2002. My first IT job with a big company. Fortune 500 medical device company. Chicago regional office. Boss was in Seattle. First initiative they handed me: evaluate copiers. Meet with Konica, Canon, Xerox the usual vendors. Because I was physically embedded in the office — not a remote help desk — I ended up golfing with the division heads. One of them ran order management. Told me the monthly process: 100-page legal order packets, photocopied, stuffed in FedEx envelopes, overnighted to field sales reps and to the hospitals. $50 an envelope. Six to eight at a time. That's why they needed new high-end copiers. OCR had just come out. Multifunction copiers in 2002 could scan to PDF, which meant no more printing. The finance guy mentioned a nightly Access export of all order data out of Siebel. I got my hands on a demo Canon copier. Built an HTML page in Dreamweaver. Scan the packet, name the file with the order number, it lands on a server. The page joins that to the Access database and links to the PDF. Field reps pull packets via VPN from the HTML page. $1M per office per year. 7 offices. Saved. Two hours of work. I got scolded by the IT Director and CIO back in Seattle — no change management, no approval chain. The people on the ground made me their hero. Trip to Cabo San Lucas. Met my wife. Life changed forever. --- The Seattle help desk had the same technical knowledge I did. They never found this problem because they were waiting for tickets. I found it at the golf course, listening to Jeff complain about FedEx bills. @Jake Van Clief dropped a Substack today: "You need to productionize your opinion, not just your process." The printer story is what that looks like before the productionizing starts. You need someone close enough to the workflow to have an opinion worth productionizing. The AI ROI gap most businesses are hitting right now — it's the same gap. The person deploying the technology isn't in the room with the people who actually know what's broken. They're managing the firewall. Attending vendor calls. Evaluating copiers.

After reading the full history of thinking machines: Vote below, then tell us in the comments: which part of the timeline surprised you the most?