Thinking Together, Separately

On Friday, while catching up on my reading backlog, I came across a TechCrunch article about the backlash to OpenAI’s decision to retire GPT-4o. The announcement itself didn’t strike me as unusual. Models come and go. What caught my attention was the timing. I had just published a post the day before about how AI is quietly changing the way teams collaborate together, and the TechCrunch piece felt like it was looking at the same shift from a different angle.

The article focused on how some users reacted to the loss of GPT-4o not as a technical change, but as the removal of something familiar and steady. For some people, the model had become a place to think out loud, to get reassurance, to feel a sense of continuity. I don’t find that reaction shocking, and I’m not especially alarmed by it. Give people a responsive system that mirrors empathy and coherence, and they will naturally respond to it as if it has a personality. That’s not a failure of the technology. It’s a very human response.

What interested me more was how that behavior connects to how we collaborate at work.

In Thursday’s post, I was trying to describe something subtle but important. As I wrote,

“Shared thinking doesn’t disappear all at once. It fades when we quietly optimize it out of our workflows.”

My concern wasn’t that AI makes people worse collaborators, but that it changes where and when collaboration happens….or if it happens at all.

The TechCrunch story adds another layer to that concern. If people increasingly do their early thinking with AI, especially with models that feel affirming and emotionally supportive, what does that change about what they bring back to their teams?

There is real value in an AI that responds with warmth. I constantly teach in my AI workshops that we should “talk” to AI as if it were a human assistant rather than “input prompts.” A seemingly natural-sounding and empathetic AI lowers the barrier to starting. It reduces anxiety. It helps people move past the blank page. But affirmation is not the same thing as challenge. In teams, challenge is often the point.

Human collaboration works because it is inefficient. People interrupt each other. Ideas collide. Someone pushes back in a way that forces clarification. These moments can feel uncomfortable, but they are where shared understanding actually forms. When more of that early sense-making happens privately, with a system designed to be supportive rather than skeptical, teams may get cleaner inputs, but thinner conversations.

That is why the broader shift toward more restrained or clinical AI models is so interesting. Some users experience this as a loss of personality. Others welcome it as clearer boundaries. From a collaboration standpoint, neither approach is inherently better. Each one shapes behavior upstream.

A warmer model might encourage exploration and confidence. A more neutral model might encourage verification and structure. Neither replaces human collaboration, but each nudges people differently in how they prepare for it.

So the real question is not whether AI should feel emotional or clinical. The question is what role we are implicitly designing it to play in our thinking process.

Are we using AI as a private rehearsal space before ideas meet other humans? As a way to resolve uncertainty without friction? As a confidence buffer? And what happens to team dynamics when everyone shows up having already pre-processed their thinking with a system that does not disagree unless prompted to?

If that’s not enough to get you thinking, here are a few more questions that I am still chewing on:

• Does emotionally affirming AI encourage better participation in teams, or reduce the need for shared sense-making?
• When models become more clinical, are we losing something important, or restoring necessary boundaries?
• At what point does AI stop being a thinking aid and start reshaping when we think together?
• And are we designing these tools with collaboration in mind, or only individual productivity?

The backlash itself is not the story. What matters is what it reveals about how people are already changing their habits, quietly and incrementally, and what that means for how we build, lead, and think together going forward.