Walking and talking with AI in the woods

In my past several years of working with AI, my greatest unlock has come down to a form factor: specifically, Bluetooth headphones (with a mic).

It's not been the many images I've generated or the assistance I've gotten planning architectures and building apps - despite doing so daily.

The single greatest leverage I've gotten from AI comes from being able to: learn, work through, consider and clarify while away from my desk.

Learning: now coming to you, anywhere you want

I enjoyed a deep dive on JavaScript engines, their API boundaries, which companies sponsor their development and exactly how they work while blowing leaves in the backyard.

My favorite: A long walk with advanced voice mode to digest, debug, defrag

My favorite thing to do is to take a long hike along the river and through the trails up into the woods, ideally while chewing through a business problem, client project, the next evolution of my site.

I start these walks with my mind occupied by various competing priorities and interests.

Initially, I like to dump everything that's on my mind and vent a bit and then use the LLM as a sounding board to help organize my thoughts, prioritize and tackle items accordingly.

Meanwhile, I'm logging miles on foot and getting some fresh air.

OpenAI’s Advanced Voice Mode for ChatGPT currently offers the best experience here.

I’ll close the screen of my Pixel phone, and carry on chatting hands-free. The biggest issue I have now is that OpenAI's Advanced Voice Mode is currently limited to one hour unless you pay $200 a month.

Flagship LLMs are ideal for this kind of work because they have read most works, speak most languages and know most technical concepts and architectures.

I have an acute sense of what I understand fully and what remains murky, so I can constantly ask questions to expand my knowledge and reduce the "fog of war" around certain concepts.

I solidified my knowledge of neural networks, their architectures and how they actually look on disk over the course of two weeks of walks.

The result? I was able to sit down and build my first neural-network backed application quickly and easily.

As I continue my trek, I'll sort out the order of priorities to tackle, and I'll continue talking through plans to improve or streamline a system or ask for critical feedback on a campaign or conversion funnel I'm building.

Walking itself helps boost creativity and reduces mental clutter—an ideal combination for diving into technical topics. Pairing that physical flow with voice-based AI conversations means I can dig into the details of ML architectures, discuss software engineering patterns, and refine new project ideas in real-time without being tethered to a screen.

Ultimately, I'll arrive at the summit of my walk: bleachers overlooking the surrounding area. I climb these and pace back and forth for a while, developing clarity.

At the end, I'm more focused, relaxed and clear on my exact next steps.

A powerful habit I expect to spread

The feedback loop of conversation with a model that has essentially read everything can elevate an ordinary walk into an inspiring learning session.

By the time I return to the parking lot, I often already have solutions I’ve been chasing for days.

I suspect that this habit will spread as more and more folks find it helpful, and as a result I expect we'll see additional roducts and services tailored to the untethered crowd: those now preferring to do as much of their work as they can away from their desks.