I realised something few days back. Our human brain has a lots and lots of questions that are keep on raising according to the situation, what we see,what we hear and what we think. But most questions are not meant to be asked ,hypothetical or we don’t know to whom we have to ask that. At that time, this artificial intelligence becomes the go to person for humans.Even in decision making situations or anything related to studies, career we just need a support or like a company.
This framing lands for me, especially "partner in thinking, not just a tool." What I keep noticing is that most AI-as-partner setups still optimize for agreement: they retrieve what's similar, surface what confirms. (Your piece on similarity vs. what "belongs together" touches the same nerve.) But the moments where my own thinking actually moves are the ones where two things I believe don't cohere. I've been building a small system around exactly that — a second mind whose job isn't to answer but to point at my own contradictions. Different axis from your human+AI pairing, but I keep landing in the same place: the second mind earns its keep when it disagrees with the first, not when it echoes it. Subscribed — curious where you take this.
My initial thoughts about AI are that it can be an extremely useful tool for scientists and scholars who already have well developed reasoning, cognition, and a deep knowledge about specific subjects but also a broad knowledge of the world. On the other hand, it could easily be a crutch for those just learning their way in life, and in any academic subject.
I also think that it will, like most technology, widen the gap between creators of ideas and content and consumers of that content.
This all sounds really interesting! I'm curious if you've thought about the potential negative effects of all this technology in our lives. I wrote about this idea a bit in a fun little story the other day.
That’s a great question. I tend to think AI is similar to previous technological shifts like the internet. There are real risks, especially for children, but I’m not convinced that simply banning access until a certain age is the answer.
The technology is likely to become a normal part of life, and eventually everyone will need to learn how to use it. The more important challenge is teaching people to use these tools in ways that amplify their own abilities rather than replace them.
The difference between empowerment and dependency matters. Used poorly, AI can become a shortcut that weakens critical thinking. Used well, it can help people learn faster, explore ideas more deeply, and become more capable than they would be on their own.
We had similar fears when the internet arrived. Some concerns turned out to be justified, others less so. What mattered most was learning how to navigate the new technology responsibly. My guess is that AI will be much the same.
Sounds great Andreas, something I’m currently exploring myself. Your letters have given me some new ideas to consider. Very interested to see what you create more of
Thank you for starting this kind of conversation. I really think we need to have an open discussion about AI and what the benefits and risks are, just like any other technological advance that has been successful throughout history. I honestly think there can be balance between AI and human intelligence that allows us to work with AI without depending solely on it. It's a great tool for resources, bouncing off ideas. But it's also important to not be so dependent on it by conducting your own research.
After working with AI in depth for the past year, I really connect with your area of research. I’d add that there is a hand in how these two minds operate that can be unseen, the builders of the models. I’m excited for your journey here and to learn from your research.
Hey Andreas — "when does AI help us think better and when does it make us more passive" is the question nobody in the productivity-hype space wants to sit with. Glad you're starting there instead of with benchmarks.
Your research angle on continual learning is exactly the gap most practitioners feel but can't name. I build custom AI agents for businesses and the single most common client frustration is "why doesn't it remember what I told it last week?" The deployed-model memory problem isn't academic for them — it's the difference between an agent that feels like a colleague and one that feels like a new temp every Monday morning.
The Two Minds framing is honest. Most AI newsletters pick a side — either techno-optimist cheerleading or cautionary doom. Holding both the biological and the artificial as actual thinking partners without collapsing one into the other is harder to write and more useful to read. Following to see how you do it.
Thanks, really glad it resonated! The memory problem is something I think about a lot in my research, so it's interesting to hear it's just as real from the practitioner side.
Staying honest about both sides is something I want to keep doing, even when it's harder to write. Would love to follow what you're building too!
cool
it was an interesting readd!! enjoyed it muchh!!
Thank you!!
Really interesting
Thank you! 🙏
I realised something few days back. Our human brain has a lots and lots of questions that are keep on raising according to the situation, what we see,what we hear and what we think. But most questions are not meant to be asked ,hypothetical or we don’t know to whom we have to ask that. At that time, this artificial intelligence becomes the go to person for humans.Even in decision making situations or anything related to studies, career we just need a support or like a company.
Very interesting! Great point and distinction of where artificial intelligence can play a big role. Let’s connect
Okay!!!Subscribe for subscribe
This framing lands for me, especially "partner in thinking, not just a tool." What I keep noticing is that most AI-as-partner setups still optimize for agreement: they retrieve what's similar, surface what confirms. (Your piece on similarity vs. what "belongs together" touches the same nerve.) But the moments where my own thinking actually moves are the ones where two things I believe don't cohere. I've been building a small system around exactly that — a second mind whose job isn't to answer but to point at my own contradictions. Different axis from your human+AI pairing, but I keep landing in the same place: the second mind earns its keep when it disagrees with the first, not when it echoes it. Subscribed — curious where you take this.
Thank you Giuseppe, great feedback!
My initial thoughts about AI are that it can be an extremely useful tool for scientists and scholars who already have well developed reasoning, cognition, and a deep knowledge about specific subjects but also a broad knowledge of the world. On the other hand, it could easily be a crutch for those just learning their way in life, and in any academic subject.
I also think that it will, like most technology, widen the gap between creators of ideas and content and consumers of that content.
This all sounds really interesting! I'm curious if you've thought about the potential negative effects of all this technology in our lives. I wrote about this idea a bit in a fun little story the other day.
https://deferredmaint.substack.com/p/the-kids-are-not-alright?utm_source=share&utm_medium=android&r=3gps3o
That’s a great question. I tend to think AI is similar to previous technological shifts like the internet. There are real risks, especially for children, but I’m not convinced that simply banning access until a certain age is the answer.
The technology is likely to become a normal part of life, and eventually everyone will need to learn how to use it. The more important challenge is teaching people to use these tools in ways that amplify their own abilities rather than replace them.
The difference between empowerment and dependency matters. Used poorly, AI can become a shortcut that weakens critical thinking. Used well, it can help people learn faster, explore ideas more deeply, and become more capable than they would be on their own.
We had similar fears when the internet arrived. Some concerns turned out to be justified, others less so. What mattered most was learning how to navigate the new technology responsibly. My guess is that AI will be much the same.
Sounds great Andreas, something I’m currently exploring myself. Your letters have given me some new ideas to consider. Very interested to see what you create more of
Nice Andy, thanks for commenting. I just subbed. Excited to see what you are working on too.
Thank you for starting this kind of conversation. I really think we need to have an open discussion about AI and what the benefits and risks are, just like any other technological advance that has been successful throughout history. I honestly think there can be balance between AI and human intelligence that allows us to work with AI without depending solely on it. It's a great tool for resources, bouncing off ideas. But it's also important to not be so dependent on it by conducting your own research.
After working with AI in depth for the past year, I really connect with your area of research. I’d add that there is a hand in how these two minds operate that can be unseen, the builders of the models. I’m excited for your journey here and to learn from your research.
Hey Andreas — "when does AI help us think better and when does it make us more passive" is the question nobody in the productivity-hype space wants to sit with. Glad you're starting there instead of with benchmarks.
Your research angle on continual learning is exactly the gap most practitioners feel but can't name. I build custom AI agents for businesses and the single most common client frustration is "why doesn't it remember what I told it last week?" The deployed-model memory problem isn't academic for them — it's the difference between an agent that feels like a colleague and one that feels like a new temp every Monday morning.
The Two Minds framing is honest. Most AI newsletters pick a side — either techno-optimist cheerleading or cautionary doom. Holding both the biological and the artificial as actual thinking partners without collapsing one into the other is harder to write and more useful to read. Following to see how you do it.
Thanks, really glad it resonated! The memory problem is something I think about a lot in my research, so it's interesting to hear it's just as real from the practitioner side.
Staying honest about both sides is something I want to keep doing, even when it's harder to write. Would love to follow what you're building too!
We sound very aligned across these themes, Andreas. Let’s connect.
Hey Peter, sure! I just followed you back
Excited for what's coming!