From capability to purpose

Four Directions of Ambition

You have a model, a task and a way to check the work. What should you attempt? Sometimes the useful gain is time. Sometimes it is a better goal, a deeper investigation or a method you would not usually borrow.

Adapted from Felix Rieseberg (2026), who describes four types of ambition: Speed, Altitude, Depth and Range. The explanations, worksheet and workplace examples here are TFIS applications of his framework.

Choose the directions that help your task. You can combine them, but using all four is not a requirement. Keep one human owner responsible for the goal, the evidence and any commitments.

One report. Four possible gains.

Imagine preparing a monthly business report. These fictional alternatives change different aspects of the work. Decide what would count as an improvement before asking AI to help.

Speed

Complete the same work in less time.

Can I reach the same standard with less effort?

Prepare the existing monthly business report faster.

Check the gain. Compare total preparation, checking and correction time against the manual report, using the same acceptance standard.

Altitude

Move from the individual task to the outcome it serves.

What should happen because this work exists?

Use the report to identify which operating decision needs attention, then test whether a shorter decision brief serves that purpose.

Check the gain. Agree the decision and acceptance criteria first. A changed goal needs its own baseline; it is a different task from formatting the old report.

Depth

Inspect the underlying data, tools and problem using professional methods.

What would an expert inspect, and what would show the result is right?

Trace a disputed metric to its raw records, inspect definitions and missing data, and check the analysis with a suitable tool or qualified reviewer.

Check the gain. Retain source records, assumptions, checks and unresolved uncertainty. Expert vocabulary or the model approving its own answer is insufficient.

Range

Bring relevant knowledge, tools and techniques from other fields.

Which other discipline has a useful method for this problem?

Borrow an operations method to identify a reporting bottleneck, or an editorial method to make the decision brief easier to inspect.

Check the gain. Name the borrowed method, why it transfers, where the analogy breaks and a test that could reject it. More personas do not establish better work.

Write a brief you can inspect.

Take ten minutes with one task you understand. Copy these prompts into your own notes. Use an approved tool and information you are allowed to share.

  1. Task and baseline. What do you do today? Record the current output, quality standard and total human effort.
  2. Intended outcome. Who needs the result, and what decision or action should it support?
  3. Useful direction. Choose Speed, Altitude, Depth or Range. Explain why it matters here. Leave the others aside unless they serve the goal.
  4. Evidence. What result would show improvement? Name sources, a check and a result that would make you reconsider.
  5. Boundaries and owner. Name the responsible person, allowed inputs and tools, time or cost limit, and the point where work should stop or ask for a decision.
  6. Review and retain. Compare intended and observed benefit. Include failures, review time and uncertainty. Save a useful correction with its source and conditions for later reuse.
A fictional first brief: “Our monthly report takes two hours including review. Test Speed on one approved sample: produce the same report with every figure traceable to the source sheet. I will check each figure and record preparation plus correction time. Stop if a source is missing. Prepare a draft; I decide what is shared.”

If the result is faster but wrong, it has not met that brief. If investigation reveals that the report serves the wrong decision, write a new Altitude brief and evaluate the changed goal separately.

Place ambition in your practice.

Ambition chooses a direction for your goal. SSA-CMM describes current capability. The five pillars of sovereignty support responsible execution. Anthropic's four fluency decisions—Delegation, Description, Discernment and Diligence—guide how you work with AI. These serve different purposes; this exercise adds no maturity score.

  • Fluency: select a recurring task and apply the four fluency decisions.
  • Agents: define a bounded workflow and collect evidence of its outcome.
  • Accumulated Curiosity: connect inquiry, retained knowledge and demonstrated capability.
  • Loom: carry task intent and evidence into a context pack, then review the result.

Source

Felix Rieseberg. (October 13, 2026). How to be more ambitious: A tutorial for humans with models. Updated October 16, 2026; accessed October 22, 2026.

The article supplies the four types, the goal beyond the task, the emphasis on raw evidence and professional methods, outside validation, and cross-field techniques. TFIS supplies the examples and exercise on this page. This is practical guidance; it does not validate an assessment or guarantee productivity gains.