Trust in Enterprise AI adoption, meta-uncertainty, UI for multi-agent Humans + AI, professional services hypotheses, and more

📖In this issue

• Framework: Trust in Enterprise AI adoption

• Open source UI system for human-centered multi-agent AI

• The challenges from AI for major professional service firms

• A week of model releases means more power, lower cost

• Podcast: Vivienne Ming on hybrid collective intelligence, building cyborgs, meta-uncertainty, and the unknown infinite

💡Framework: Trust in Enterprise AI adoption

Much of my work these days is in the transformation of organizations to Humans + AI. Perhaps more than anything else, this requires trust, across an array of different dimensions.

This simple framework maps some of the layers and dimensions of trust required to faciliate and accelerate AI adoption. Of course the next step is the journey of building trust across these dimensions, but the necessary first step is acknowledging and defining where trust needs to be built.

🧠🤖Humans + AI

Open source UI system for human-centered multi-agent AI

Multi-agent systems should be designed to include human as well as AI agents. A new open-sourced interface from Microsoft does exactly that. In Magentic-UI, a lead Orchestrator coordinates specialist agents - including humans - using 6 simple collaboration models as a starting point.

Magentic-UI: Towards Human-in-the-loop Agentic Systems
Key insights and summary

The challenges from AI for major professional service firms

I mapped the stock performance of the largest listed professional service firms, which are the only ones we have any decent data on, with most of the others partnerships or tightly held. They are all down substantially this year, the larger ones more. I have made a list of possible reasons for this very poor performance, there are quite a few but these aren’t the only challenges.

List of hypotheses for the poor performance of professional service firms 

A week of model releases means more power, lower cost

I’m sure you’ve noticed that this week saw the release of Google’s very impressive world-building model Genie 3, OpenAI’s open source open-weights model gpt-oss and of course GPT-5, over 2 years since GPT-4 was released. Plenty to say, but just one point to share here: the pricing of GPT-5 is very aggressive, giving more power at significantly lower cost. Thus the increasing importance of building Humans + AI work and thinking…

GPT-5: Key characteristics, pricing and model card

🎙️This week’s podcast episode

Why you should listen

Vivienne’s accomplishments are extremely impressive, but it is her attitude to AI and technology that are truly inspiring. She believes we all have infinite potential, and we need to nurture that in everyone.