8 years of looking up

Eight years.

I started this blog in February 2018 because two rockets landed at the same time and I couldn’t stop watching the replay. I wrote about it from my apartment, late at night, with the kind of breathless wonder that you have when you’re seeing something for the first time and you know it matters but you can’t fully explain why.

Eight years later, I’m still on the roof. Still looking up. Still trying to explain why it matters.

What’s changed

In 2018, AI was GPT-2, which could write a convincing paragraph and nothing more. Now AI reasons through problems, writes code, sees images, holds a million tokens in memory, and occasionally says things that make me forget I’m talking to a machine.

In 2018, the most advanced robot was Atlas doing a backflip in a controlled lab. Now humanoid robots work in factories, learn tasks from YouTube, and cost less than a car.

In 2018, Starship was a stainless steel prototype that looked like a water tower. Now it launches routinely, catches itself on mechanical arms, and has tested orbital refueling.

In 2018, self-driving was a novelty in one city. Now it’s a profitable business in eight cities with half a million rides per week.

In 2018, the most advanced chips were at 7nm. Now they’re at 2nm. Ten atoms wide. The physical limit of silicon is visible.

In 2018, JWST hadn’t launched. Now it’s operational a million miles away, and it might have found biosignatures in an exoplanet atmosphere.

What hasn’t changed

The roof. The stars. Orion in winter. The sense of being small in a vast universe that doesn’t care about us but somehow produced us.

My wonder. I worried, somewhere around year three or four, that I’d get used to it. That the progress would become normal. That I’d stop feeling the thing I feel when a rocket lands or a robot learns something new.

I haven’t. The wonder is intact. If anything, it’s deeper, because I understand more now. I know how ASML’s EUV machines work. I know why TSMC’s yields matter. I know what mixture-of-experts architecture means and why orbital refueling changes the Mars equation. The more I understand, the more astonishing it becomes.

Ignorance produces awe. Understanding produces deeper awe. That’s been the surprise of eight years of writing this blog.

What I’ve learned about predictions

I’m wrong as often as I’m right. The things I’m most confident about are the things I’m most wrong about. The future punishes certainty and rewards curiosity.

I’ve learned to say “I don’t know” more often. Not as a hedge. As an honest assessment. The future is genuinely uncertain. Anyone who tells you they know how it goes is selling something.

What I’ve learned about writing

The short posts work better than the long ones. Not always. But the posts that resonate most are the ones that capture a single feeling. The thought about latency. The question about Mars. The parking lot where no one parks anymore.

People don’t come to this blog for analysis. They come for the feeling. The sense that someone else is sitting on a roof, looking up, wondering the same things they wonder.

I’ve learned to trust the wondering. To publish the unfinished thoughts. To ask questions I can’t answer. That vulnerability is what makes this blog mine and not content.

Eight more years

I’ll be here. On the roof. Watching.

The next eight years will be more dramatic than the last eight. I’m sure of that. The convergence of AI, robotics, space technology, and semiconductor engineering is accelerating. The threads are weaving into something I can’t fully see yet.

But I’ll be watching. And writing about what I see. And wondering, as always, what it means.

The stars don’t answer questions. They just present them. Eight years ago, that felt like a limitation. Now it feels like enough.

Related thinking: