It’s not headless, it’s more arms.

Sun, 04 Oct 2026 12:00:00 -0500

Not the headless horseman, the octopus. When Salesforce says headless, they mean API-first: the capability gets embedded wherever work happens. It’s not going away, it’s going more places. The head’s still there making the calls. What changes is reach, more arms into more surfaces, each one sensing and acting on its own, all still one animal. Same brain, more arms.

And that’s exactly why design matters more here, not less. The output is fragmenting across different areas, and that’s where you start to lose control. Every surface it lands on is a place your work has to present itself, hold its logic, and stay consistent. Headless doesn’t retire the interface. It scatters it, and scattered is harder. Presenting your work well in someone else’s context, and keeping it consistent across all of them, is the whole job.

Agent count is a flex, not a feature.

Sun, 04 Oct 2026 12:00:00 -0500

Yes, we run multiple autonomous agents. Most serious systems do. But the number is plumbing, the same way nobody cites their microservice count. Show the outcome, not the machinery. If someone wants to see which agent did what, it’s in the reasoning trace. Otherwise it’s noise.

Takes by Wes Jones

It’s not headless, it’s more arms.

Agent count is a flex, not a feature.