Are We Ready?¶
I was preparing a presentation for Item Consulting — their internal conference, that Friday — and I kept coming back to one question.
Are we ready for what's happening?
Over the previous month I'd been sharing the details here: building a 200,000-line production Java library in eleven days. Not theory, not demos. Production code. 7,461 tests. Manufacturing compliance. Eight binary format parsers for PCB design files.
The first reaction was always: "That's impossible."
The second: "That only works for you."
But I was already seeing the pattern repeat. Different domains. Different codebases. Different developers. Same systematic methodology. Early results: 10-30x productivity in weeks, not months.
The timeline I was seeing¶
- Today: ~5% of developers using systematic methodology (vs just AI tools)
- 12 months: ~30%
- 24 months: Table stakes
- 36 months: Taught in bootcamps
The gap between using AI tools (1.5–5x) and systematic methodology (10–30x) is the real opportunity. And that window is 12–18 months wide before it closes.
If development fundamentals changed — estimation models, team sizing, velocity assumptions — and we're still using approaches designed for a world that no longer exists... are we ready?
What the comments unlocked¶
The thread that came back was more interesting than the post.
Daniel Bentes put his finger on something precise:
"Definitely not ready. Most people don't have the mindset necessary to work this way. The other issue is identity. AI taking over skills that humans self-identify with as part of their identity — something they studied and crafted over years — will be the most difficult thing to overcome."
Two distinct barriers. Mindset is the first one people name — learning to work differently — but Daniel went one level deeper: identity. The skills you've spent years building aren't just capabilities. They're part of how you understand yourself as a professional. When AI takes them over, the threat isn't just to your job description. It's to your sense of authorship.
Erland Glad Solstrand replied from the same angle. He'd independently published an article asking "but how does it make me feel?" — about the specific loss of ownership and joy that happens when AI writes code and the human reviews. He'd been sitting with the same question.
The direction of the relationship¶
That exchange sharpened what I was already thinking about SDD — Skill-Driven Development.
The identity problem isn't inherent to AI-assisted development. It's specific to one configuration of the relationship: AI leads, human reviews. In that mode, ownership disappears because you didn't drive the work — you approved it. The craft doesn't operate through you. It operates on you.
Flip the direction: your skill sets the course, AI executes. You're still the author. The architecture decisions are yours. The judgment calls are yours. The output runs at a velocity you couldn't achieve alone, but the direction is entirely yours.
The craft doesn't disappear. It operates at a different altitude.
That's the distinction we'd been calling Skill-Driven Development. It doesn't solve the mindset shift Daniel describes — that work is still required. But it at least keeps identity intact as the entry point, not the casualty.
Still a genuine question¶
Standing in front of thirty developers at Item Consulting that Friday, the question was still real.
Are we ready?
Not as a rhetorical challenge — as an honest assessment. The timeline I sketched is probably directionally right, but the pace of adoption will be uneven. The people who figure out how to keep themselves as the authors — the skill-drivers — will navigate the transition cleanly. The ones who cede authorship to the tools will find the identity problem Daniel named, and it won't be comfortable.
The window is 12–18 months.
The question is what you do while it's open.
Originally published on LinkedIn, February 5, 2026.
This post is part of the AI-Augmented Development series.