Notes
Shorter than a lesson and less structured. Technique that does not fit the curriculum yet, model behaviour worth writing down, and the occasional correction.
- 21 August 2026Technique
"You are an expert" probably makes it worse
The most repeated prompting advice in existence has the weakest evidence behind it, and on knowledge tasks the research says it actively costs you accuracy.
2 min read - 21 August 2026Model behaviour
The prompting advice that stopped working
Prefilling the assistant turn was standard, taught everywhere, and now returns a 400 error. It is the cleanest example of why some technique expires and some does not.
2 min read - 20 August 2026Technique
Asking a model to check its own work usually does not help
The instruction "now review your answer and improve it" feels like it must work. The research says it often makes correct answers worse - and explains exactly when it does help.
3 min read - 19 August 2026Model behaviour
The 3 watt-hour number was wrong by about ten times
The most-repeated figure about AI energy use turned out to be a large over-estimate. What the measurements actually say, and why the correction matters more than the number.
2 min read - 19 August 2026Model behaviour
The strawberry problem is a tokenizer problem
A model that can write a working compiler cannot count the letters in a word. That gap tells you something useful about where to aim your prompts.
3 min read - 18 August 2026Model behaviour
AI detectors systematically flag non-native English writers
The tools institutions use to catch AI-written work misclassify second-language writing as machine-generated. If you write clear, plain English, this is your problem too.
2 min read - 12 August 2026Technique
Why "don't mention pricing" makes it mention pricing
Negative instructions are the most reliably disappointing thing beginners write. The reason is mechanical, and the fix takes one sentence.
3 min read