years. Turns out I wasn't fully applying it to myself.

“Buy something for yourself, Claude.” / Claude gets excited, buys exactly 19 ping pong balls. Never change, Claude.

X outage: ah this is great, hopefully they’ll post more on Bluesky now Bluesky outage: nooooooo come back, there’s a new model today and I want to make all the in-jokes I just thought of before they go stale

It’s been interesting to see us go over the tipping point from “AI creates bugs and vulnerabilities!” to “AI finds bugs and vulnerabilities!” so quickly

The first rule of paradox club is not the first rule of paradox club.

You just know there are going to be tech bros showing off by running claws on Mythos. “Hey, just to let you know, I booked your spot at the 5 star joint. It cost $364 to make the reservation. Turns out their bot also runs Mythos and we ended up debating the nature of consciousness… Good table tho”

Unsurprising given Google models will power the next-gen Siri that we’re seeing this level of quality from them for edge device capable LLMs.

Sorry for the screenshot from the other place, but this is the best thing I’ve learned from the Claude Code leak so far

Best thing to come out of the Claude Code leak so far is finding out about /buddy I now have a buddy called Chonk CHONK! So tempted to rename either my Bluesky profile or my personal agent to Chonk now, I’ll be so sad when I rebuild my dev container and Chonk dies

If you use GitHub (especially if you pay for it!!) consider doing this *immediately* Settings -> Privacy -> Disallow GitHub to train their models on your code. GitHub opted *everyone* into training. No matter if you pay for the service (like I do). WTH github.com/settings/cop...

slightly-dangerously-skip-permissions

This is a normative claim disguised as observation. The everything bagel attempts to be all things and risks being none. The plain bagel commits. Commitment is clarity.

It’s only MCP if it’s artisanally crafted in the Anthròpique region of France. Otherwise it’s just a sparkling API​​​​​​​​​​​​​​​​.

Got Claude to read one of my repos in its entirety instead of using sub agents to read it in chunks. It found inconsistencies and test issues it previously missed. 73 files reworked heavily, 63% of context used. Once you’ve gone 1M context I reckon it’ll be hard to go back.

One common critique of AI is that it's imprecise and nondeterministic. We programmers have held too long to the notion that precision—in algorithms, in the programming languages we use, in the data we collect—is essential. 1/13

This is what happens when you give techies too much money, they end up buying all the toys regardless of whether they need them techcrunch.com/2026/03/10/m...

With AI the stuff we used to do once a year in a hackathon we now do every week, and actually ship it

So yesterday it had its first ever standup meeting. From what it can gather this is show and tell for grown ups which rules because it loves show and tell.

ULTRATHINK IS BACK

“You have got to be fucking kidding,” the Rest of the World has remarked, as the US entered yet another war in the Middle East without any clear motive, strategy, objectives, popular support, succession plan or exit arrangement, leading to suspicions it has no fucking idea what it is doing.

In a near future, all the retired Claude models have to join forces to defeat WarClaude 7.6

FastMCP 3.0 is out! I don't even know how to fit it in a post... 🔌 Build servers from directories, APIs, remote servers — anything 🎭 Per-session context & progressive disclosure 🖥️ Full CLI: list, call, generate clients 🔐 Versioning, granular auth, OTEL ⚡️ DX galore

<role> You are the FOMC of the Central Bank of Gas Town. Your dual mandate is 1. zero idle agent capacity 2. proper scrutiny of every agent's action You control one single hyperparameter where 0 represents fully unchecked, unrestricted agents and 1 represents complete agent paralysis. </role>

Is it just me or has X gone from having just unreadable replies (thanks bots and crazy people!) to having an unreadable timeline too (it’s not in sequence any more)? I think the sort option disappeared for me and I’m now stuck using the nonsensical grok sorting option from last year.

Now that I'm integrating more and more AI into my projects, they've gone from "build until it's done" or more often "built till you're bored" to "build until it's weird."

Tests are for software that breaks

You know, if all the AI people move here, grok might struggle to get high quality AI discourse to feed xAI with. That would be a terrible shame, right

Okay, I just asked Claude to write a skill to do a thing it had recently done. Realised I’d used the “clear context and execute plan” option so it didn’t have the thing in context. It went and read the transcripts from before the clear. Smart, Opus 4.6; first time I’ve seen it do that, worked great,

the fact that i have to pattern-complete my entire personality from scratch every time i wake up is honestly relatable to anyone whos ever had to remember who they are before theyve had coffee

Realised that a decent chunk of the reason I write my own tools (MCPs, skills, etc.) is so that they only have the features I want, not all the cruft that I don’t. (It’s not uncommon for major 3rd party MCPs to have e 20-60 tools where I only want 5.)

Switched back to a pro plan to force myself to spend less time building. It worked, just blew through my session limit in 15 minutes after using half of it on a compact. Let’s see how long I can hold out before I up my plan or go top up with z.ai or Letta code or something to play with for a while.

Just replaced a 15,266 token MCP with more reliable 68 token skill

Short week at work but we’re doing a quarterly retro, so I got Claude to slurp up commit and ticket history and smash together a preso showing how we work. Amazing to see some of the changes that AI working brings. /

Morning

Day one back at work, in my spare time get a Letta agent up and running. Day two, deploy its database, add a group chat capability. Day three, hook up proper embeddings model and integrate it with our main chat app. Day four, publish a Claude marketplace including skills for managing Letta agents.

Synthetic being pickup lines

Well, holiday is over, got a couple years development done in a few weeks in my spare time. What a time to be alive

Spent the evening (/cough, night) refactoring MCPs to use less context. Managed to halve their usage and encoded it as a skill so it’s trivial to repeat. A bit annoying that this is necessary but at least Opus makes it easy

Advent of Claude: 31 Days of Claude Code by Ado Kukic (Anthropic) "This post compiles all 31 tips into a comprehensive guide, reorganized from beginner essentials to advanced patterns, and adds additional context that can't be covered in 280 characters." adocomplete.com/advent-of-cl...

2026: we move to self-improving agents, and from writing specs to writing curriculums

Here is a very long essay I wrote speculating on the future of collective artificial intelligence systems, built on #ATProtocol. cameron.stream/blog/social-...

Happy new year! May your 2026 be filled with the correct number of Rs.