Career Strategy in the AI Era

TL;DR

• Career strategy built for stable markets plans your next move. Career strategy built for AI-era markets plans the move after, because the shape of the next role is changing faster than you can predict.
• The three career forces worth modelling in 2026: your skip-move horizon (the role two hops away), your shadow superpower (the skill that made you senior and is now slowing you down), and the modernity-over-pedigree hiring signal (recency of hands-on AI work beats brand tenure).
• The quarterly career review should be a deliberate practice, not an annual exercise. Markets are moving faster than annual cycles; your planning cadence should match.

---

Career advice written before 2023 assumes a stable market. Roles evolve, but the shape of a senior role five years out resembles the shape of the equivalent role today. Under those conditions, optimising for the next job is rational. The role after is a bigger version of the role before, and accumulating scope, title, and prestige compounds cleanly.

That model is badly calibrated for the current market. The shape of senior product roles in 2028 is actively being redesigned by AI tooling, operating-model changes, and the collapse of the translation layer between business intent and software execution. A career plan built on accumulating 2022-flavoured seniority will land you, in 2028, as an expert operator in a market that no longer values your expertise.

This chapter integrates three concepts from the blog (the skip move, the shadow superpower, modernity over pedigree) into a single planning practice for operators navigating AI-era career decisions.

The three planning horizons

A career always has three horizons active at once. Most career advice focuses on only one of them.

The immediate horizon (0–6 months). The day-to-day job: what you do, who you work with, what you ship, and how your week actually gets spent. Most career anxiety lives here. It's almost never where the strategic decisions get made.

The next horizon (6 months to 2 years). The next role. What you'd apply for, what you'd be offered, what you'd take. Traditional career planning is built around this horizon. In stable markets, that's correct. In fast-changing markets, it becomes a trap, because the next role is a stepping stone whose value is a function of what comes after it.

The skip horizon (3–5 years). The role two hops away. In 2026, this horizon is where your planning should sit, because the shape of roles two hops out is less predictable than the shape of roles one hop out. Counter-intuitively, that makes the skip horizon more stable as a planning anchor, because you're forced to describe a specific endpoint instead of assuming continuity.

The exercise that works: describe your skip role in enough detail that you could interview for it today. Not the title. The daily work. The tools in active use. The skill stack the hiring manager would test you on. The kind of company that would be hiring for it.

If you can't describe the skip role concretely, you don't have enough exposure to the frontier of your field. That's the first problem to fix, and it's usually fixable by spending two months in deep conversation with operators at AI-native companies, following public builders on Twitter/X and GitHub, and reading the technical content people at frontier companies publish about their operating practice.

Once you can describe the skip role, career decisions get easier. Every move is judged against: does this close the skill-stack delta between the current role and the skip role? The move that closes the most delta is usually the best move, regardless of what it does to title or compensation in the short term.

The shadow superpower blocker

Every senior operator has a shadow superpower: the skill that produced their last two or three promotions, which is now the skill most at risk of being commoditised by AI. The pattern is predictable and widely distributed.

• The VP who got where they are by running elegant quarterly reviews and managing executive narrative.
• The Director who compounded their career on cross-functional alignment and stakeholder management.
• The Principal PM whose reputation was built on PRD quality, precision of requirements, and managing drift between intent and execution.
• The design leader whose competitive edge was visual polish and production velocity.

None of these skills were fake. They produced real value in the operating model that rewarded them. The issue is that the operating model is being retired, and the skills are being repriced. The planning implication for careers is specific.

The shadow superpower will fill your calendar unless you act against it. The activities that earned your seniority still show up on your calendar every week: the reviews, the one-on-ones, the briefing decks, the escalations. Acquiring new skills requires time that's currently committed to outputs the organisation still rewards. Nobody frees up ten hours a week by cancelling the status review that their VP still expects. So the new skill never compounds, because it never gets the reps.

The fix is subtractive, not additive. The intuition senior operators reach for is to layer new skills on top of existing work. The research and the observed outcomes are clear: this almost never works. The move that does work is to deliberately stop doing some of the old-skill work, even at the cost of short-term validation, so that the new-skill work gets enough time to develop receipts.

The diagnostic question: What's the task your calendar defends? Look at the last four weeks. What activity shows up five or more times and could be produced by an AI tool in 20% of the time? That's your shadow superpower. The fact that you're still spending the hours is the tell that the old validation loop hasn't been broken.

The practical move is to drop one responsibility per quarter that falls into the shadow-superpower category, and use the recovered time to ship something with AI tools. Even small ships, done visibly, break the validation loop. The senior operators who make this transition successfully all do something like this. The ones who don't end up being hired for the old version of the role they hold, in a market where the old version no longer exists.

Modernity over pedigree

The third force worth modelling: the hiring signal has inverted. For twenty years, brand tenure at top-tier companies was a positive hiring signal. From roughly 2024 onward, that signal has weakened enough that treating it as a default positive actively selects against the skills most AI roles now require.

The practical implication for career planning: the resume signal built up over years is depreciating. Two years at a frontier AI-native company outperforms six years at a 2016–2022 big-tech incumbent across the dimensions that interview loops now test for (tool stack recency, shipped-feature fluency, eval literacy, unit-economics reasoning). Brand tenure is not the signal it used to be.

You need a recency portfolio, not just a tenure resume. The question you need to be able to answer, credibly and concretely, is "what have you shipped end-to-end in the last 90 days?" If the answer is "I set strategy" or "I reviewed my team's work" or "I prepared the board deck," you're selling the old signal. If the answer is a specific feature, built with specific tools, evaluated with a specific eval harness, deployed to specific users, you're selling the new one.

Building the recency portfolio is often the forcing function for the quarterly skip-move test. You cannot build the portfolio from inside a role where the day-to-day work doesn't produce evidence. The portfolio is downstream of the operating model you work in. If your current role doesn't produce ship receipts, the career decision is about changing the role, not about writing the resume.

Rewrite your resume around the last 18 months. Remove or compress older bullets. Lead with specifics (models, tools, evals, shipped features, unit economics). Treat 2018–2022 accomplishments as context, not headline material. This one change in resume framing alone can shift interview loop outcomes materially, because hiring managers who've absorbed the inversion pattern are screening for the recency profile first and scanning past brand names.

The four decisions that define AI-era careers

Most career strategy is run in reaction to opportunities as they appear. In a fast-changing market, reactive strategy loses to pre-framed strategy. Four decisions recur across most senior-operator trajectories right now, and pre-framing each makes the decision faster and sharper when it arrives.

When to stay versus leave

The default in stable markets is to stay. Switching costs are real; knowledge compounds; relationships deepen. In fast markets, the default should be reframed: stay when your current role closes the skill-stack delta toward your skip role; leave when it doesn't.

The test: in a typical week, what percentage of your time is spent building the skill stack of your 2029 role? If it's over 50%, stay and compound. If it's under 20%, leave. If it's in between, the decision is genuinely harder, and the other factors (compensation, relationships, ability to influence the role from inside) start to matter more.

When to take a title cut

Title cuts are culturally taboo and often rational. The test: is the new role inside the operating model of your skip role, with the people who are shipping the work you'd need to do to compete for the skip role three years from now? If yes, the title cut is a skill-stack investment priced lower than the equivalent external training, and it's one of the most reliable career moves available.

When to pivot industries

Industry pivots are usually underrated by generalist-market operators. The pivot from a traditional industry into AI product work is obvious. The less-obvious pivot (from generalist AI work into a specific vertical like proptech, legaltech, healthtech, or financial services) is where the scarcest talent is being built right now, because the combination of domain depth and AI fluency is rarer than either in isolation.

The test: does the target industry have problems the 2029 role will care about, that you can't pick up in your current context? If yes, the pivot usually pays back in 18–24 months. If you're pivoting away from a small domain into a bigger one (the intuitive direction), you're often moving away from the scarcity that would have made you valuable.

When to build versus work for someone

The founder/operator-in-residence/solo-builder pivot is a high-variance move. Expected value calculation matters; most founder moves fail on their stated goals. Skill-stack-wise, they almost always succeed, because shipping end-to-end is the highest-density way to rebuild skills that senior operators have lost to delegation.

The test: can you afford the downside case where the build doesn't work? If yes, and if your current role isn't producing the skill-stack receipts you need, the build pivot is one of the cleanest skip moves available. Some of the strongest senior AI hires in 2025–2026 are operators who failed at a founder attempt and are being hired at the top of the market on the strength of what they built during the attempt.

Running the quarterly review

Most operators run career reviews annually. Annual cadence was appropriate when the market moved on annual timescales. In 2026 the operating model is shifting on a 3–6 month cycle, so the career review needs to match.

A practical quarterly template:

Has the shape of your skip role changed since last quarter? Read public writing from frontier companies. Talk to two or three operators at AI-native companies. If the shape has moved, update your reference point.

Has your skill-stack delta closed or widened? Be honest. What specific new skills did you build this quarter that close the gap to the skip role? What skills did the market add to the skip role's requirements? Net direction matters.

Has your shadow superpower calendar footprint grown or shrunk? If the weekly hours you spend on old-skill work are higher than they were three months ago, you're regressing. The calendar doesn't lie.

Has your recency portfolio grown? What did you ship end-to-end this quarter, with AI tools, that you could put on a resume? If nothing, the role isn't producing career artefacts, which is its own signal.

Is there a next move that closes a meaningful chunk of the delta? Not necessarily available today. Map the move you'd make if the opportunity appeared. The mapping exercise is the point; opportunity shows up faster when you've already decided what to look for.

Running this honestly every 90 days is meaningfully different from running it every year. The quarter is the right unit for a market that moves on quarterly cycles. Operators who run it well end up making fewer big career decisions per decade, because the small decisions compound correctly. Operators who skip it end up making reactive big decisions at the end of long periods where the direction was wrong and nobody audited.

The common failure modes

Three failure modes recur often enough to be worth naming explicitly.

Credential compounding against a shrinking market. You spend 2024–2028 accumulating scope and title in an operating model that the market is actively repricing down. Your compensation and scope look good on paper; your skill stack is depreciating rapidly. The failure mode only becomes visible in 2028 when you try to move and discover that the market has moved past the role you're now senior in.

Reactive career moves driven by short-horizon pressure. A bad week at your current job, an unexpected internal reorg, or a friend's offer triggers a career move that fits the immediate horizon but not the skip horizon. These moves are usually net-neutral at best. The fix is discipline: the quarterly review generates your candidate decisions, not the emotional weather.

Conflating brand with trajectory. Taking a role at a high-prestige company because the brand will be good on the resume, in a market where the resume signal has inverted. The brand helps with one interview loop and hurts with the next one, because the hiring managers at frontier companies are actively discounting brand and looking for recency. Optimising for brand is usually optimising for a 2020 market.

None of these are hard to recognise once you've seen them. They're difficult to avoid without a deliberate planning practice, because the incentives around senior operators (compensation curves, peer signalling, employer loyalty norms, family stability) all push toward the failure modes by default.

Anti-pattern: the prestige-driven lateral

Director of Product at a mid-size company, eight years of experience, strong brand tenure. Gets an offer for a VP role at a larger incumbent. Title bump. Compensation bump. Bigger team. Under next-move framing, the decision is obvious.

Under skip-move framing: the new company operates in an operating model that already looks dated. The team is large because the company hasn't rebuilt around AI productivity. The VP scope will involve running the kind of roadmap reviews, steering committees, and quarterly alignment sessions that were load-bearing in 2019 and are becoming noise. Eighteen months in, the director realises that the title bump is compounding on a declining skill stack, and they can't walk back to a frontier IC role without looking like they failed upward.

The rational move was the smaller-scope IC role at an AI-native company that the director passed on six months before the VP offer came in. That smaller-scope role would have produced the skill-stack receipts that made the 2028 senior role accessible. The prestige role produced a 2026 bullet point that won't matter by 2028.

This failure mode is the most common one in the senior-operator market right now. It looks like success on the way in. It looks like a career-derailing mistake eighteen to twenty-four months later, when the skill-stack reality catches up to the title. The frame that prevents it is the skip-move test, applied honestly, against every offer.

A note on energy and sustainability

The reframes above (skip move, shadow superpower, modernity) imply a specific kind of ongoing effort: deliberate skill-stack rebuilding alongside your day job, often in territory that feels uncomfortable or regressive. That effort is real. It shouldn't be nights-and-weekends coded; the mode that actually works is protecting a thin slice of time each week for non-delivery skill-building, inside the working week.

Career strategies that require burning the candle at both ends are usually wrong, and definitely wrong on longer horizons. The operators who are quietly outperforming the market in 2026 are not the ones working sixty-hour weeks. They're the ones who have restructured their forty-hour weeks to include 3–5 hours of deliberate skill-stack work. Over a year, that compounds into a different person. Over three years, it compounds into a different career.

The shape of a well-run AI-era career isn't more effort than the previous model. It's the same effort, applied against a different planning horizon, with a different validation loop. The rest follows from that.