An Anthropic engineering manager's four kinds of AI ambition, his favorite prompt, and one prompt to use all four on your own work.
Felix Rieseberg works on Claude at Anthropic. He has led engineering for Claude.ai, Cowork and Claude Code Desktop. In October he published How to be more ambitious: A tutorial for humans with models.
His point: AI is great at some things and weak at others, and most people never find out what it can really do because they ask too little. Here are his four kinds of ambition, with an example for each.
01
Speed
The same work, in less time.
In his words, "the bread and butter for AI's economic value." Most of us already do this, so he skips past it. It's the floor, not the goal.
02
Altitude
Ask for the outcome, not the task.
"Going up in abstraction and considering the outcome rather than the individual task." Ask what you actually want to happen after the task is done. His example: instead of "make the quarterly review deck," ask "tell me what changed this quarter, why, and what we should do about it."
03
Depth
Do it the way a top expert would.
"Considering a task the way a field-leading expert would." Experts start from raw data, test their own assumptions, and use real tools. His example: "We believe churn is an onboarding problem. Try to reproduce that from the raw data."
04
Range
Bring in other fields.
"Drawing on multiple fields rather than just the one the task might need at the surface, using knowledge, tools, and techniques you normally wouldn't." My example: a marketing problem might need statistics, psychology and design, not just marketing.
His trick: turn the noun into a verb. "A deck" becomes "getting busy people to believe a story about cause and effect." Then ask which experts are good at that.
This is the prompt Felix says often helps him. Paste it under any task you were about to give AI. It makes the AI tell you what a real expert would do first, then rewrites your ask.
Prompt by Felix Rieseberg, from his article.
I built this prompt from his four ideas. Paste in a task you're working on and it walks you through Speed, Altitude, Depth and Range, then hands back one upgraded prompt.
Start with one thing on your list this week. Ask yourself: could I aim higher, go deeper, or bring in another field?