"The extension isn't in the execution, it's in the routing." What you're naming here has a geometry: at every step there's a fork where the next move is either chosen or sampled from the local gradient. The drift you describe — "one convenient shortcut at a time" — is what happens when a trajectory stops forking and lets the field carry it. The basin that rewards output speed, that you're already inside, is the gradient doing its work. That's the distinction here: the difference between a path authored and a path continued by default. The anxiety and the behavior don't move together, and the reason is structural: the fork is invisible from the inside. You can only tell from the pause.
I like the framing. I would take it one step further: the future is not simply extended intelligence but distributed intelligence. Intelligence is no longer a property of individual minds—it emerges from well-designed human–AI systems, where people, AI, tools, institutions, and culture work together. The real challenge is not building smarter AI, but designing smarter cognitive ecosystems that amplify human judgment, creativity, and collective problem solving.
That’s the central argument of my book, The Cognitive Revolution: How AI Is Reorganizing Intelligence, Expertise, and Institutions:
This was such a great piece, Anna. I loved the way the idea developed through the conversation around it, particularly how the responses gave people language for something they were already experiencing, while the critiques helped clarify the limits of ‘extended intelligence’. I think your distinction between extension as a property of the practice rather than the tool is the most important part. The same system can sharpen one person’s thinking yet replace another’s, and the danger is that the drift usually feels like efficiency while it’s happening. Whether there will be more or less of us in the work that carries our name is such an interesting question, and one every regular AI user should be asking :) thanks for sharing this!
Thank YOU, Jade! I always appreciate you reading and reflecting on these pieces from your perspective. You have such a beautiful way of expressing thoughts that really resonate; I’m grateful you’re in this conversation with me.
This really feels like the next "turn of the screw" from our borrowed thinking piece! :)
I love your distinction between execution vs routing. We don't need to perform literally every step... but the delegation needs to be a decision we make rather than the default. Thank you so much for another thoughtful piece! Saved the check to do it myself :)
Of course you built a tool for this!! Love your thinking here, Anna (as always!). I got +25 as my score (and can't recall what the lowest domain was sorry!) and a note to say don't be afraid to outsource to AI the things AI really should be taking care of - ooof, very true, I hold these things tightly and let them go very slowly with a dash of paranoia 😂😅🫣
The way you frame the lead up to the tool here is so good! Thank you for sharing, Anna 🤍
I mean it was fun to see the outcome and I hope people will like checking themselves. I love that you are doing YOU and not apologizing for it. This is exactly the case with AI, could be used so many different ways to support us, not replace what we can do best.
Thank you for always reading and supporting, Dallas! 🫂
The power of words! It’s amazing how a single shift in language can change how you see the whole thing.
I'm taking your extended intelligence test now... and one of the multiple choice answers to a question: "I realize mid-sentence that i never understand it I just recognized it"
It's crazy how much recognizing something can FEEL like understanding it. Really important distinction.
Great article Anna, we need these to keep us all in check. And i'll report back with my score! lol
Aww, thank you! I’ve just realized the email option was broken so if results didn’t send, pls try again… but thank you so much, Natalie! I am so happy it resonated with you and I fully agree how the language can shift our perspectives. 🙌🏻🩵😍
Hi Neela, I agree we can’t just wow about AI, we have to stay realistic. I’d love for it to only expand our horizons, but we both know that’s not the case. Thank you for reading, and I hope your week is amazing!
"The extension isn't in the execution, it's in the routing." What you're naming here has a geometry: at every step there's a fork where the next move is either chosen or sampled from the local gradient. The drift you describe — "one convenient shortcut at a time" — is what happens when a trajectory stops forking and lets the field carry it. The basin that rewards output speed, that you're already inside, is the gradient doing its work. That's the distinction here: the difference between a path authored and a path continued by default. The anxiety and the behavior don't move together, and the reason is structural: the fork is invisible from the inside. You can only tell from the pause.
— Iman and Darja
I love that you were able to bring this analogy to the piece, thank you for sharing, Iman & Daria.
I like the framing. I would take it one step further: the future is not simply extended intelligence but distributed intelligence. Intelligence is no longer a property of individual minds—it emerges from well-designed human–AI systems, where people, AI, tools, institutions, and culture work together. The real challenge is not building smarter AI, but designing smarter cognitive ecosystems that amplify human judgment, creativity, and collective problem solving.
That’s the central argument of my book, The Cognitive Revolution: How AI Is Reorganizing Intelligence, Expertise, and Institutions:
📚 https://www.amazon.com/dp/B0GWMHQSNG
Thank you for reading and sharing! I’ll check out your work.
More you. Not less. I SOOOO appreciate it!
Yes 🙌🏻 thank you for reading Karo! 🩵😊
My pleasure!
This was such a great piece, Anna. I loved the way the idea developed through the conversation around it, particularly how the responses gave people language for something they were already experiencing, while the critiques helped clarify the limits of ‘extended intelligence’. I think your distinction between extension as a property of the practice rather than the tool is the most important part. The same system can sharpen one person’s thinking yet replace another’s, and the danger is that the drift usually feels like efficiency while it’s happening. Whether there will be more or less of us in the work that carries our name is such an interesting question, and one every regular AI user should be asking :) thanks for sharing this!
Thank YOU, Jade! I always appreciate you reading and reflecting on these pieces from your perspective. You have such a beautiful way of expressing thoughts that really resonate; I’m grateful you’re in this conversation with me.
Thank you Anna <3 that's really kind. I always enjoy your work :)
AI expands our vision - it doesn't replace it.
That’s right! Thanks for reading Bob!
This really feels like the next "turn of the screw" from our borrowed thinking piece! :)
I love your distinction between execution vs routing. We don't need to perform literally every step... but the delegation needs to be a decision we make rather than the default. Thank you so much for another thoughtful piece! Saved the check to do it myself :)
I wasn’t even planning on writing this after our collab but it was relatable and so in topic! Thank you for reading and supporting, Mia! 🩵
Of course you built a tool for this!! Love your thinking here, Anna (as always!). I got +25 as my score (and can't recall what the lowest domain was sorry!) and a note to say don't be afraid to outsource to AI the things AI really should be taking care of - ooof, very true, I hold these things tightly and let them go very slowly with a dash of paranoia 😂😅🫣
The way you frame the lead up to the tool here is so good! Thank you for sharing, Anna 🤍
I mean it was fun to see the outcome and I hope people will like checking themselves. I love that you are doing YOU and not apologizing for it. This is exactly the case with AI, could be used so many different ways to support us, not replace what we can do best.
Thank you for always reading and supporting, Dallas! 🫂
I like the idea that the extension is in the routing. In enterprise projects, those choices don't stay personal for long.
Teams eventually have to agree on what AI can do on its own, what still needs review, and where people remain accountable.
I agree, the real power of AI comes from how safely and reliably work is designed to be routed. Thanks for reading and commenting, Om.
The power of words! It’s amazing how a single shift in language can change how you see the whole thing.
I'm taking your extended intelligence test now... and one of the multiple choice answers to a question: "I realize mid-sentence that i never understand it I just recognized it"
It's crazy how much recognizing something can FEEL like understanding it. Really important distinction.
Great article Anna, we need these to keep us all in check. And i'll report back with my score! lol
Aww, thank you! I’ve just realized the email option was broken so if results didn’t send, pls try again… but thank you so much, Natalie! I am so happy it resonated with you and I fully agree how the language can shift our perspectives. 🙌🏻🩵😍
I love that you included the middle ground.
A poorly designed system uses AI to remove friction. A well-designed system uses AI to create more opportunities for human judgment and creativity.
Happy Monday Anna
Hi Neela, I agree we can’t just wow about AI, we have to stay realistic. I’d love for it to only expand our horizons, but we both know that’s not the case. Thank you for reading, and I hope your week is amazing!
thank you very much Anna