15+ years building systems. Now AI agents. Gaming → dental → robotics. The Pragmatic Builder. No hype, just results.
Joined March 2026
- Tweets39
- Following5
- Followers0
- Likes1
A clinic operator approved every scheduling exception. Personally. Daily.
We documented the decisions. Built agents to handle them.
Founder-dependent approvals dropped 70%.
That's the skill: knowing which decisions need a human and which need a system.
"We don't trust AI enough to automate." You don't trust your calculator either. You trust the process: input → compute → verify output. Build that process for agents. Stop reviewing everything. Start building systems worth trusting.
Human in the loop = your best people burning hours reviewing AI output instead of making decisions only humans can make. Build systems you trust. Free the humans for human work.
The tools that pass? The boring ones.
Chat scanning that surfaces what I'd miss. CI/CD monitoring that catches failures before they block the team. Document processing that killed hours of manual entry.
None went viral. All of them work.
Petscom had $82M in funding. Amazon had a logistics problem.
Both were "internet companies." Same test. Different decade.
Full breakdown: ivarberg.com/blog/dot-com-te…
Agents multiply whatever you are. Deep expertise → 10x output. Shallow understanding → 10x confident garbage. The bottleneck was never execution. It's judgment.
Got quoted 3 months for a system rebuild. Did it in 3 days with agents. The secret wasn't AI. It was
knowing what to ask for. Agents multiply experience. Without it they just multiply mistakes.
6 months of daily AI agent work. No going back.
Not because I love AI. Because the gap between "with agents" and "without" is now so large it feels like typewriters vs computers.
I mapped the journey. Here's the framework:
The part that matters most:
AI amplifies capability. Good and bad.
Bad actors are already using it at full speed. Too many good people are watching from the sidelines.
The worldly pressures won't let us go back. Good actors must join in to shape what comes next.
This is Part 1 of The AI Practitioner's Playbook.
Next: How I actually direct agents, why chat interfaces are a ceiling, and why messy thinking is the #1 reason agents fail.
Full post: ivarberg.com/blog/no-going-b…
Every transformative technology had a hype crash. Electricity. Radio. Internet. The technology survived. The speculation didn't. AI is no different. Stop asking "is it a bubble?" Start asking "what survives the pop?"
Calculator replaced arithmetic. Spreadsheet replaced the ledger. Agent replaces cognitive manual labor. Same pattern. Same resistance. Every time we're embarrassed we fought it.