LLMs

A guide on getting productive with Claude Code, Anthropic’s agentic coding tool. This is an update to my original post from mid-2025, incorporating newer patterns around plugins, issue tracking, and skills. Installation You’ll need an either an API key or a Pro/Max subscription. On first start, claude will ask you which you want - I recommend the latter to avoid unmanaged costs. To use a subscription, be logged in to https://claude.ai in a browser and follow the instructions. ...

By chance I read two articles at a similar time that join up nicely. First an academic paper The potential existential threat of large language models to online survey research, by Sean Westwood.1 Westwood demonstrates that autonomous AI agents can complete surveys with a 99.8% pass rate on attention checks, logic puzzles, and other standard bot detection methods. His bot maintains a consistent persona, can remembers its prior answers and calibrates its vocabulary and spelling errors to match its assigned education level. It strategically refuses superhuman tasks to avoid revealing itself.2 and only a small number of synthetic respondents can be sufficient to flip a poll’s predicted winner. ...

On Tuesday I did a talk to our internal AI group, along with talks by excellent colleagues, on Context Engineering. The very next day, Anthropic released a blog post on Effective harnesses for long-running agents and updated their prompting best practices guide. Note to Anthropic: please can you release relevant material the day before I do a talk rather than afterwards! I talked about managing context size, how LLM accuracy declines as the context window fills past 50% and my mental model of LLMs as amnesiac pedants, and how this drives Claude Code features like it’s heavy use of todo lists, summarising context into markdown files etc. ...

On Friday Simon Willison published a look at Jesse Vincent’s1 Superpowers plugin for Claude You can read Simon’s and Jesse’s posts to see what this super cool plugin does, but I wanted to understand how to create a plugin and this is quick summary of my notes. First up, Jesse uses a complex structure with three repos: Marketplace Plugin Skills It’s setup this way to allow for contributions, but if you just want to play with creating a plugin, it’s not needed. ...

Trying out a new format this week. Instead of links, here’s some shorts. Too little for a post, too much for a link. This week we have Power variation in AI data centres Adopting new CLI tools Historical maps Power variation in AI data centres (Link requires registration and full article isn’t free, but there’s a massive preview which is all I have access to) A new to me issue with AI workloads - training runs are very, very spiky in their power use. ...

The Anthropic Cookbook is an excellent resource for learning how to write LLM code and it’s continually being updated, most recently with example code for agents. The Cookbook comes as a set of Jupyter notebooks. I prefer Marimo to Jupyter, mainly because it supports reactive cell reloading i.e. if you change an earlier notebook cell that a later cell depends on, the chain of dependencies is calculated and everything is re-run correctly which saves a lot of mental energy and stupid mistakes when working with notebooks. ...

HEPI, a UK higher ed thinktank, released the results from their 2025 survey of AI use by students. There’s a lot in the report, so here’s a few quick highlights. The big headline is that basically all students are using genAI: In 2025, we find that the student use of AI has surged in the last year, with almost all students (92%) now using AI in some form, up from 66% in 2024, and some 88% having used GenAI for assessments, up from 53% in 2024. The main uses of GenAI are explaining concepts, summarising articles and suggesting research ideas, but a significant number of students – 18% – have included AI-generated text directly in their work. ...

Piotr Mazurek and Felix Gabriel have an amazing post up on LLM Inference Economics from First Principles, which I found on Bluesky. They go into a huge amount of detail about how inference works and how that affects processing speed. But I saw the graph above and thought “we can get energy from that”. And so I asked chatgpt-o3: “Looking at these graphs of throughput at different batch sizes on a 4xH100 80gb cluster, what ranges of power per token do they equate to?” ...

Note: This post has been replaced by Claude Code - The Missing Manual, which includes updated patterns around plugins, issue tracking, and skills. Notes that I wrote for work on how to get started with Claude Code, Anthropic’s agentic code tool. Anthropic’s own documentation is good, but I wanted a single page that I could share with people to help them get rapidly up to speed. Once past this stage, their Claude code best practices page is excellent. ...

A very late set of weekly links! This time it’s heavy on AI and it’s impact on research and teaching. Plus cyborg cicadas. First up: A NYMag article on LLM use at Universities went viral (at least in my education and AI heavy feeds!) last week. So that’s not the link as you’ve probably already read it. Instead, here’s Benjamin Breen on how AI makes the Humanities more important and also wierder and Will the Humanities Survive Artificial Intelligence? from the New Yorker. ...