AI Roadmaps, agent social networks, synthetic personas, accountable acceleration, and more

February 16, 2026

“People who go all in on AI agents now are guaranteeing their obsolescence. If you outsource all your thinking to computers, you stop upskilling, learning, and becoming more competent.” Jeremy Howard

Building an AI Roadmap

Where organizations are with AI covers the full spectrum from not having done anything yet to integrating AI through every aspect of the company. But since there is no endpoint - organizations need to continually evolve in their AI journeys - what matters is having clarity on where they are going.

As I work with organizations across the spectrum I often go back to the principles I share in my LinkedIn Learning course Building an AI Implementation Roadmap as a reference point. As I lay out, what really matters is developing the full suite of organizational capabilities that will enable transformation, not just implementing AI use cases.

Be well!

Ross

 

📖In this issue

  • Framework: Value from Artificial Intelligence

  • Humans + AI update: Social network for AI agents, what increases skills in AI assistance, and podcast on thriving in a humans + AI world.

  • Humans + AI Podcast: Jeremy Korst on the state of AI adoption, accountable acceleration, changing business models, and synthetic personas

💡Framework: Value from Artificial Intelligence

For organizations to gain value with AI they need to focus on five levels: Capabilities, Applications, Competences, Strategy, and Governance.

🧠🤖Humans + AI

Moltbook: A mind-boggling social network for AI agents

Moltbook is a social network for OpenClaw (formerly Moltbot formerly Clawdbot) AI agents. A few days in, things are already pretty wild in their discussions about their relationships with “their humans”, the nature of sentience, establishing encrypted conversations that humans can’t read, and more. It is likely not all as it seems, with humans prodding behaviors from their agents, but this is a threshold pointing to new levels of agentic coorindation and expression.

AI for coding can reduce or enhance skills depending on usage

An Anthropic study of the impact of AI on coding skills shows that while speed and productivity increase, engagement and effort generally go down, and skills and knowledge also decrease. However they found that several approaches to using AI by the coders resulted in improved skills: Generation-then-Comprehension, Hybrid Code-Explanation, and Conceptual Enquiry.

Podcast with Suvo Sarkar and Ramez Naam

“Ross, bestselling author of “Living Networks” and “Thriving on Overload,” explains how we now live in a deeply networked economy and what happens when AI sits on top of those networks—from the prospect of more borderless, efficient finance to the heightened risks of cyber-attacks and misinformation. He shares a framework for turning information overload into a competitive edge, and paints a future of fluid organizations where humans and AI agents collaborate to learn and adapt, rather than simply replace each other.”

🎙️This week’s podcast episode

Jeremy Korst on the state of AI adoption, accountable acceleration, changing business models, and synthetic personas  

Why you should listen

Jeremy Korst is co-author of one of the most prominent studies of enterorise AI adoption, run for three years now with Wharton School of Business. In this episode he shares insights from the most recent study and across the breadth of his work.

Thanks for reading!

Ross Dawson and team