NonTrivial

AI is automating the very skills we spent years learning. So what skills actually matter now?

In this episode, I argue that we need to operate at higher levels of abstraction, moving beyond procedures and specialization toward conceptual thinking and transferable models.

I introduce three essential skills for working effectively with AI: extraction, synthesis, and direction, and explain why our education system has largely trained us to do the opposite.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.

Multitasking has a bad reputation, but constantly focusing on one thing can be far worse. In this episode, I explore why switching between tasks can improve efficiency, creativity, and depth by helping us escape unproductive paths and discover better solutions.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.

We're taught that learning requires a plan: study the fundamentals, follow a curriculum, master the prerequisites, and only then start doing. But what if that approach is actually holding us back?

In this episode of NonTrivial, I argue that the deepest learning comes from exposure. Jumping into unfamiliar situations, experimenting, making mistakes, and discovering how things work for yourself.

Using examples, I explore why following someone else's structure can produce fragile understanding, and why developing your own internal structure leads to deeper, more intuitive knowledge.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.

We're taught that compromise is a virtue. But what if it's often a failure to find a better solution?

Compromise typically means weakening two objectives until neither party gets what they really wanted.

There's another approach: find the intersection.

In this episode we explore how the molecular mechanism behind human vision reveals a surprisingly powerful approach to solving everyday problems. Using conical intersections and mathematical optimization we uncover five practical principles for finding solutions without unnecessarily sacrificing what matters.

From relationships and careers to negotiations and product development, the lesson is simple:

Don't weaken your objectives. Change the conditions until they can coexist.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.

What if the best way to build something is to start with the entire finished thing, even when you know it's going to be wrong?

Most people approach creation sequentially. Finish the first part, move to the next, and gradually work toward the final product.

But there's a fundamental problem with this approach: you can't understand how the pieces should work until you see how they interact.

In this episode of NonTrivial, I explain why your first draft should always be the entire thing, whether you're writing a book, building software, starting a business, or developing an idea.

Using concepts from machine learning, Fisher information, and sensitivity analysis, I explore how starting with a naive version of the final product helps you identify the information you actually need, rather than wasting time perfecting pieces that may not belong.

The goal isn't to get it right the first time. It's to build something complete enough to reveal why it's wrong.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.

What if trying to do everything properly is exactly why you never finish anything?

Going deep doesn't have to mean going slow. In this episode, I explain why mastering every detail from the beginning is often the wrong approach, and how working in layers lets you build deeper understanding while moving faster.

The key is simple: Get your depth through structural iteration, not sequential completion.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.

Thinking harder doesn’t always make you smarter. At some point, without new information, thought becomes rumination.

In this episode, I explain why action changes the probability distribution of your future, why overthinking is computation without new data, and why the best way to make better decisions is often to stop thinking and perturb the real world.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.

Following your passion doesn’t mean finding a job that perfectly matches it. Your passion should be the generator: the thing that builds your skills, produces ideas, and makes you unusually good at what you do.

In this episode, I explain why you should protect your passion from the market, keep creating, and let the overlap with the economy emerge naturally. Then build a commercial arm around what the market actually values.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.

For centuries, we’ve treated mathematics, logic, and analytical reasoning as some of the highest expressions of human intelligence.

Then AI came along, and started automating them first.

In this episode, I argue that this isn’t a coincidence. Mathematics and computer programming are unusually closed, symbolic, rules-based forms of thought. They can be extraordinarily powerful, but the very properties that make them precise also make them particularly amenable to machines.

Meanwhile, many of the things we traditionally dismiss as “soft” such as intuition, abstraction, analogy, deciding what matters, navigating complexity, and connecting ideas across domains, must represent a deeper kind of intelligence. And AI proves it. 

In this episode:
• Why math and coding are being automated so quickly
• Why symbolic reasoning is fundamentally different from navigating complexity
• What AI reveals about IQ and analytical intelligence
• Why the mathematician and programmer of the future may need a completely different skill set
• What humans should be doing when machines can manipulate the symbols for us

AI didn’t just automate some of our smartest work. It forces us to reconsider what “smart” meant in the first place.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.

We spend our lives chasing closure; in our careers, relationships, beliefs, identity, and definitions of success. But reality rarely fits the neat version we have in our heads.

In this episode, I explore why chasing closure creates so much frustration, why some things in life need to remain open, and how finding closure at the right level can give us something stable to navigate by.

Support the show

Members: Join for full access to additional talks, practical exercises, member Q&As, and more at dekyon.io.