Daniel Morgan

An independent weekly on applied AI. Est. 2024

Your weekly AI
changelog.

Each week I go through the AI release notes, papers, and launches and write up the bits that are actually useful at work. Five minutes to read.

ai.daniel-morgan.com / rankings
This week

Claude Opus 4.7 retakes #1 in coding; Nano Banana 2 overtakes in image.

Three rankings moved this week. One model got quietly deprecated. Full scores and a prompt I've been using, inside.

Live rankings · Δ week
01Claude Opus 4.7Coding+2
02GPT-5.4 ProResearch+0
03Nano Banana 2Image+3
04Dreamina Seedance 2.0Video−1
Latest
models tracked
1,000+
hrs testing AI
Weekly
issues
What's inside

What each issue covers.

01 weekly

Releases

A short note on each new model, update, or launch from the week, and whether it's actually useful to reach for.

02 weekly

Papers and posts

A few papers or posts I found interesting that week, each summarised in a paragraph so you can decide if you want to read the full thing.

03 weekly

Prompts and workflows

Prompts and workflows I've been using at work recently, with what's gone well and what hasn't.

04 weekly

My take

What I think is overhyped this week, what's underrated, and what I'd skip.

Daniel Morgan
Daniel MorganSenior Engineer · London
The author

Ten years shipping software.
Now writing weekly about applied AI.

I'm a senior engineer in London. I've shipped AI into production at three companies and I spend a lot of time testing models against real work. Each week I write up what's changed, what I've tried, and what I think of it.

From the blog

Recent articles.

Browse archive →