Step 1. Build “AI” (LLM models) that can’t be trusted, doesn’t learn, forgets instructions, and frustrates software engineers Step 2. Automate the use of these

Step 1. Build “AI” (LLM models) that can’t be trusted, doesn’t learn, forgets instructions, and frustrates software engineers Step 2. Automate the use of these LLMs into “agents” Step 3. ??? Step 4. Profit