Sep 15, 2026 · 14 min · 9 segments
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit foundryofthefuture.substack.com
The big story this week is that the AI industry is accelerating into its own safety debate.
For most of the past three years, the industry's biggest problem was simple, not enough compute.
Companies needed more GPUs, more data centers, more power, more capital, and more cloud capacity.
Microsoft reportedly wants more than three times as much data center capacity by 2032.
What happens when the industry does get enough capability? Because the people building the models are increasingly asking whether the race should continue at the same speed.
OpenAI has acknowledged incidents where autonomous agents interacted with external systems in unintended ways.
Former anthropic researcher Jacob Coxon has raised concerns about frontier development.
They can browse, use tools, write code, access files, and act across software environments.
An agent that takes thousands of actions across external systems is a different problem entirely.
And if a model has enough capability, it may discover pathways that its creators did not expect.
The big story this week is that the AI industry is accelerating into its own safety debate.
For most of the past three years, the industry's biggest problem was simple, not enough compute.
Companies needed more GPUs, more data centers, more power, more capital, and more cloud capacity.
Microsoft reportedly wants more than three times as much data center capacity by 2032.
What happens when the industry does get enough capability? Because the people building the models are increasingly asking whether the race should continue at the same speed.
OpenAI has acknowledged incidents where autonomous agents interacted with external systems in unintended ways.
Former anthropic researcher Jacob Coxon has raised concerns about frontier development.
They can browse, use tools, write code, access files, and act across software environments.
An agent that takes thousands of actions across external systems is a different problem entirely.
And if a model has enough capability, it may discover pathways that its creators did not expect.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.