The Senior Jump, and the One After It

The Senior Jump, and the One After It describes two thresholds that are usually explained as promotions and are actually changes in what your output is. Below the first, your output is code you wrote. Above it, your output is decisions other people make, and the code is a side effect. Everything that feels strange about senior work follows from that substitution.

Key takeaways

• A threshold is a change in the unit of your output, not an improvement in the same unit. Doing junior work excellently for longer does not cross one.
• The 75 published progression frameworks collected on progression.fyi, from companies including Dropbox, GitLab, Etsy, Spotify, Monzo and CircleCI, all describe scope widening from task to system to organisation. Their individual publication years run from 2007 to 2024, so all of them predate the change in what the first rung contains.
• The senior jump is where you stop being the person who types. The measure is whether the team makes better decisions when you are not in the room, which is uncomfortable precisely because it is not about your throughput.
• Influence without a title is the same mechanism arriving earlier, not a separate skill. Both resolve to your reasoning being available to people who are not obliged to take it.
• Being on record is the price. Influence requires stating a position in writing, dated, before the outcome is known, which is why the decision records from Chapter 7 are the load-bearing artifact of this chapter too.

Read this after Chapters 6, 7 and 8, because specifications, reviews and converted ambiguity are the three artifacts a threshold is judged on, and before Chapter 11, which is what happens when the same reasoning is visible outside your employer. Chapter 12 is the alternative to crossing thresholds inside a company at all.

An engineer is told she is not yet senior. She closes more tickets than anyone on the team, her code is the cleanest in the repository, and she has never missed a date.

The feedback she is given is vague, and it is also correct. Nothing she produces changes what anybody else does. When she is on holiday, the team makes the same decisions it would have made anyway, only slightly slower.

The following quarter she writes two specifications, one decision record, and a narrowing of a problem nobody could define. Nothing about her code changes. Her scope does. The details here are composited from ordinary organisational shapes rather than one company, and the shape is exact.

A threshold is a change in what your output is

The reason progression feels arbitrary is that people measure it in the wrong unit.

Below the first threshold, your output is completed work: tickets, changes, features, correctly done and on time. Improving inside that unit makes you a better engineer and crosses nothing. Which is why the fastest person on a team can stall for years while feeling that they are obviously ready.

Above it, your output is decisions. Some are yours, and more of them are other people's decisions that were better because of something you wrote, said or established. That unit is harder to see, which is why the evidence for it has to be manufactured deliberately in the form of artifacts.

The consequence is uncomfortable and worth stating flatly. A threshold is crossed by changing what you produce. Not by producing more.

Seventy-five published ladders agree, and all of them predate the shift

There is a public corpus of career frameworks. It is more useful than any single company's document.

The index at progression.fyi collects 75 published career and progression frameworks in one place, including those of Dropbox, GitLab, Medium, Etsy, Spotify, Monzo, CircleCI, Rent the Runway and Basecamp. They are real internal artifacts, freely readable. The striking thing about reading several is how much they agree. Scope widens from task to system to organisation, influence replaces authorship, and the language shifts from doing to enabling.

Their individual publication years run from 2007 to 2024, and the index carries no site-level update date. That is the finding, in my reading of it. Every published ladder describes the transition this chapter is about. Every one of them was written before the first rung stopped consisting mainly of producing code.

So the ladders are not wrong. They are correct about the shape and out of date about the entry conditions. The behaviours they describe at the senior rung are the behaviours that now start earlier, because the work that used to fill a junior's week is the work that got cheapest first.

Task, system, organisation

Three scopes, described as behaviour rather than as title, because titles differ between companies and behaviour does not.

The third column is the one people resist. Each threshold requires giving something up that was the source of your standing. The loss is real, not rhetorical.

The senior jump is when you stop being the one who types

Here is the judgement version, stated so you can disagree with it.

In my experience the senior threshold is crossed at the point where the team's decisions improve when you are not present. Not when you can be relied on to finish hard things, which is the previous rung, and not when you are formally consulted, which is a courtesy. When your absence for a week does not degrade the quality of what gets decided, something you built into the team is doing the work. That is the test.

The behaviours that produce that are specific. You write down the reasoning behind decisions so it is available without you. You review in a way that transfers a rule rather than fixing an instance. You specify precisely enough that the work does not need your interpretation. All three are the Chapter 6 and Chapter 7 artifacts, used as teaching rather than as process.

I have been wrong about this in one direction, repeatedly. I used to treat being needed as evidence of value. It is the opposite. Sustained indispensability at a detail level is the clearest sign that a threshold has not been crossed.

The one after it is decisions by people you do not review

The next threshold has the same structure and a wider radius, which is why it does not need its own theory.

At the senior level your influence travels through people whose work you see. Above that, it travels through people whose work you do not see, in teams you are not part of, via standards, defaults, written positions and the occasional intervention. The mechanism is identical. The loop is much longer.

Two things change practically. The first is that you stop being able to verify most outcomes yourself. So you invest in things that fail visibly: defaults that are safe when ignored, checks that run without anyone remembering them, documents that are found at the moment of need rather than at induction.

The second is that being wrong costs more and shows up later. Which is why the writing discipline matters more at this level rather than less. A position with a date on it is worth more than a position stated confidently in a meeting.

Influence without a title is the same mechanism, earlier

People treat influence without authority as a separate skill and it is not. It is the same substitution, available before anybody has agreed to it.

The mechanism is that your reasoning becomes available to people who are under no obligation to take it. What makes it accepted is not seniority. It is that the reasoning is checkable: it names the trade-off, states what would change the answer, and turns out to be right often enough to be worth reading. A specification that saved a team a week. A decision record that predicted the failure that happened. A review comment that transferred a rule. All three work identically whether or not you have the title.

The one asymmetry is patience. Without a title you cannot compel a decision, so everything runs through being persuasive in writing and being demonstrably right over time. That is slower and it produces the evidence that the title, when it arrives, is justified by something.

What the market is paying the threshold for

There is labour-market evidence pointing the same way, and it needs its limits attached.

Hosseini Maasoum and Lichtinger's working paper, posted 31 August 2025, uses US résumé and job-posting data covering nearly 62 million workers across 285,000 firms from 2015 to 2025, identifying AI adoption through postings for dedicated AI-integrator roles. Their finding is directional: generative AI appears to act as seniority-biased technological change, disproportionately affecting junior relative to senior workers. I read the abstract via a search extract rather than the paper itself, so I am reporting the design and the direction. No percentage from me.

The complementary finding is about what the tool does at the top of the skill distribution. Brynjolfsson, Li and Raymond, in the Quarterly Journal of Economics in May 2025, found the most experienced and highest-skilled workers saw small speed gains and small declines in quality, while less experienced workers improved on both.

Put those together and the threshold argument gets an economic reason rather than a cultural one. If assistance adds least where skill is highest, then above a certain level your own throughput is no longer the lever. What remains is deciding what should be produced and raising the quality of decisions around you.

The objection: my company does not have a ladder

Common, and it changes the tactics rather than the argument.

Plenty of organisations have no written framework, or have one that is decorative and applied inconsistently. In that situation the thresholds still exist, because they are about what work gets given to whom, and that is decided regardless of documentation. What you lose is the map, so you have to build one from the frameworks on progression.fyi and from watching who actually gets the undefined work.

There is a second version of this objection that is harder. Some companies will take threshold behaviour and never change the role. That is Chapter 9's diagnostic. The answer is the same: artifacts are portable, and internal goodwill is not.

My read is that the absence of a ladder is neutral information on its own and becomes negative in combination with one thing: no written record of why past decisions were made. A company that documents decisions can recognise judgement, whether or not it has a rubric. A company that documents nothing has no mechanism for noticing.

Chapter summary

A career threshold is a change in the unit of your output rather than an improvement in the same unit, so doing task-level work excellently for longer never crosses one. Below the first threshold your output is completed work. Above it your output is decisions, including other people's decisions that were better because of something you wrote, which is why evidence has to be manufactured deliberately as artifacts. The 75 published frameworks collected at progression.fyi, from companies including Dropbox, GitLab and Monzo, agree that scope widens from task to system to organisation and that influence replaces authorship. Their publication years run from 2007 to 2024, so all of them describe the transition while predating the change in what the first rung contains. Each scope requires giving something up: the system scope costs you being the fastest producer, and the organisation scope costs you being in every decision. The senior threshold, in my experience, is crossed when the team's decisions improve in your absence, produced by writing reasoning down, reviewing to transfer rules rather than fix instances, and specifying precisely enough to remove interpretation. The next threshold has the same mechanism with a longer feedback loop, which favours defaults that are safe when ignored and positions written with dates on them. Influence without a title is that same substitution arriving earlier, accepted because the reasoning is checkable. And the market points the same way, with Hosseini Maasoum and Lichtinger reporting generative AI as seniority-biased technological change across nearly 62 million workers.

Chapter 11 takes the same reasoning outside the building. Community, Speaking, and Public Work is about why visible work changes which opportunities reach you at all. It is also the one chapter where I am the evidence, which requires care about the difference between what I can show and what I merely believe.