Journal on Tend and Flow

Skill Safety

Tue, 18 Aug 2026 08:15:42 +0100

OWASP (opens in a new tab) recently published a community-driven project to catalogue the current types of threats posed by skills, calling out the high level of malicious skills being shared publicly out in the wild. Their security checklist is here (opens in a new tab). This is version one, written for developers and technical staff, and will no doubt change and improve over time.

I thought there was real value in having a non-technical version. In many organisations, non-technical users are writing their own skills and, most likely, sharing skills that others have written. Feel free to repurpose my checklist below.

AI Transparency Is Not Enough

Wed, 12 Aug 2026 12:44:31 +0100

The general direction of AI legislation is to make AI use more transparent and maintain a clear line of accountability for the content. The EU’s AI Act is now in force and while it mostly talks about the responsibilities of organisations making and deploying AI systems, it does also cover businesses using those systems (‘deployers’). In a very broad summary of the Act (opens in a new tab), deployers need to (a) follow the provider’s instructions in using the system, (b) ensure a sufficient level of AI literacy among staff involved in the operating of the AI system, and (c) inform individuals when interacting with the system (“unless it is obvious from the context”). That last point extends further to specific use cases involving media content and so-called ‘high risk’ systems being used to make decisions involving the public. The EU law now sets the baseline for transparency, but it is not governance.

The New Site Is Live

Mon, 10 Aug 2026 12:03:50 +0100

I just finished a retrospective (opens in a new tab) with Claude and asked it how I did. It started as a funny idea to see what would happen, but it actually turned up some useful stuff!

Some context

Last week, I rebuilt the site for Tend and Flow. It had been a placeholder and mid-Summer seemed like a perfect time to do it properly. I used Opus 5 in VSCode as my main workbench (everything is happening in either VSCode or Cowork these days!). Based on my brief to make the attack surface as small as possible, we ended up deciding to orientate around markdown files (it’s a low volume site). I used Claude Design to decide the type, palette and wireframe. We built the first version to check the content workflow. Then, again in Design, created and implemented a more thematic aesthetic based on my branding. After the release of v1, I created a specific maintenance skill for ongoing checks, did the tidy up and compacting. Nothing particularly innovative, but it took about three days (of which 60% of the time was really spent on content).

Designing Governance

Thu, 06 Aug 2026 17:03:57 +0100

The flood of policy dressed up as governance

I have become troubled by how much AI governance turns out to be policy wearing a framework’s clothing. Something goes wrong, or nearly does, and the response is reliably the same: a new document, a committee, a tweaked RACI chart. Everyone relaxes. Governance has been done.

Except it hasn’t. Nobody opens the document again until the next incident forces them to. This is not laziness, the people writing these things mean well and work hard. It is a design failure, and it becomes considerably more expensive the moment you point an AI agent at a real business process.

Vacations in Chaos

Thu, 30 Jul 2026 10:40:00 +0100

Chinese AI Takes the Stage

Fri, 24 Jul 2026 12:17:44 +0100

Background

This week we read about the Chinese creating a new global forum ‘World Artificial Intelligence Cooperation Organisation’ WAICO (opens in a new tab). A genius move to be simultaneously (a) using state funding to gain advantage, (b) leveraging the gains made by the US, and (c) inviting the global south to join them and share open source models and compute. Forces the US to be protectionist (the funding model demands it), which hastens the collapse of the US’s lead in AI. While reading this, I’m also watching videos of Unitree’s (opens in a new tab) amazingly capable robots. This could be the moment China takes the lead on AI and robotics. Assuming it might be (at least directionally), what’s the future of rights and personal freedoms? If it is distinctly Chinese flavoured, what does a once-libertarian society look like in 100 years?

Hot and Humming

Sat, 18 Jul 2026 12:46:24 +0100

Background

In the midst of a UK heatwave, I continue to read news reports that data centres raise the surrounding temperature (heat islands) and also emit a humming noise. Together with the political battles about planning permissions between local communities and national government, this is a global mess. I don’t want to live next door to a hot and humming data centre, and neither do you. So how do we get this back onto a sensible, long-term, sustainable development path?

AI Bubbles

Sat, 11 Jul 2026 12:58:10 +0100

Background

This week, I think I must have read (or listened to) about 5 well-researched, convincing articles and podcasts talking about the irrational valuations in AI and infrastructure companies. Having lived through property bubbles, web 1.0 bubbles, telco infrastructure build-out bubbles, bank bailouts, etc etc…you can’t help but think we’re in an AI bubble right now. And if so, when does it pop? I imagined a mischievous, AI-personified Pandora (of ancient myth (opens in a new tab)) having a thoroughly good time. Are her gifts good or bad? Bubble or not?

World Cup Independence Day

Sat, 04 Jul 2026 13:13:57 +0100

Background

It’s World Cup season and with it comes the stress, the pain, the anguish, the celebration. I’m no football (soccer) pundit but it’s entertaining. And for the World Cup - hosted by the US (and Canada and Mexico) - to happen on the occasion of the 250 year anniversary of the USA’s Day of Independence from their colonial oppressors (who invented the game of football). Glorious. How determined one country has been to be the best, at everything, at any cost while thinly disguising some pretty structural weaknesses. ‘O say can you see…’

Is AI Dismantling the Skills Ladder

Sat, 13 Jun 2026 19:45:01 +0100

“Apprentices spend months just learning how to wring out hot towels. After that, they are taught how to prepare the fish. And then, finally, they are allowed to cook the rice. It is not an easy path.”

Jiro Dreams of Sushi (2011), IMDB

The framework is this: mastery develops in three stages. The Apprentice follows instructions - recipes and algorithms, the explicit steps that produce a known result. The Journeyman has internalised enough of those recipes to begin modelling behaviour - adapting, blending, responding to context with a working knowledge of what good looks like. The Master operates from principles. They are no longer following a map or copying a pattern. They are navigating from a deep, internalised understanding of why things work the way they do.

Trillionaires

Sat, 13 Jun 2026 13:50:56 +0100

Background

This week, with the IPO of SpaceX, we have the occasion to (celebrate?) the world’s first trillionaire. With so much inequality in the world, it is frankly disgusting that our society and systems have allowed this situation to happen, and glorify someone who believes the future is somewhere other than earth.