Braincandy

Deep dive into topics that matter... and topics that I fell deeply into

Braincandy

Aug 14, 2026

•

10 min read

Your AI is learning. Are you?

The danger is not that AI removes the work. It is that the learning stops landing on you.

Chuck DeVries

Braincandy

The first worker AI replaced was never hired

Aug 5, 2026

•

14 min read

The first worker AI replaced was never hired

The layoffs are arriving before the automation, and our instruments cannot see it.

Chuck DeVries

Braincandy

The second-order game: agent memory is the new lock-in

Jul 17, 2026

•

4 min read

The second-order game: agent memory is the new lock-in

When the bartender remembers your drink it's not really because he's your friend

Chuck DeVries

Braincandy

Jul 10, 2026

•

3 min read

The Loop That Taught the Model

A model becomes what you reward. So does an organization.

Chuck DeVries

Braincandy

Reinforcement Loop, Part 2: why your best model loses to a worse one

Jun 25, 2026

•

4 min read

Reinforcement Loop, Part 2: why your best model loses to a worse one

Why best model may not be the best

Chuck DeVries

Braincandy

Reinforcement Loop, Part 1: the loop in one picture

Jun 18, 2026

•

3 min read

Reinforcement Loop, Part 1: the loop in one picture

Most AI rollouts get one good quarter, then flatten. The reason is a loop that never closes.

Chuck DeVries

Braincandy

Capability to Consequence: the one shift behind every AI headline

Jun 11, 2026

•

3 min read

Capability to Consequence: the one shift behind every AI headline

Stop asking what AI can do. Start asking what changes when it touches real work.

Chuck DeVries

Braincandy

Your AI is learning. Are you?

The first worker AI replaced was never hired

The second-order game: agent memory is the new lock-in

The Loop That Taught the Model

Reinforcement Loop, Part 2: why your best model loses to a worse one

Reinforcement Loop, Part 1: the loop in one picture

Capability to Consequence: the one shift behind every AI headline

Trouble Worth Making