In practice, for many useful applications, many of the various obvious problems with AI agents (drift, hallucination, compounding errors) are more solvable than they are in theory Clever prompting, tool use, constrained topics,
LLM judges & organizational process close some gaps
PROMPT ENGINEERING
-
Practical Solutions for AI Agent Reliability and Error Reduction
By
–
-
Space Curvature Technique Enhances AI Model Prompting
By
–
"space curvature" also has a nice effect in prompts pic.twitter.com/B5G0ssaarH
— fofr (@fofrAI) 28 juin 2025"space curvature" also has a nice effect in prompts
-
Non-technical experts excel at AI prompt engineering and product development
By
–
Many of the best prompters I have met who are creating actual useful products in organizations are not technical. In fact, coders often struggle with non-deterministic systems in a way that teachers and managers do not. Broaden AI development beyond engineering.
-
Two Paths to Mastering AI: LLM Understanding and Instruction Design
By
–
All the technical language around AI obscures the fact that there are two paths to being good with AI:
1) Deeply understanding LLMs
2) Deeply understanding how you give people instructions & information they can act on. LLMs aren’t people but they operate enough like it to work -
Creating AI Character Memories and Personality Development
By
–
Recording the show, managing her prompts and memories, all of that is going to be 100x easier now. In our next episode, we talk about how to make her a fully-realized person even though she has very limited history and backstory. And from there we construct her first "memory".
-
AI UX design needs embrace variance and branching exploration
By
–
When there is a lot of natural randomness and discovery in an AI use case (image creation, innovation), the focus should not be on single-threaded conversation that becomes self-reinforcing through autoregression, but embracing variance, randomness & branching. Calls for new UX
-
Grok Solution Similar to Claude Code
By
–
Assuming a solution similar to Claude Code or Gemini CLI but designed for Grok
-
Seedream generates realistic faces from creative prompts
By
–
faces from seedream with prompts like: > 100 yearbook photos
> 100 photos from a party photo booth
> 100 mugshots -
AI Translation Tool for Language Learning: Human Control First
By
–
I used AI in most of my initial conversations with my now wife to translate from English to Japanese, and we're incredibly happy. But, I always made sure the AI was saying what I wanted it to say, not telling me what to say. And only giving me minor feedback. Never allowed it
-
Prompt Engineering for Shopping Deal Evaluation with LLMs
By
–
The prompt was “is this a good deal?” (Prompting is really not that hard for most requests) And it is an interesting case where hallucinations don’t matter that much (though I didn’t find any). Most online shopping guides are out of date, riddled with conflict-of-interest & odd)