I had to do therapy on my AI
The author discusses their experience of trying to train a language model (LLM) to assist with code reviews. Despite multiple attempts to refine the instructions, the LLM consistently ignored key directives. The author reflects on the challenges of working with AI and the need for clear communication.
- ▪The author attempted to use an LLM to assist with code reviews but faced challenges with the AI ignoring instructions.
- ▪After several iterations, the author realized the importance of clear and explicit prompts for effective AI assistance.
- ▪The experience highlighted the complexities of working with AI and the need for a structured approach to communication.
Opening excerpt (first ~120 words) tap to expand
← Tin's Posts · May 21, 2026 · 5 min read I had to do therapy on my AI I've had to do "therapy" on my LLM agent last week. Entertaining stuff. I've been writing a skill (as I do), when I realized the LLM was ignoring my instructions. Here's the setup. I'm working on a codebase that has a lot of boilerplate. We're using certain cloud providers and SDKs, all wrapped in Typescript typing and exposition. Lots of code, lots of verbose typing, all lines of code that describe far less wordy concepts. Stuff that's straightforward enough to review but makes the PR look intimidating and creates a mental load that makes it difficult to start.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Tinthe.
Discussion
0 commentsMore from Tinthe
