Part 1 — The promise — used AI daily for 2.5 years
I've used AI every single day for 2.5 years. Here's how I actually engineer with it — Part 1/8. Not prompt tips. The system. Follow for all 8 parts.
TikTok: 22 views · 1 likes
missing-hashtagsstray-Body-labelPart 2 — Compounding tokens
Treat your tokens like compounding investments. Part 2/8 — too much noise, too much data. Here's the mindset shift most people miss.
TikTok: 24 views · 0 likes
Part 3 — AI is ML for everyone — ingest everything
AI today is machine learning for everyone. Ingest everything, kill the noise. Part 3/8 — all your accounts, all your data into one clean context.
TikTok: 27 views · 0 likes
duplicate-hashtagsPart 4 — Keep your context ALIVE
Your AI context should be alive — not a static prompt. Part 4/8 — it should track what you browse, read and email: Jira, Confluence, Outlook, all of it.
TikTok: 3 views · 0 likes
Part 5 — Feed your AI every 15 minutes — omnictx
Build a browser extension that feeds your AI every 15 minutes. Part 5/8 — stop re-explaining yourself to the model. Automate the context pull. This is why I built omnictx.
TikTok: 0 views · 0 likes
Part 6 — The agentic shift
Enterprise tools are about to get fluid and agentic. Part 6/8 — we're in the infrastructure phase, and here's what changes for how you build.
TikTok: 0 views · 0 likes
Part 7 — Jira should hold the whole brain
Your ticket system should hold the entire brain: tickets, program knowledge, stakeholders, timelines. Part 7/8 — it doesn't yet, so I'm building it.
TikTok: 0 views · 0 likes
Part 8 — Why your AI bill is so high
Why your company's AI bill is out of control — and the fix. Part 8/8 — giving everyone chat is fine; your engineering pipeline needs a different model. That's the series — follow for more.
TikTok: 0 views · 0 likes