If you use AI for coding, this might save you money.
One of the biggest challenges I've faced when coding with AI agents is token cost. As projects grow, context windows get larger, and the bill can add up quickly.
Robert John | Don’t Fear AI
@trojrobert
Putting ML in production @Adtriba. Working on Data like a farmer work on seed, I used ML to produce the fruit AWS Community Builder Ex @Bosch @MercedesBenz int
- You don't need 50 AI Skills You need one that's written properly Most people write Skills wrong. They dump a wall of instructions into a file. The problem is almost never the instructions. It's the description. It's the structure. It's the missing "why."
- Europe cannot afford another lost technology generation. We missed mobile. We lost cloud. And unless we move faster, we risk losing AI as well. Europe will increasingly depend on decisions made elsewhere, limiting its ability to shape its own economic prosperity, and future.
- Most companies don't have an AI problem. They have an AI spending problem. Last week I stood in front of a room full of builders at LLMday at House of AI in Hamburg and I asked one question: "Do you actually know what makes your AI agent expensive?"
- MiniMax M3 is reportedly hitting near Opus level benchmarks at ~1/20th of the cost That’s not an incremental improvement that’s a shift in scaling economics Why it works: it uses full attention over massive contexts, it uses sparse attention to focus compute only where it matters
