Tell me more. What do you not like about Uncharted Territories, and what do you think a better name would have? This question is for anybody who might have an opinion
The Game of Thrones analogy is the sharpest insight in this piece — and it applies far beyond entertainment. The limiting factor was never production budget. It was the storyteller. After a decade in energy trading and risk management, I’ve watched the identical pattern play out in market research and analysis: the explosion of data and AI tools has made it trivially easy to produce professional-looking commodity reports. And most of them are worthless — because the analytical framework underneath is generic.
Your point about insights versus information is the one the entire media industry needs to absorb. In energy markets, I can get AI to produce a perfectly formatted report on oil price movements in under a minute. It will be accurate, well-structured, and completely useless — because it will tell me what happened without telling me why it matters or what it means for positioning. The insight layer — connecting a refinery outage in the Gulf to a fertiliser supply constraint in Southeast Asia to a food price risk in Africa — requires a mental model of how the physical world actually works. That’s the part AI can amplify but cannot originate. And it’s the part that makes the difference between content people skim and content people act on.
The authenticity point deserves more attention than it typically gets in AI discussions. In commodity trading, the most trusted research voices aren’t the ones who are always right — they’re the ones who are transparently wrong when they miss, and who explain why they missed. That accountability creates compounding trust that no amount of polished, AI-generated analysis can replicate. When I read a piece by someone who says “here’s what I got wrong last month and here’s what I’ve updated in my model,” I pay ten times more attention to their next call than I do to a flawless-looking report from an anonymous desk.
Your framework of what quality means — insightful, entertaining, truthful, honest, unbiased, authentic, personalised, at scale — is essentially a product specification for what trusted media looks like in an AI-saturated environment. What strikes me is that six of those eight qualities are fundamentally about the human behind the content, not the tools producing it. AI handles scale and personalisation. Everything else — the courage to take unpopular positions, the honesty to correct mistakes, the depth of worldview that produces genuine insight — those are human moats that widen as AI commoditises everything else.
One dimension I’d add from building AI-assisted research systems: the media companies that will win aren’t just the ones using AI to produce more content. They’re the ones using AI to deepen research on fewer, higher-stakes topics. The real unlock isn’t volume — it’s depth. Using AI to synthesise across hundreds of sources, surface non-obvious connections, and stress-test arguments before publication. That produces the kind of insight density that makes readers feel like every piece made them meaningfully smarter — which is, as you correctly identify, the only thing that sustains attention in a world drowning in content.
The vision here is ambitious in exactly the right way. Appreciate the intellectual honesty about what AI can and cannot do — and the clarity about where the human remains irreplaceable. Following closely.
Indeed, it's a product spec, because that's the first section in my internal strategy doc!
I think some of them are human indeed, but maybe not all. AIs could be much more, if well structured. They aren't today. Which is why merging humans with AI is the best next step.
Depth is indeed valuable, but I need to change UT to reap it. Right now I touch so many topics that my depth is sporadic, and that prevents from getting enough supporters on any given topic.
In any case, each one of your thoughts made me think. Thanks again.
Wondering about some of your assumptions, eg, bias. Is it really possible to be unbiased? While the aspiration may be needed, it might be more honest and authentic to regularly self explore what biases actually are at play.
Great insight and you have my support. Most of the mechanics of what you are proposing seem doable with people, time and resources, mainly AI. I see one one aspect that you haven't fully explained or explored here and that is the world view used to measure and push the direction of the content.
I've used and enjoyed pushing Grok and Claude to explore political events and more interesting climate change projections. The biases embraced by both AIs is difficult to shake because in many instances the news feeds they source are biased by volume alone. Getting an AI world view that is stable (won't drift) in my view is the most critical aspect of your project. It would be used as a yard stick to craft the narrative.
If we're not careful AI will be used to roll back our ability to communicate all the way to the Neolithic. This applies now: we will be limited to face to face contact since everything else can be maliciously faked with AI. For example who still answers phone calls? But also over time. We are basically about to un-invent writing. AI is poisoning our more permanent records, like research, books. Go to Amazon and a lot of searches for books already yield AI slop. And many Google search results almost completely consist of AI slop content farms.
(Also reminder to people that, if you spend hours and hours scrolling 30 sec videos on Tik Tok you're little more than a meat bag plugged into the Matrix. So don't do that, and if you know people who do, maybe gently get them to stop)
An interesting element to the "content explosion" is that the people creating most of the content are still limited in their imagination on what is actually interesting. We seem to be getting "lots more of the same" rather than people wandering off to create their own unique niches in the information ecology. I think that's the trick to survival. The tools aren't coming up with good ideas (how can they, LLMs are trained on what is already there) - but they will hyper-accelerate those that are creatively different in what they approach.
In general substacks articles are far too long and in old fashion writing style. There is so much high quality content an so little time!!! We need short paragraphs almost bullet like points. Give your point of view or info in no more than 5 minutes reading time. It si not a question of attention span as time available. Axios is in my opinion how info/opinion should be transmitted. New times demand a new writing style.
It will be a good starting point to get mone recognition, which in turn will can be levarage into merchandizing, books, paid engagements and product placements.
There are several channels that started in “niche” areas that have grown to be very successful. They found something that was very interesting to say and that could be understood by large segment of the general population.
I would check veritasium, specially this video where is shows how it grew and became monetized https://youtu.be/piHGnG4LsmQ?si=pUNwQA4Za0ThK72N. It also is a good example as how to grow with support of other people so that it does not take all of your time.
Other specialized channels that grew are
MentorPilot
Practical Engineering
Production cost could be a lot more managable. Feedback is more immidiate that writing a book.
I understand you perspective here and I absolutely want and ethical and well regulated Ai ecosystem. But as one of the many thousands of performers whose employment is about to be massively slashed, or eradicated entirely by those production companies fighting for supremacy and maximum profits for their share holders, disagree with the notion that’s these companies will be pumping out the best quality content for a long time. And when they do it still won’t have been worth the human cost. Our only hope is that the general public have the stamina to keep pushing back against producers who want to eliminate human performers from their shows. But knowing how easily swayed people are, that’s no hope at all. 😢
The chart captures something most people feel but can’t articulate — AI doesn’t just add more content, it reshapes the entire quality distribution. The middle collapses. What’s left is either genuinely great or commoditized noise. The bet worth making is on the quality end, because that’s where human judgment still has a moat.
One small nuance: The middle gets extremely bloated. I read tons of AI articles now that are OK good. That shifts the bar though, and that level of quality then becomes the commoditized noise you discuss.
Get a better name than "Uncharted Territories". That was ok for the curious, but if you are thinking mainstream, abet something more positive.
Tell me more. What do you not like about Uncharted Territories, and what do you think a better name would have? This question is for anybody who might have an opinion
What about "The Explorer"? It resembles The Economist and keep the spirit of the current name. An Explorer goes into a Uncharted Territory
I can see how the term "Uncharted Territories" can imply niche newsletter status.
From my vantage point, the tension that must be navigated is the following:
On the one hand, maintaining intellectual curiosity-led rigour, UT was founded on (for me)
with
On the other hand, a broadening capacity that signals trust also beyond Tomas as a person, and thus acquire mainstream relevance more easily.
Adaptation from my head: “The world (or territories) of the 21st century”
To be discussed further
Hi Tomas, thanks for your article. I saw a link for hiring but not one for investing, are you planning to include one?
Yes sorry, probably. I should. Thanks for reminding me!
The Game of Thrones analogy is the sharpest insight in this piece — and it applies far beyond entertainment. The limiting factor was never production budget. It was the storyteller. After a decade in energy trading and risk management, I’ve watched the identical pattern play out in market research and analysis: the explosion of data and AI tools has made it trivially easy to produce professional-looking commodity reports. And most of them are worthless — because the analytical framework underneath is generic.
Your point about insights versus information is the one the entire media industry needs to absorb. In energy markets, I can get AI to produce a perfectly formatted report on oil price movements in under a minute. It will be accurate, well-structured, and completely useless — because it will tell me what happened without telling me why it matters or what it means for positioning. The insight layer — connecting a refinery outage in the Gulf to a fertiliser supply constraint in Southeast Asia to a food price risk in Africa — requires a mental model of how the physical world actually works. That’s the part AI can amplify but cannot originate. And it’s the part that makes the difference between content people skim and content people act on.
The authenticity point deserves more attention than it typically gets in AI discussions. In commodity trading, the most trusted research voices aren’t the ones who are always right — they’re the ones who are transparently wrong when they miss, and who explain why they missed. That accountability creates compounding trust that no amount of polished, AI-generated analysis can replicate. When I read a piece by someone who says “here’s what I got wrong last month and here’s what I’ve updated in my model,” I pay ten times more attention to their next call than I do to a flawless-looking report from an anonymous desk.
Your framework of what quality means — insightful, entertaining, truthful, honest, unbiased, authentic, personalised, at scale — is essentially a product specification for what trusted media looks like in an AI-saturated environment. What strikes me is that six of those eight qualities are fundamentally about the human behind the content, not the tools producing it. AI handles scale and personalisation. Everything else — the courage to take unpopular positions, the honesty to correct mistakes, the depth of worldview that produces genuine insight — those are human moats that widen as AI commoditises everything else.
One dimension I’d add from building AI-assisted research systems: the media companies that will win aren’t just the ones using AI to produce more content. They’re the ones using AI to deepen research on fewer, higher-stakes topics. The real unlock isn’t volume — it’s depth. Using AI to synthesise across hundreds of sources, surface non-obvious connections, and stress-test arguments before publication. That produces the kind of insight density that makes readers feel like every piece made them meaningfully smarter — which is, as you correctly identify, the only thing that sustains attention in a world drowning in content.
The vision here is ambitious in exactly the right way. Appreciate the intellectual honesty about what AI can and cannot do — and the clarity about where the human remains irreplaceable. Following closely.
Very very astute. Thanks for sharing.
Indeed, it's a product spec, because that's the first section in my internal strategy doc!
I think some of them are human indeed, but maybe not all. AIs could be much more, if well structured. They aren't today. Which is why merging humans with AI is the best next step.
Depth is indeed valuable, but I need to change UT to reap it. Right now I touch so many topics that my depth is sporadic, and that prevents from getting enough supporters on any given topic.
In any case, each one of your thoughts made me think. Thanks again.
You missed April Fools Day with this.
Ha!
I'd invest, a crowdfunding scheme could work
Yes I am interested in investing! Also consider platforms such as microventures.
OK, I'll organize something. Investing won't be for immediately, but I'll make it happen.
This is beautiful. Thank you for your efforts, and for this terrific vision.
Wondering about some of your assumptions, eg, bias. Is it really possible to be unbiased? While the aspiration may be needed, it might be more honest and authentic to regularly self explore what biases actually are at play.
It's impossible to be unbiased, but it's possible to design a system that removes bias! Hence why I need to build tools.
Excellent point. Ongoing transparency about the biases that have impacted decisions and direction is far more useful than the pretense of objectivity.
Great insight and you have my support. Most of the mechanics of what you are proposing seem doable with people, time and resources, mainly AI. I see one one aspect that you haven't fully explained or explored here and that is the world view used to measure and push the direction of the content.
I've used and enjoyed pushing Grok and Claude to explore political events and more interesting climate change projections. The biases embraced by both AIs is difficult to shake because in many instances the news feeds they source are biased by volume alone. Getting an AI world view that is stable (won't drift) in my view is the most critical aspect of your project. It would be used as a yard stick to craft the narrative.
What do you think?
This is the 1st priority of the project!
Good luck.
If we're not careful AI will be used to roll back our ability to communicate all the way to the Neolithic. This applies now: we will be limited to face to face contact since everything else can be maliciously faked with AI. For example who still answers phone calls? But also over time. We are basically about to un-invent writing. AI is poisoning our more permanent records, like research, books. Go to Amazon and a lot of searches for books already yield AI slop. And many Google search results almost completely consist of AI slop content farms.
(Also reminder to people that, if you spend hours and hours scrolling 30 sec videos on Tik Tok you're little more than a meat bag plugged into the Matrix. So don't do that, and if you know people who do, maybe gently get them to stop)
Robots will eventually include face to face
Agreed otherwise!
An interesting element to the "content explosion" is that the people creating most of the content are still limited in their imagination on what is actually interesting. We seem to be getting "lots more of the same" rather than people wandering off to create their own unique niches in the information ecology. I think that's the trick to survival. The tools aren't coming up with good ideas (how can they, LLMs are trained on what is already there) - but they will hyper-accelerate those that are creatively different in what they approach.
True!
Vamos! 💪
In general substacks articles are far too long and in old fashion writing style. There is so much high quality content an so little time!!! We need short paragraphs almost bullet like points. Give your point of view or info in no more than 5 minutes reading time. It si not a question of attention span as time available. Axios is in my opinion how info/opinion should be transmitted. New times demand a new writing style.
Really enjoyed reading this and I agree with your diagnosis. Good luck!
I would recommend the Youtube route.
It will be a good starting point to get mone recognition, which in turn will can be levarage into merchandizing, books, paid engagements and product placements.
There are several channels that started in “niche” areas that have grown to be very successful. They found something that was very interesting to say and that could be understood by large segment of the general population.
I would check veritasium, specially this video where is shows how it grew and became monetized https://youtu.be/piHGnG4LsmQ?si=pUNwQA4Za0ThK72N. It also is a good example as how to grow with support of other people so that it does not take all of your time.
Other specialized channels that grew are
MentorPilot
Practical Engineering
Production cost could be a lot more managable. Feedback is more immidiate that writing a book.
I understand you perspective here and I absolutely want and ethical and well regulated Ai ecosystem. But as one of the many thousands of performers whose employment is about to be massively slashed, or eradicated entirely by those production companies fighting for supremacy and maximum profits for their share holders, disagree with the notion that’s these companies will be pumping out the best quality content for a long time. And when they do it still won’t have been worth the human cost. Our only hope is that the general public have the stamina to keep pushing back against producers who want to eliminate human performers from their shows. But knowing how easily swayed people are, that’s no hope at all. 😢
There will be a lot of slop.
I want to fight that.
I believe AI can help fight that, so that quality improves, doesn't shrink
The chart captures something most people feel but can’t articulate — AI doesn’t just add more content, it reshapes the entire quality distribution. The middle collapses. What’s left is either genuinely great or commoditized noise. The bet worth making is on the quality end, because that’s where human judgment still has a moat.
One small nuance: The middle gets extremely bloated. I read tons of AI articles now that are OK good. That shifts the bar though, and that level of quality then becomes the commoditized noise you discuss.