What happens when the coding becomes the least interesting part of the work

Obie Fernandez · article · 2025-12-12 source ↗ Reader 🖥️

Agent Security Coordination Cost Domain Expertise as AI Force-Multiplier Pair & Mob Programming Taste & Judgment

Senior engineering value has always lived in mostly-tacit, pre-code judgment (pattern-recognition, abstraction level, blast radius, reversibility) rather than typing, which is why coding agents cannot yet substitute for it — but once senior engineers and their employers realize this, the organizational result will be a drastic contraction toward small teams of highly accountable, AI-empowered seniors who absorb roles that used to belong to juniors, PMs, business analysts, and designers.

mixed · mixed — Introspective/expository on what senior engineering judgment actually consists of, pivoting into an argumentative, practical prediction about team and organizational headcount.

Structure

Part	Role
The 'silent checklist' that runs before typing — questions about shape, direction, blast radius, reversibility, none of which mention code	evidence/introspection
'Why the most important stuff stays invisible' — senior thinking is tacit, rarely documented, and barely represented in frontier model training data	premise
'Agents force you to say the unspoken parts out loud' — pairing with an agent reproduces the beneficial effect of pairing with a good human partner	evidence
'The senior toolkit is bigger than people think' — blast radius, sequencing, reversibility, social cost, false-confidence detection as additional tacit tools	elaboration
'Why this subject flies over some colleagues' heads' — reaction to agents depends on which part of the job you valued	explanation of disagreement
The organizational prediction: 20-person teams shrink to 4-5 highly accountable AI-empowered seniors, absorbing PM/BA/designer roles	conclusion / extrapolation

Key terms

Senior thinking
The largely tacit, fast, quiet judgment a senior engineer brings to a task — about shape, direction, blast radius, reversibility, social cost — which is almost never articulated, documented, or standardized by teams.
Blast radius Agent Security
The scope of consequence if a change turns out to be wrong — how far it would spread and how hard it would be to contain or undo.
Pairing with an agent
Working with a coding agent (e.g., Claude Code) in the same disciplined, conversational mode as pairing with a human — explaining not just what to build but how you're thinking about the problem, which forces tacit reasoning into the open.

Propositions

explicitThe real initial effort in programming happens before the first line of code is written — pattern-recognition against years of experience finds the solution — making the actual typing/coding mostly follow-through, which is why coding agents are a game-changer for someone whose value was already in that pre-code judgment.conf: high
supportsWe don't write code anymore · p1 — Both independently describe the center of the engineering job moving from typing code to the judgment/framing work that precedes or surrounds it.
explicitSenior engineering judgment runs on a largely tacit 'silent checklist' about shape, direction, and consequences (what kind of change is this, does it belong near the surface or deep in the system, what's the blast radius, is it reversible) rather than about code syntax — and teams rarely document or standardize this because it's hard to pin down and because talking about patterns/frameworks is easier than talking about instinct and timing.conf: high
explicitThis tacit 'senior thinking' is barely represented in frontier models' base training data — published books and blog posts 'only scrape the surface,' and essentially no corpus of senior engineers' spoken-aloud pairing sessions exists — which is why agents can recognize the elements of senior thinking without having any intrinsic determination to apply them.conf: high
supportsAgents don't know what good looks like. And that's exactly the problem. · p1 — Both argue current agentic AI operates below the level of professional judgment — Fernandez's training-data-absence account gives a mechanism for why.
explicitPairing with a coding agent forces you to say unspoken parts of your reasoning out loud (abstraction level, risk tolerance, whether something is a deliberate temporary kluge), and that act of explanation reproduces the same benefit that pairing with a good human partner has always had: it slows you down enough to notice when your own instincts are off.conf: high
supportsOn Pair Programming · p4 — Both describe pairing (human or agent) as live, continuous reasoning-exposure that catches errors and sharpens design as the work happens, not after.
explicitWhether coding agents feel liberating or beside the point to a given engineer depends on which part of programming they actually valued: those who enjoy the physical act of typing code experience agents as irrelevant, while those whose value is judgment, tradeoffs, and intent experience agents as clearing space for exactly what they care about.conf: high
explicitOnce senior engineers — and especially their employers — collectively realize that typing code was never the scarce part of the job, the organizational result will be drastic headcount contraction: a 20-person team could become 4-5 highly accountable, AI-empowered senior engineers who also absorb the project-manager, business-analyst, and (depending on domain) designer functions, because even the most productive juniors and mid-level engineers start to look like net negatives.conf: high
supportsDesigning the AI-native engineering organization · p3 — Independent convergence on roughly the same AI-era team size (Fernandez: 4-5 seniors; that doc: 3-4 people) as the bottleneck shifts away from raw building capacity.
implicitWhat people care about most in their job is also, in the author's framing, exactly the part most exposed to automation for those who care about typing rather than judgment — so the apparent safety of 'that's not what I care about' is actually a marker of vulnerability, not insulation.conf: medium
implicitcontestedThe argument assumes the organization's current supply of senior engineers — and the irreplaceable tacit judgment they individually hold — is a stable, already-adequate resource that can simply be redeployed more efficiently; it does not consider where the next generation of seniors will come from once the junior/mid roles that historically served as the training ground for that same tacit judgment are the first ones eliminated.conf: medium
contradictsWhy Mentoring Matters More in the AI Era · p2 — Fernandez's prediction implicitly treats the senior-judgment supply as stable and redeployable; the HBR piece argues directly that eliminating the junior on-ramp that builds judgment destroys the leadership pipeline that produces future seniors — the opposite assumption.

Arguments

Senior engineering judgment is not currently replaceable by coding agents, and the organizational consequence of that fact is a drastic headcount contraction toward small teams of highly accountable seniors, not business as usual.

Explicit premises:

Most of what a senior engineer does is pre-code pattern recognition framed around shape, blast radius, and reversibility rather than syntax
This tacit knowledge is barely represented in frontier model training data because it is almost never written down or recorded
Coding agents execute the mechanical/typing part of the job extremely well, which was already the less valuable part for senior practitioners
Once senior engineers and their employers recognize this, the economically rational move is to concentrate work in a few highly accountable, AI-empowered seniors and remove the roles whose value depended on the old division of labor

Implicit premises:

Senior judgment can be exercised just as well by a 4-5-person senior team using AI as it could by the original 20-person team's division of labor, with no quality or coverage loss from the roles removed
The supply of senior engineers carrying this tacit judgment is effectively adequate going forward and does not need active replenishment from a junior/mid pipeline

Inferences:

If true at scale, this predicts a structural hollowing-out of the junior/mid tier well before any agent itself acquires 'senior thinking' — which would eventually choke off the supply of future seniors, since that judgment is explicitly described as learned over 'years and decades' doing exactly the work being eliminated. The author states the mechanism that implies this but never draws the consequence.

Narratives

When pairing with someone less experienced, the author slows down enough to explain parts of his 'senior thinking' but deliberately not all of it, because fully articulating it 'would bog us down too much.'
→ Shows concretely that senior tacit knowledge stays underdocumented even in the most favorable transmission context available — live pairing with a junior — setting up why pairing with an agent (which has no relationship to protect and no patience to spare) forces out loud what pairing with a junior colleague does not.

Questions addressed

✅ Can coding agents replace the judgment-heavy part of senior engineering work?
Not currently, and not soon — that judgment is tacit, poorly represented in training data, and agents have no intrinsic determination to apply it even when they recognize its elements; the author doesn't expect this to change without something like real AGI.
✅ Why do some programmers not see the point of coding agents, or actively dislike them?
Because the part of the job they personally value is the physical act of writing code — exactly the part agents are best at replacing — so agents feel irrelevant or threatening rather than liberating, depending on what the engineer valued in the first place.
⭕ What happens to the junior/mid engineers, project managers, business analysts, and designers whose roles depended on the old division of labor once seniors can direct AI directly?

Gaps

incompleteThe piece predicts collapsing a 20-person team into 4-5 AI-empowered seniors who absorb the PM/BA/designer roles, but its own argument undercuts this: senior judgment is explicitly described as learned 'over the course of years and decades' by doing exactly the junior/mid work the piece says will disappear first ('even the most productive juniors and mids are going to start feeling like net negatives'). If that pipeline closes, the piece never explains where the next generation of seniors — the only people it trusts to hold this irreplaceable tacit knowledge — will come from.
Read next: extract-hbr-mentoring-matters-ai-era, which names this exact tension (the on-ramp for professional judgment disappearing alongside the flattening of the senior population expected to do the mentoring) as a structural risk to the leadership pipeline.
incompleteThe claim that cutting a team to 4-5 seniors removes the need for a project manager, business analysts, or designers treats those roles as pure overhead that disappears once seniors can ship directly, without engaging what those roles actually absorbed besides shippable output — stakeholder alignment, prioritization trade-offs, user research, cross-team coordination. That work doesn't vanish just because seniors can type code themselves; it has to land somewhere.
Read next: extract-kill-your-org-chart-for-ai (a related claim that at least names 'the development of better judgment' as management's remaining purpose) and extract-ai-native-engineering-org (bottleneck shifts to alignment/decision-making, not away from coordination entirely).
uninformedThe claim that typing code 'feels like a waste of time' when billed at an expensive hourly rate treats the economics purely from the senior consultant's individual vantage; it doesn't address the same calculus for salaried engineers inside a company absorbing the AI spend, or whether clients will keep paying the same hourly rate once they also realize how much of the work is agent-executed.