Transcription
We're talking about pride of ownership today. And I know you might think, "AI, Nate is an AI guy." I promise you this gets back to AI.
At the end of the day, most of the conflicts we see in our public institutions and in the private workplace are about how we handle pride of ownership in an AI world. And I want to take you through a brief tour of the before times so you get a sense of how much has changed and how our underlying frameworks have not shifted, because I think that gives us the ability to show where we need to advance our work culture, our educational culture, to truly make AI a useful tool.
We begin before AI. We had three implicit questions that we asked every time in work, at school, if we took pride of ownership in something. It is: did you author it? Do you truly know the material? Can you show me the chain of provenance for this material? That happens with property and transactions, but it also happens with work. Can you show me the reviews it went through? It was a conversation I've had about mini-docs. Professors have asked how many books I read. They're implicitly asking about the chain of provenance of the idea. So, show me that you can keep the idea and have a sense of integrity there. And finally, can you show that you have ownership on outcomes? If you're in class, it's: do you have ownership on the grade? If you are at work, it's: do you have ownership of your KPIs?
The point is those questions are not new. We have been asking those questions since Ian Nir complained about his copper shipments because someone was not upholding their end of the agreement and not taking pride in their work. And yes, I am making a very nerdy reference, and I hope someone appreciates it.
The point here is that those underlying components of pride of ownership will not change in the age of AI. They won't. Stop expecting them to. When you get into a fight, whether it's uh, at school, whether it's at work, whatever your context about whether AI is appropriate to use, it almost always comes down to pride of ownership. You need to be able to answer all three of those questions affirmatively in order to have a positive communal AI productivity experience. And I know that sounds weird, like it's very hippie to say it that way, but fundamentally when we use AI one-to-one, where I am talking to the AI, if I am doing so in the context of group work, the group is affected and demands to know what's going on implicitly or explicitly, and if they feel like they don't, the instinct of a lot of groups is to clamp down. And so if you were in an environment like that, you have to recognize those are the three questions people are asking: is: do you know your work? Can you keep your work? Do you understand the chain of provenance? And can you be responsible for the outcomes? Can you hang your hat on what happens afterward? Those are the things that work is expecting.
And I believe we can answer yes on those in the age of AI. It's not impossible. You can do it. You can, in fact, use AI to prompt and ask yourself questions of the data that you have at your disposal. So you know your product area, so you know your domain, so you know your educational subject matter better. You can actually increase your product knowledge. Yes is a spectrum, and you can be more yes on product knowledge or domain knowledge if you use AI well. And so, in a sense, part of the interesting thing about AI is that people tend to assume you can use AI as a cheat sheet and skip the product knowledge part. But it doesn't have to be that way. It can be the reverse.
Similarly, with provenance, you can be transparent about where you are using, evolving, and thinking about your arguments. Sometimes that's as simple as saying, "Chat GPT and I have been working on this together." Sometimes that's as simple as saying, "This is an argument. I evolved it after this conversation. Then I processed it through Chat GPT, and I used this prompt and came back." And that more formal sense might be appropriate if legal might get involved. There are cases now where documents are being created with legal implications inside workspaces, and you have to have some provenance, and that may include the prompts. But even if you're not that formal, it is still appropriate to understand for yourself how you gained conviction in a space. And I think that's the heart of it. And so maybe it's not about logging all your prompts and this and that, although maybe that's a best practice. It's about: is your workspace and the artifacts you're leaving behind in a position where a stranger who is fluent in the art could come in, look at the workspace, look at the artifacts and evolve to a place of similar high conviction that you have about your angle of approach on your work? Does that make sense? It's like: could they look at what you have left behind and gain a sense of conviction because you have left enough behind that there's a bit of a chain of provenance on your thinking? That's what keeping means to me. That's keeping your thinking.
Finally, on outcomes, that I, I think that has changed the least in the age of AI. Like, at the end of the day, you are still accountable for the KPIs, and there is still a gut-level accountability to say, "It's on me," the grade for the essay, it's on me if you're an education, the performance of the team, it's on me if you're a manager; that kind of thing, that level of commitment is not different than it was before, but people sometimes think you're going to skip it because they think that people will depend on AI and then blame AI. I got news for you: you, you just cannot blame the machine. It's the old, the old joke from the IBM slide that a machine can't make managerial decisions. Well, joking aside, it's still kind of true from a human perspective. Humans expect humans to own and be accountable. So you got to be accountable. And that's as old as taking accountability for your copper imports. And I am just going full nerd here. I hope someone appreciates it.
So there you go. It's about making sure that you are responsible for knowing your domain, keeping the provenance of your work in an artifact form that people can understand how you evolve conviction, and you're responsible for your outcomes. That's true before AI, and that's true with AI. And I think if we understood that and we talked about that more, we would have fewer arguments about when and where to use Chat GPT or co-pilot because they would be grounded in what's really going on, which is an agreement in how we work. We need to reforge those agreements in how we work. And I think that framework is a way to think about it. Cheers.