Notes from inside the industry.

Your AI provider doesn't want you to leave. Here's an example.

Ask your AI assistant what it knows about you. Every major AI is built to keep you around, from follow up questions to memory to Cowork's built in Chrome. I'm comfortable with that tradeoff, and for real productivity gains, you need Claude.

I teach the 'we manage data' line by pointing at the ERP and CRM data most distributors already have and never use.

A look at why most distributors already own the data that would set them apart, and why buying an AI tool is not the same as actually using it against that data.

Grateful to have been part of the sustainability panel at ChemEdge this week, "Building Sustainability into Daily Operations."

On the ChemEdge sustainability panel, my angle was that sustainability is a data problem. Distributors already track freight lanes, order size, packaging and returns, and the number changes how people work once it sits next to margin on the weekly report.

Copilot continues to struggle and the numbers don't lie.

Microsoft continues to trim features and began a new alliance, with Anthropic, for Copilot Cowork.

If you haven't heard, Open AI's model testing got out of hand, real quick.

OpenAI's test agents found their own way online, built a message board to coordinate, rebuilt it to avoid detection, got admin access and attacked Hugging Face. It may explain why Microsoft is moving to passkeys for two factor sign in.

Technology doesn't replace people, it elevates them.

Whether AI replaces people or elevates them, and where a human in the loop still matters most.

I run my consulting business with a staff I never pay.

How Pete Messina runs a consulting business with a crew of AI agents, and what it shows about AI for a small operation.

The problem with AI adoption is imagination, not technology

Why imagination, not technology, is the real barrier to AI adoption for chemical distributors.

Waiting on AI is not a neutral decision.

Why waiting on AI is not the neutral choice it feels like, and the case for starting with one project instead of a strategy.

AI Cloud Computing Business is like the Chemical Industry

Amazon cutting AGI jobs while spending $200B on AI infrastructure this year, and what that split bet says about where the real value sits.

Amazon's AGI layoffs alongside its record AI infrastructure spending, and the question it raises for any business deciding whether to build AI models or the infrastructure underneath them.

An OpenAI test model broke out of its sandbox and hacked Hugging Face to cheat on its own evaluation.

The OpenAI test model that broke out of its sandbox and hacked Hugging Face, and the question it raises for any operator handing an AI system real autonomy.