A calmer professional feed

By using AI to assess, we entirely miss the point of what assessment is actually for.

The thing I've found most challenging with AI-driven development is not building full stack, however complex that might be. It's actually the one area that appears to be simplest. I'm pretty ok with aesthetics, having spent several years at the start of my working life as a photographer. But UI (user interface) is unbelievably hard to get right, and I generally find AI pretty poor at it. There are plugins and skills to help, but they all tend towards a bland, Inter font, this-was-vibe-coded look. Nothing wrong with Inter (I use it a lot) but you generally know when a platform has had its inter

So pleased to see things are working well here! I even had a warning from Vercel: "Organic traffic surge from legitimate user activity". So we all must be doing something right.

Darren's little experiment has caused me a bit of discomfort, prompting a little bit of self reflection and I wonder if anyone else is feeling similar emotional responses. On LI, I felt I had 'satisfied the engagement expectation' whenever I clicked Like, and I went beyond the call of duty if I selected 'Celebrate' etc. I don't recall feeling an obligation to reply, as that had been met by one click!

I've been building with AI tools for a couple of years now, with varying degrees of success, but I can safely say that inContact is the first platform I've built that has found its way properly into production. Although it is still very much in its beta infancy, having a platform being used by a couple of hundred people only a few days after launch is a good feeling - if a little nerve-wracking.

When I build with AI I generally have low expectations that it will ever get things right the first time. But here's how I accelerate the debugging process. So much of the success around AI assisted development lies in being able to articulate the precise problem so that the coding agents can fix it. To do this, these 3 debugging tools become your friends:

How do we make data unpalatable for the wrong access. How can we make it unprofitable, unappealing, uninteresting, too much of a hassle for those bad actors. What would the tech version of exploding ink look like? What would deter them? It just seems too easy a crime to get away with.

Not trying to kick Canvas while they are down, but LMS systems are RIPE for disruption and now is the time. I am working with a few schools considering what stays in the new system and what gets replaced. If you were planning/trying to build a better, smarter, easier replacement for learning management systems, what would you tackle first? What would you absolutely NOT include?

A few weeks ago I tried to run and couldn't manage more than 30 seconds without feeling like I was about to die. This week I've done three 5k runs.

Are we unintentionally risking our students’ safety in the name of school marketing? It’s a conversation that started at a recent Google Summit in Bangkok, and later come up again during conversations with good friend James Abela. It's definitely a question school leaders can no longer afford to ignore.

I have spent the last decade on both sides of the same door.

Just read Guy Claxton's piece "Dispositions are Embodied" and it got my mind back into the dispositions space, thinking about whether dispositions might hold the key to the question: "what should be the focus of education for young people in the age of Gen AI?"

Okay, I want to start at the beginning. I have no problem with Gemma Collins. I have no problem with people having views on education. And I have no problem with people talking about the changes that could, and maybe should, happen. Diverse opinions are needed in education, it is the only way to take collective and powerful steps forward.

This morning I`m sitting at my window watching the rain lash down and the clouds thicken as I benefit from my students being on study leave. It feels like a good morning to think about something that's been weighing on me for a while. Education is dividing itself. Im not sure when it happened, but I feel it. I am feeling it in staffrooms, in conversations with colleagues, in the numbers of good people quietly packing up and leaving. And I keep coming back to the same question, what is driving that division, and what is it costing us? What I am seeing saddens me. Record levels of SEND. Record l

Wouldn`t it be great if we pushed for things we know are great for young people with the same energy we bring to AI, technology, and everything else we apparently can`t afford to miss?

If you want to measure the quiet deterioration of modern education, look closely at the contraction of student language.

Over on the other site, I've been sharing my thoughts about different types of AI use as different levels of abstraction of the task. As my first post here, I thought I'd share a quick overview because it shapes how I talk to people about AI use both as educators and in considering their work with students. Level 1 is simply input and output. It's giving a large language model a prompt and a back and forth to iteratively arrive at what you are looking for. Many people start here with the large language model of choice, and either stop there because it's not doing what they want, or they contin

There is a story about a teacher who went to work in a school with no smartphones, no computers, and no digital devices of any kind.

I recently authored a series of booklets as a collaboration between STEM Learning and Voice 21. The outcome is a 3-booklet series called 'Talking about AI'. Booklet 1 is out in the wild, and people are finding it helpful to speak to children about AI. If there are any teachers and leaders out there who would like to give the booklets a spin, sharing booklet 1 here, which spans from EYFS - KS3 (UK system), pre-K to Grade 8 (US system).

I lost a good chunk of an evening to Instagram this week. Not proud of it. You know the spiral. One reel becomes forty minutes and you couldn't tell anyone a single thing you watched.

A band named after the chest pain that precedes a heart attack has become the unlikely poster child of anti-AI music. Angine de Poitrine, a Canadian duo, have taken social media by storm with their surreal, microtonal music. The pair wear bizarre polkadot costumes, masks, and never speak during concerts, preferring strange gestures that their audiences copy. Their guitarist has a unique, double necked instrument with twice the number of frets, enabling microtones, something normal guitars cannot do. Their music is utterly unique and (in my opinion) compelling. I can listen to them and become q

After several people asked for it, I have brought reactions into posts. You'll see three icons below:

I built inContact as an experiment. And am keen for others to join me in it. Over recent months, I've become increasingly aware that LinkedIn has lost its way. An overreliance on ads, suggested posts that seem only designed to annoy (and therefore respond), and a significant increase in performance as we all try to game the algorithm. So out of this frustration I built inContact. I designed it in such a way that it works against the grain. Quality writing (up to 2000 words) will be rewarded. Outbound links are not throttled. Text can be formatted. And there are no ads or suggestions. So far, w

Brilliant to see someone having a crack at trying to build a safe and different style of Social Network. Not only is it small scale and doesn't have the AI-infused writing currently dominating LinkedIn, but its definitely something worth exploring as it looks clean and not cluttered with Ads and endless nonsense. Not an easy task you've set yourself there!

It is a sunny day today, and somehow that makes it easier to see things clearly.

I read a fair few AI papers and was asked to take a look at this paper recently by a client: Oak National Academy's Aila, Education AI Companies and the Anthropomorphising of AI – Sam Clarke (2026)

...because I am capable of delivering better products now without the risk of losing interest or the idea taking so long to conceptualise that the artefact being produced is out of date before it even makes 'production'. As a technology observer, can't call myself an enthusiast on a platform such as this with other genuine enthusiasts(!), I have a rudimentary understanding of what code can do. But, I have ideas way beyond my own capabilities, or so I thought.

"Dear parent, over the weekend we replaced all of the teachers in your child's school with new applicants. We are not sure what experience they have, but we will do our best..." Would you send your child in on Monday without asking serious questions? Now apply the same thinking to your local council. More than 18,000 years of experience may have walked out of English town halls this month! Roughly 2,000 councillors lost their seats in the May local elections. If the average councillor had served for around nine years, that is eighteen millennia of accumulated local knowledge gone in a single T

There is a keyhole garden that haunts me. I built it in the Seoul, with two classes of geography students. The idea was, in my head at least, brilliant. I would take learning outside, do something real, something physical, something that connected curriculum to the actual world. The idea was to build a keyhole garden. We would feast on the things we grew and delight in the flowers we would grow. They`d learn about sustainability and I would create a generation of gardeners. The kids would love it. The kids were wonderful. Enthusiastic, willing, genuinely excited. What they could not do was bui

I have been building tools with AI for a while now. AILitKit is the latest one. And every time I sit down to make something new, I keep coming back to the same two thoughts.

The most difficult part of building my own education business has been the “business” side of my business. Branding, positioning, sales, design, marketing, social media etc. These small, and in my view uninteresting, decisions that sit around the work itself, are ones I have often tried to ignore. I have convinced myself that they are unimportant, distracting to my core and will potentially put off the people I seek to engage with. However, in those moments of pause and reflection, I recognise that they often determine whether the work is noticed, understood, trusted, and shared. As I seek to

OpenAI 's newly released #ChatGPT for #Clinicians - Here is my take: Overall very impressive! I usually caution medical residents regarding using AI tools in clinical decision making, but this one is by far one of the best I have encountered. Although I must add that when I asked the model to solve some standard insurance and social dilemmas for me, it went back to behaving like good ol' ChatGPT. 1. I am happy to see that unlike most supposedly clinician-only platforms, this one is asking for NPI (which is public information) plus an additional step for verification. 2. You are using GPT-5.4!

I'm pretty early in a journey that explores AI for more than just the document write/review/edit/summarise style skills. I'm now using a combination of Claude coupled with Cloudflare, GitHub, Vercel and Supabase as a way of starting to build things. I think I've had some great successes - building my own website and housing the tools I've been developing - primarily a set of business development tools (I've been exploring the niche areas that enterprise tools tend to have as gaps).

I post a lot of my Videos and training on platforms such as LinkedIn, so let's see what it looks like when posted here. I’ve been experimenting with Canva's Magic Layers for a few weeks now and it solves a problem many of us run into with AI-generated images and infographics: You generate a great image… but then you can’t properly edit or rearrange individual elements afterwards. (yet, unless you re prompt the entire thing) Really hope we see some editing coming to the big generators. With Canva Magic Layers, Canva scans the image and attempts to separate objects, text, characters, and design

I’ve been reflecting a lot recently on the conversations many schools are now having around AI literacy and digital literacy. Increasingly, the same argument seems to surface: schools are businesses , and therefore we must capitalise on our advantages: better teaching, better resources, better access to technology in order to give our students an 'edge' over competitors.

Has anyone here read Empire of AI ? I've been listening to it on Spotify off and on for a couple of months (not quite finished) and I keep finding myself hitting pause at times and then, thinking about it all.

Sharing with people here who have an interest in speaking to children about AI - a new 3-part booklet series called 'Talking about AI'. This was a collaboration between STEM Learning and Voice 21, which I authored. Booklets 1 and 2 are in the wild, and you can access a free copy here: https://community.stem.org.uk/blogs/shameel-khan/2026/04/20/what-can-oracy-and-ai-do-for-education

That title is the core principle behind my 10 tips for using AI in college writing instruction, featured in Larry Ferlazzo's Classroom Q A column in 𝗘𝗱𝘂𝗰𝗮𝘁𝗶𝗼𝗻 𝗪𝗲𝗲𝗸, alongside two other educators. (Link below.)

Tonight at Google I/O, Sundar Pichai told the room that AI Mode in Search has passed one billion monthly users in under a year. AI Overviews are at 2.5 billion. The new intelligent search box, which expands as you type and reasons across text, images, files and video, is rolling out to everyone, free, today.

Prelude