Time-sensitive item up front: Cameron Tonkinwise, whom I quote incessantly on the Picnic, is doing a talk about Design Futures at the University of Technology Sydney with Lucy Kimbell. You can sign up to attend live (all 5 of my Australian readers; say hi to each other in the comments!) or to download the recording after the session runs.

There is a vicious cycle in Product that I’ve written about before: we all hate meetings, because our meetings are poorly designed. But instead of designing the meetings to be better, we go all-in on a culture of handoffs. Then when our vision reaches production in a mangled and unrecognizable state, we end up having those meetings anyway, and they are all the more painful for being put off for so long.

Designers are the most guilty of this, but I’ve worked with many PMs who throw a Jira ticket over the fence and expect their vision to materialize in working code sometime within the next two weeks.

You can see how, in an environment where people don’t know how to talk to each other, LLMs feel like liberation. Suddenly you don’t have to fix the thorny problem of working together with other humans. You can pretend that you are the expert on every part of your business, and replace everyone else with an app that your company pays for (at least for now). Said app will give you such quick and confident (but not necessarily correct) answers that it will start to feel like asking humans is now obsolete.

Meetings are the only real influence we have

But don’t get too comfortable. Studies show that this magical feeling drives us further into isolated modes of working — and as we do so, we give away any influence we might have had over how work actually gets done.

You are probably feeling the organizational effects of charismatic technology … Talking to other people and making decisions about how to work together is not novel, making it hard to believe that this is part of the necessary response to the charisma machine - it feels too banal, too small, too simple. Yet this is precisely how beneficial new uses of this technology will be come to pass, in the context of decent work, through figuring it out together.

Wylie’s piece collects several articles (by Elizabeth Ayer, Kevin Baker, and Marc De Pape) that are themselves worth reading. The picture that Wylie paints by bringing these sources together is one that will feel very familiar to us all: automation obscures the logic behind business processes by making it appear to be beyond reproach. A signal that efficiently passes through a company towards its designated recipient without mediation is a signal that cannot be questioned in any meaningful way by the workers who carry it out.

And unlike prior kinds of automation, AI automates the driver rather than the carriage — we are still doing the same tasks as before, but now you are responsible for how the robot “decides” to execute them. In a “human in the loop” system designed to prioritize efficiency, the human is always going to be brought in only after the AI has already made a mess of things:

The one I worry about is that the agents attempt to remediate, and their attempt makes things worse. And it’s only after these failed remediation attempts that humans enter the loop. … The original problem was already so complex that the agents couldn’t handle it, and they have now made it worse by trying to remediate.

As a result, our ability to control what is done in our name continues to shrink, while the scope for which we are accountable continues to expand.

Making meetings work is work

Meta’s AI org was staging a maximalist experiment in whether this could all really work: a system where AI-powered managers can handle many more direct reports, where meetings are cut down in favor of a widget throughput vision of productivity. I say “was” because it was a well-publicized disaster, and the company has already committed to rolling it back.

But more modest versions of this experiment (a low simmer compared to Meta’s rolling boil) are now taking over all knowledge work. The hyperfocus on the individualistic mode of productivity is leaving workers depressed, because it turns out that interacting with other humans was what made the work meaningful — to an identity-forming degree that we are now struggling to replace.

So why do we keep running these experiments? Perhaps because we collectively decided that beacuse meetings don’t work, they can’t work. After all, even masterfully-facilitated workshops by professionals at the top of their game end up going nowhere. And usually, this is because their conclusions are rejected by widget-mode incentive structures that rule the broader organization:

A powerful conversation is not evidence of system change. What happens under the temporary conditions of a facilitated gathering may have very little effect on the structures, incentives, networks and relationships that operate when people leave. It is easy to be seduced by what happens in a good conversation to believe that change and transformation can occur there.

A good meeting (remember: a workshop is just a meeting with rules) facilitates more than just the action. Its success also requires social alignment and internal synthesis, which that take place outside of the official agenda (and even outside of the calendar block).

If we want to have a real say in the future of work, this is what we need to be thinking about: how do we come together, and how must the outcomes of that collaboration continue to resonate once we are apart?

— Pavel at the Product Picnic