If you spend time on X or LinkedIn, it can feel like every company is suddenly “cooking,” “moving at warp speed,” or “shipping 10× faster” with AI. The stories are loud, polished, and often impressive.
But for most people inside real organizations, work still feels… normal. Sometimes slow. Often messy. Human.
Here’s the truth: AI is accelerating some teams, but the idea that the entire industry is moving at superhuman speed is exaggerated. Most work still moves at human speed because most work still depends on hidden information that AI can’t access.
1. Yes, some teams really are moving faster
There are companies shipping faster than ever. They have:
- deep AI expertise
- internal agent platforms
- big token budgets
- strong internal tooling
- top‑down mandates to use AI everywhere
They’re automating boilerplate, refactors, tests, documentation, and migrations. They’re collapsing iteration cycles. They’re parallelizing work in ways that weren’t possible before.
These gains are real — but they’re also narrow. They apply to work where all the relevant information is visible to the model.
And that’s the catch.
2. Most work doesn’t live in the open
AI can only accelerate work when the knowledge it needs is:
- in the codebase
- in the docs
- in the spec
- in the training data
- in public sources
But most real work depends on hidden information — the stuff that never makes it into any dataset:
- hallway conversations
- private Slack threads
- political dynamics
- tacit knowledge
- undocumented decisions
- “why we don’t do it that way”
- “what legal said last time”
- “what the customer actually meant”
This is the information that actually drives decisions, tradeoffs, and sequencing.
And AI can’t accelerate what it can’t see.
Hidden information is the real speed limit.
3. Engineering already automated the easy stuff
Long before LLMs, teams had:
- CI/CD
- infra-as-code
- automated testing
- containerization
- linters and static analysis
The low‑hanging fruit was picked years ago.
LLMs add efficiency, but they’re not replacing decades of automation overnight.
The bottlenecks today are almost never mechanical.
They’re human.
4. Human-paced work sets the pace of the whole system
Even if an agent can generate 1,000 PRs, humans still have to:
- interpret ambiguous requirements
- negotiate priorities
- align stakeholders
- navigate compliance
- make judgment calls
- communicate decisions
- absorb change
You can’t ship a feature faster than legal can review it.
You can’t close a deal faster than a customer can evaluate it.
You can’t align a roadmap faster than stakeholders can agree.
The system moves at the speed of its hidden information — not the speed of its automation.
5. Why the hype feels louder than reality
Three reasons:
A. The frontier companies post the most
They share demos, metrics, and internal wins.
They have incentives to signal speed.
B. The median company is quiet
They’re experimenting, not evangelizing.
They’re cautious.
They’re dealing with real customers and real constraints.
C. You’re comparing your internal messiness to someone else’s highlight reel
And that’s never a fair comparison.
6. What’s actually changing — and what isn’t
AI is great at:
- reducing mechanical toil
- drafting and summarizing
- generating boilerplate
- refactoring
- writing tests
- compressing iteration cycles
AI is not great at:
- accessing hidden information
- resolving ambiguity
- navigating politics
- replacing judgment
- building trust
- making tradeoffs
- aligning stakeholders
The real story: AI accelerates tasks, not entire companies.
7. You’re probably not behind
Most companies are moving at a reasonable pace.
They’re adopting AI thoughtfully.
They’re balancing risk, cost, and value.
They’re integrating AI into existing workflows, not rebuilding everything from scratch.
If your work involves humans, ambiguity, or hidden information, you can’t move at warp speed — and that’s not a failure. It’s the nature of the work.
The companies that win won’t be the ones that panic their way into “10× velocity.”
They’ll be the ones that adopt AI deliberately, sustainably, and with clear judgment.
8. A more grounded path forward
Use AI where it helps:
- repetitive tasks
- drafting
- analysis
- code generation
- documentation
Don’t force AI where it doesn’t:
- strategy
- relationships
- judgment
- alignment
- anything dependent on hidden information
Clarity beats velocity.
Understanding beats hype.
Integration beats theatrics.
You’re not behind
AI is powerful, but work still happens at human speed — especially the parts that matter most.
You don’t need to cook. You just need to build well.