Walk into the headquarters of a large American company in 1995 and the IT floor is the floor with the rooms. Servers behind glass. A nameplated team that owns the change-management calendar. Code is written there. Code is shipped from there. The rest of the building consumes what that floor produces.
Walk in now. The IT floor still exists. It is still important. It is no longer where most of the company’s software is written.
The center of gravity of building has moved off the engineering floor. Engineering knows. The shape of engineering is what is changing.
What we see, inside the Fortune 500
Six of the ten enterprise customers we’ve worked with over the last year now have an internal-tools function inside marketing. Five have one inside finance. Three have one inside legal. None of these functions existed two years ago. Each of them is staffed primarily by people whose job title does not contain the word engineer. Each of them ships software, into production, into the hands of internal users and sometimes external ones.
At one of those customers, the marketing internal-tools function ships more software per quarter, by line and by deployed surface, than the marketing-technology team inside engineering does. The engineering team is not threatened by this. They architected it. They provided the rails, SSO, observability, the deployment pipeline, the secret store, the on-call rotation that catches what breaks. The engineering team does not write the software. They make sure the software written by everyone else is allowed to ship.
The new shape of engineering
Engineering, inside an enterprise that builds, is increasingly a platform. The platform team builds the rails. The platform team owns the press. When the marketing manager who built a tool with Lovable is about to send the URL to a customer, the platform team is the thing they reach for, through Relay, in our customers’ case, to confirm what they built will hold.
Three things follow from that. First, the software engineer’s job is denser, not lighter. They review more diffs, from more authors, in more domains, per week than they used to. Second, the company’s engineering footprint expands rather than contracts. It expands in the direction of governance, not generation: review, observability, security, audit. Third, the failure mode of the new company is not bottleneck-at-engineering. It is bottleneck-at-review. The marketing manager has shipped the tool; engineering has not yet seen the tool; the tool is in production. This failure mode is what the press is for.
What the org chart will look like in 2030
We don’t make precise predictions, but the rough shape is legible. A company of 10,000 employees in 2030 will have, by our estimate, between three and five times the number of distinct software systems in production it has today. Most of those systems will have been written by people not in engineering. The engineering team will be larger than today, not smaller, but it will be doing a different job. Engineering managers will manage review queues more than they manage authorship. Staff engineers will spend most of their day reading other people’s code and saying this can ship or this cannot. Platform teams will outnumber feature teams.
And the press, the moment a non-engineer asks a software engineer a single, time-boxed question about a single, time-boxed thing they have built, will be the smallest unit of work in the company. Smaller than a story. Smaller than a code review. Smaller than the meeting it would have replaced.
What this means for buyers right now
Three things, in the order we’d weight them. One: invest in the rails. The rails are the difference between an enterprise that lets non-engineers build and an enterprise that has non-engineers shipping unsanctioned software through the back door. Two: invest in review capacity. Either inside, by repurposing software engineers from feature work to platform work, or outside, through services like ours. Three: write down what you are willing to let ship without an engineer in the room, and what you are not. The line is the policy. The press is the enforcement.
The founders, May 2026.