Language models are non-deterministic. Given the same inputs, same outputs are not guaranteed, and they can make mistakes. In AI Driven Development, this means we have to treat them a bit like fallible humans: Review the code. Run automated lint and test suites. Test everything end-to-end.

Better Prompting with Pseudocode was my talk from the 2023 Prompt Engineering Conference focused on why pseudocode works so well to prompt language models, especially for AI Driven Development. www.youtube.com/watch?v=TADr...

Prompting AI with Pseudocode improves response scores up to 38%. SudoLang is the most well tested and well documented pseoudolanguages in the industry. I talked to @devtools.fm about it last year. We talked a lot about AI Driven Development. www.youtube.com/watch?v=2EAJ...

Effective AI Driven Development requires a systematic approach. You can teach the AI to write code your way, use your preferred libraries and frameworks, build UI with your components, and follow your best practices. I'm writing a book on the topic. Early access here:

leanpub.com/effortless-p...

The past 18 months have seen the most rapid change in human written communication ever

By. September 2024, 18% of financial consumer complaints, 24% of press releases, 15% of job postings & 14% of UN press releases showed signs of LLM writing. And the method undercounts true use.