Blogging in the Age of AI

Here, have some AI art that took two seconds to generate that’s better than any art I’ve ever made in my life.

For those of you who are unaware, https://ai-2027.com/ has come out, collecting some of the best forecasters and thinkers about AI to analyze where we are and what the next few years look like.

To be blunt, it doesn’t look good, assuming one is in favor of the human species.

The forecast, relying only on the assumption that the current rate of progress continues, forecasts that human-level AI will be reached in 2027. This means an AI that can do functionally any online task that a human could, a drop-in replacement for any remote worker at a tenth of the cost, is only two years away.

Then, of course, the major AI labs that have that technology turn it to creating an even better AI, and then turn that AI to creating an even better AI, and-

Well, that’s the singularity.

Not a science fiction novel, just the logical end state of Moore’s Law and other scaling effects that produce exponential increases in the amount of intelligence being deployed.

While I recommend reading the whole page (even though it is quite long), this post isn’t exactly about that. It’s about living (and writing) in the world in the meantime.

Blogging in the Age of AI

I’m not an expert in most fields. No one is. There are too many fields, and those fields are too deep, for any human to master more than a few over the course of their lifetime.

Not so AI, which has functionally digested the entire internet by this point.

One of the ways I like to imagine current AI is that it is a college senior - or, shortly if not already - a graduate student, in every field of study humanity has ever had.

Think about that for a second.

I’ve been posting once a week for almost two years now. Each post usually takes me a couple of hours to write, unless it’s a longer effort post. And I’m writing as a human whose knowledge is probably broader than most but still so incredibly limited.

The thought occurred to me recently that I could have AI write posts for me. I won’t, because that’s not the point of this blog and it’s not the reason I’m writing, but I could.

I could post a graduate-level analysis of any topic in existence multiple times per day, using AI.

Would that get me more readers? Would it provide more value to my existing readers?

I don’t know. Maybe.

But that’s not why I’m doing this.

So I had to think to myself: why am I doing this? What value can I possibly provide, at this point, that an AI cannot? And then how do I shape my substack around it? Because if I’m not doing anything an AI can’t, why even bother?

The Advantage of Being Human

The answer, when it came to me, seemed obvious in retrospect.

What can I do that an AI can’t?

Go be a human out in the world.

…that’s about it, in the long term. The long term meaning ‘the next two years’.

So that’s what this blog will pivot to being about. I think I’m going to focus a bit less on concepts and terminology and takes on economics, and a bit more on real-life experiences and what I’ve learned and experienced from them. It’ll take a little while to fully execute the pivot, because I’ve got a backlog, but hopefully that’ll be better for me and for you.

At the very least it won’t be something you can get from an AI for the next five minutes.