marketing isn't dead. it's becoming engineering.

everyone's shouting that marketing is dead. it's not. it's evolving. like every other white-collar role.

the role is changing, not disappearing

look at who companies are actually trying to hire right now. it's not "marketers." the fastest-growing title in the stack is the gtm engineer: postings up 205% year over year, median salary around $127k, over a third of the listings asking for sql and python. for a marketing job. and the top payers aren't agencies. they're vercel and openai, at $250k+.

clay, the tool at the center of this shift, turned "treat gtm as a systems problem" into a company last valued around $3b, roughly $5b if you believe the latest tender. an entire job category formed around one product. that tells you where the leverage went.

the job didn't vanish. it moved closer to engineering.

why it moved

the mechanism is simple. distribution became software a decade ago: ads are auctions, feeds are algorithms, outreach is an api. what changed now is that ai pushed the cost of producing the creative, the copy, and the campaign toward zero.

when production is free, making the asset stops being the job. the job becomes building the system that makes, tests, ships, and kills a hundred assets a week. that's not a brief. that's a pipeline. and pipelines are engineering.

the god-tool fantasy

we've all seen the pitch: twenty different ai marketing agents that promise to do everything for you while you sip a piña colada. they mostly don't deliver. because that isn't what teams need yet.

the receipts are piling up. 11x, the a16z-backed "ai sdr" that sold fully autonomous digital workers, got caught claiming customers it didn't have, with churn reported at 70-80%. gartner expects over 40% of agentic ai projects to be canceled by the end of 2027. and the much-quoted mit finding (contested, but echoed directionally everywhere) put it bluntly: roughly 95% of enterprise genai pilots showed no p&l impact. the same report found that tools bought and layered over existing workflows succeeded about twice as often as the ambitious internal builds.

what the world needs right now isn't another ai marketing god-tool. it's a co-pilot. something that integrates seamlessly into the current ops of a real company and makes the people there 10x more proficient.

the strongest objection

the real god-tool won't be a startup. it'll be meta. zuckerberg has already said the endgame out loud: tell the machine your business objective, connect your bank account, done. that product was promised for end of 2026. as of today it's in limited testing, and the media buyers who've touched it call the output mixed.

and look at what the money is doing in the meantime. agencies are taking the hit: wpp is down roughly 60% and cutting deep. but cmos raised the share of budget going to in-house people this year, from about 22% to 24.5%. the money leaving agencies isn't leaving marketing. it's being reallocated to people inside the company who can run the machines.

when meta finishes the loop, it'll own commodity performance ads inside meta's walls, optimized for meta's inventory and meta's margin. everything else, your owned channels, your positioning, the weird cross-channel plays, your own data, still needs someone to engineer it.

people want harnesses, not replacements

it's the little steps. maybe in another two years, the way no-code eventually arrived, we get the genuinely hands-off tools. the signal to watch is whether meta actually ships the bank-account version and performance advertisers accept it. until then, people want harnesses, webs, and co-pilots they can lay over their existing systems, operations, and strategies, and improve multi-fold.

my guess on the title itself: "gtm engineer" eventually goes the way of "webmaster." hybrid titles show up when a new stack arrives and dissolve once the stack matures. the title marks the transition. the direction underneath, marketing run as engineering, is the end state.

that's a real gap. it's also the thing i'm quietly building on the side. beta-piloting with a couple of friends' companies, still fine-tuning the product and the positioning.

what fun times.