💡 Core idea
A busy team is not a successful team. Clarity about outcomes — and about what you will not do — is what makes speed useful.
Motion is easy to buy. Tools make it easier still. The manager’s job is to make the aim legible enough that speed has a direction. As tools get more capable, Greg Brockman (@gdb) puts the constraint in one line: “the bottleneck is increasingly knowing what you want.” A team can now move fast in several directions at once. Without a stated aim, that is only motion.
Ant Murphy notes that “value” is often the word people use, not the criterion they actually prioritize. If you cannot say what will be true in the world when the work is done, you do not yet have a goal. You have a slogan. That is also what decades of goal-setting research would predict: Locke and Latham (2002) summarize that specific, difficult goals consistently produce higher performance than urging people to “do their best,” with meta-analytic effect sizes from .42 to .80. Specificity alone is not enough — difficulty matters — and people need feedback on progress, or the goal has no traction.
Outcomes, not activity
Write outcomes: what will be different for a customer, a user, or the business, by when, and how you will know. Output — documents, tickets, features, pages — is evidence at best. It is not the aim.
A one-page team strategy is enough: context, the outcome, the bets, the non-goals, the risks. A goal tree can connect company intent to team aim to individual focus without a paperwork religion. If the line from a person’s week to the team outcome cannot be said in a sentence, the tree is broken — or the work does not belong.
Alignment before autonomy
Maarten Dalmijn argues that without alignment, autonomy produces disorder. This is not an argument against giving people room. It is an argument against room without a shared aim. Publish the aim, the non-goals, and the decision rights. Then get out of the way of the work that fits.
When strategy moves, reconnect the work. A team can hit every milestone on a project nobody needs anymore. That is not a delivery success.
What you will not do
Scope discipline is the quiet half of priority-setting. Keep a kill list and a parking lot: good ideas that are not current work, written down so they do not have to be re-litigated every week.
Itamar Gilad describes a common failure: good ideas die late, when already-started work is cut. The waste is not only the abandoned effort. It is the late decision. Cut earlier, or finish. Do not let work linger in a grey zone where it consumes attention and will never ship.
Jack Vanlightly argues for raising the ambition threshold rather than accumulating more low-value work. More items is not more strategy.
Speed is not a virtue by itself
Graybeard describes speed worn as a moral posture: impatience and half-made decisions dressed up as urgency. Aviv Ben-Yosef warns that a team can be fast at building the wrong thing when the roadmap replaces vision.
James Stanier argues for slowing down in the upstream phases so later speed is real — and, separately, for fixing one bottleneck at a time. Ant Murphy puts a still sharper filter on teams that are already struggling to ship: fix delivery before adding more discovery or innovation theater.
The sequence is ordinary: make the outcome clear, make the non-goals visible, remove one constraint, then go.
Tools
- One-page team strategy. Context, outcome, bets, non-goals, risks. If it needs a deck, it is not clear yet.
- Goal tree. Company aim → team aim → personal focus, in plain language.
- Kill list / parking lot. Good ideas that are not this quarter’s work.
- Weekly priorities memo. At most three team priorities, written, short, shared.
🤖 AI-era tip: Pressure-test the wording, do not let a model invent the goals
Use an approved assistant to attack your draft goals: What would a skeptical executive ask? What is vague, unmeasurable, or actually an activity dressed as an outcome? Then you and the team rewrite and own them. Polished objectives that nobody believes are worse than plain language people will defend. Never let a tool invent the team’s aims and then treat typographic confidence as alignment.
Takeaways
- Write outcomes and publish non-goals; “value” is not a criterion until it is specific.
- Alignment first, then autonomy — or you have funded disorder.
- Cut early or finish; late cuts are how good ideas die and average work lingers.
- Speed without a decision or a vision is just faster waste.
- A weekly memo of three priorities is enough if it is true.
Evidence
- Locke, E. A., and Latham, G. P. (2002). Building a practically useful theory of goal setting and task motivation: A 35-year odyssey. American Psychologist, 57(9), 705–717. Specific, difficult goals outperform “do your best” goals (meta-analytic d = .42 to .80). Goals plus feedback outperform goals alone. On complex new tasks, a specific learning goal can be more useful than a premature performance-outcome goal.
- Brockman, G. (@gdb) (2026). the bottleneck is increasingly knowing what you want. Primary post.