The Subsidy Cliff
Current AI pricing is a subsidy, not a market rate. Organizations building on the assumption of cheap AI will hit the same cliff ZIRP beneficiaries hit in 2023.
All our posts on AI, engineering, and software craft
Topics we write about:
Current AI pricing is a subsidy, not a market rate. Organizations building on the assumption of cheap AI will hit the same cliff ZIRP beneficiaries hit in 2023.
The reassurance industrial complex tells everyone they will be fine. The reality is a structural divide forming along lines most people do not see, between those who engage intentionally and those who wait.
Every previous revolution gave institutions time to adapt. This generational transformation does not.
A raw take on what happens when you stop waiting for permission and start building what's missing — why real progress creates friction, how process becomes protection, and what the "it's in my scope" reflex really reveals about organisational priorities.
Not because it codes better. Because you refused to evolve. Here's what adaptation actually looks like.
Microservices before users. Scale before market fit. Complexity rewarded, simplicity dismissed. Engineering forgot its purpose: the product. AI is the reckoning.
Standups, sprint planning, retros—all designed for human limitations AI is now bypassing. I stopped running agile ceremonies. We ship faster. Here's the truth.
AI models are non-deterministic. Prompts are suggestions, not guarantees. CLI tools with strict heuristics give you what prompting can't: actual enforcement.
Engineers treat context windows like they're infinite, then blame AI when quality drops. Managing context is the skill nobody talks about—and everyone needs.
Vibe coding is fun. It's not serious engineering. Planning, architecture, and quality gates matter more with AI, not less. Here's why the boring work still wins.
The shift from IDE to terminal broke my attachment to code—and made me a better engineer. Why the real skill in AI development is letting go, not holding on.
For decades, seniority meant capability. AI changed that. A hypothesis on what actually differentiates engineers now—and why tenure alone is no longer enough.
Why we're here, what we believe, and why this probably isn't for everyone.