It's Time to Chop Trees With a Blunt Axe

How an accelerating technology curve changes the metagame of knowledgework

“Give me six hours to chop down a tree and I will spend the first four sharpening the axe.” - Abraham Lincoln

Does it make sense to spend hours sharpening our axe when the chainsaw will be invented tomorrow?

Maybe it doesn’t.

There’s a threshold on the technology acceleration curve where new technologies are released faster than we can master prior technology. There were nearly 8000 years between the invention of the axe and the the chainsaw. There were 4 months between the release of ChatGPT and GPT4!

Mastery Dividend Period

In other words, axe mastery paid dividends for up to 8000 years: over 100 human lifetimes! Optimization and mastery have been no-brainer investments for most of human history.

But we’ve crossed a threshold where the metagame has fundamentally changed. Mastery of tools no longer pays dividends for the rest of our lifetime—we’re lucky if a new AI tool stays relevant for 12 months! It’s incredibly disorienting when the “sharpen the axe” strategy that worked for thousands of years suddenly stops working.

There are two overwhelming sentiments on AI Twitter:

1. The Axe Sharpeners: I’m overwhelmed! How is anyone supposed to stay on top of the firehose of AI advancements? I’ll stay on the sidelines for now.

2. The Blunt Axe Choppers: I have no idea what I’m doing, but look at this cool thing I built in an afternoon 🤷

There is an anxious sense of paralysis amongst the Sharpeners who are used to finding victory via the dividends of skill mastery: Will the skills I learn today be relevant tomorrow? Why build something today when better technology will come out tomorrow? Why invest in an AI company today when everything will fundamentally change tomorrow?

Many companies that invested deeply in building ML expertise before LLMs are now struggling to compete with blunt axe choppers who are outperforming their proprietary models using GPT4. They’re desperately marketing why their fully custom approach is better, but customers can tell that the results of GPT4-powered products are often better than those that came from application-specific models or fully custom LLMs.

A Golden Opportunity

A shift in the metagame presents an enormous opportunity to generate alpha—as players grapple with a completely new set of rules. If you can adapt to this reality, you can generate huge excess returns.

I’m betting that the optimal strategy is to:

1. Dive in and get comfortable with the feeling of using tools you haven’t mastered

2. Treat the dynamic trading of tools as a the #1 skill for teams to master: Pick up new tools as soon as they are released, leaving old tools behind. Don’t get attached to the old romantic idea of mastering a single tool. Get comfortable wielding tools you don’t 100% understand.

3. Stay focused on mastering things that are stable (eg. marketing, positioning, UX, orchestration, infrastructure & data) around the rapidly shifting software foundation.

Startups are designed to succeed against established competitors in times of disruption. Small teams are able to be live players that adapt first to a new environment and write the playbooks for the decades to come.