Notes
Posts about data and AI, and about the gap between the demo and Monday morning. Most were written for LinkedIn, a couple of them on my phone at 30,000 feet. Kept here, lightly tidied, because LinkedIn forgets.
All notes
-
Learning from an AI tour of South East Asia
At SuperAI Singapore the conversation had moved from "look what the model can do" to "we handle x% of inbound calls with it". I like it when AI gets boring.
-
The ceiling and the floor
Benchmarks get the headlines. Stability, support and paginated reports are where most business value still hides, in both AI and BI.
-
BI in 2025: still pasting into Excel
Twenty-five years of "no more exports". Why the last mile of decision-making still lives outside the tool, and what would change that.
-
Hi AI: how real is real?
From GPT-3 poems to robots in Shanghai hotel lifts. Deploy at scale, fail at scale, learn at scale.
-
How big should a data team be?
Gartner's team-size ranges against a $50 million business, and a question to data leaders. With a 2026 postscript.