A web-based dashboard for tracking Claude Code token usage and costs in real time. Runs as a lightweight Node.js server that reads Claude Code's session files and serves a React dashboard accessible from any browser on your network.
If a person writes 100 lines of code there’s a (valid) assumption that someone thinks this 100 lines of code is worth writing. With AI it takes no effort to write 10,000 loc. Asking someone else to figure out if that code is worth merging just offloads the effort to someone else who didn’t ask for it.
It makes sense, because humans and AI write code different. Because humans aren't AI. The types of mistakes AI writes into code would never be done by a human.
Fair. Though I'd argue there's a difference between bitter and just telling the truth with receipts. I took the severance, built something new, and haven't looked back. If it reads as bitter that's probably on me as a writer.
I spent 2.5 years running the largest HCLS engagement at one of the big three cloud providers. 55 engineers, $18M revenue, $31M in bookings. My peers wrote glowing feedback. My customer's CIO keynoted the company's flagship conference. My review said I needed improvement.
This is about why that happens — and why it's not just my story.
You hired TCS, Infosys, Wipro, or Cognizant. The pitch was compelling: $45/hour versus $150/hour. The math was obvious. You'd save millions.
Then reality hit.
You don't have 200 engineers in India. You have 12-18 real shippers in the US and a 180-person tax.
I've seen this pattern repeat across cable, big pharma, insurance, and more. At one Fortune 500, they had 55 engineers on the books. Only 3 were trusted to deliver end-to-end without handholding. The other 52? They generated work for those 3.
Enter AI (But Not How You Think)
Here's where it gets interesting.
Everyone's talking about AI coding assistants. Cursor raised at a $29 billion valuation. GitHub Copilot is on every developer's machine. The narrative is "AI will make developers faster."
That's thinking too small.
The real disruption isn't making your 200 offshore engineers 20% faster at typing. It's eliminating the need for 180 of them entirely.
Gene Kim was right in "Vibe Coding" the implementation bottleneck is gone.
What he didn't say out loud? The organizational bottleneck is next.
Just published: "The Death of the Traditional Product Owner"
In it, I argue that 90% of Fortune 500 Product Owner roles as practiced today will cease to exist within 24 months.
Not because AI is cruel. Because they were always a workaround for a broken system.
The best POs? They're about to become the most valuable engineers on the planet. The rest? They're about to discover what Ops teams learned in 2015.
After 20 years of transforming engineering organizations (Pearson, Aetna, AWS ProServe, Comcast), I've seen this pattern before:
2014: "Containers will never replace Ops"
2016: "Serverless is just a fad"
2025: "Product Owners are essential"
We know how those movies ended.
OutcomeOps makes it safe, auditable, and compliant for engineers to own outcomes directly - even in regulated environments.
No more telephone games. No more 12-18 handoffs per feature. No more requirement translators.
Just engineers shipping business value.
P.S. If you're a Product Owner feeling uncomfortable right now - good. That feeling is your starting line.
Built as a web alternative to ai-token-monitor (Tauri desktop app - https://github.com/soulduse/ai-token-monitor) for headless servers and remote development setups.