Practical Ways the Top 1% are Using AI

Have You Seen a Measurable Impact using AI?

Sam Parr just released his AI Founder Report and boy is it a doozy.

The key findings are jaw dropping and only going to get more so. Here’s the content without signing up through Hampton.

This is not a doom column and I have zero concern AI is going to take millions of jobs tomorrow. But what I do know is:

1. AI needs to be a piece of your business

2. The companies who implement AI are going to see more success than ones not implementing

3. Someone with elite AI knowledge will grow in a non-linear fashion

The last point is important. Most small business owners cannot code and aren’t spending time learning AI.

Companies succeed by reaching prospects, nurturing leads, and converting those leads to paying customers. Domain knowledge and fulfillment capabilites matter. If you build an incredible sales campaign for a plumbing company in Omaha, they are limited by the number of jobs they can take along with the labor in the nearby area. Success is a chicken and egg problem, where owners fluctuate between supply and demand constraints.

However, the best operators are not working in the business. AI will only make life easier on demand. For physical businesses, expect to spend far more time balancing supply side factors (labor, real estate, travel, etc).

Owners Should Listen to YCombinator

YCombinator has this theory that founders should focus on tasks that don’t scale. When I first heard that I was surprised. Wouldn’t your business grow if you spend most time on things that scale?

The reality is things that scale mean they are reproducible, and thus can be easily learned.

Take the most reproducible task possible, say, a shelf stocker at a grocery store. It’s so simple that training someone would take a few minutes. Maybe there are some edge cases where 1% of shelf stocking needs explanation but overall, you can pay someone a minimum wage and let them work.

If so, you can spend 1% on those edge cases and focus the rest of your day on again, very difficult things that aren’t reproducible (and outsourceable).

Most of the time, this is considered ‘value-adding’ tasks. Things that directly make the business money. But in 2024, things that make you money (marketing, digital campaigns, cold outreach) are mostly reproducible.

Paul Graham categorizes non-reproducible tasks well, I highly recommend you read more.

Back to AI and the Replication Process:

There are hundreds of roles within a grocery store that are more complex than shelf stocking. As an owner, your job should be matching highly reproducible work with solutions (either humans or tech).

This is where AI matters. In the past, you had to hire humans for varying levels of reproducible tasks. On the low end, you have things like call centers. On the higher end: lawyers, doctors, and engineers.

Both groups had to deal with some form of outsourced labor, though the higher end group was mostly safe due to risk/reward.

Later this week I’ll share low-wage companies are more in danger of job loss than in the US.

But with AI, the general value proposition is to take unstructured data and turn it to ‘structured data.’ 

What does that mean? Something that’s structured is easily reproducible.

Like if I said, ‘here’s a shelf, if it’s empty: check if we have more of that item and then proceed to replacing it.” It’s very simple and is structured in a way that I can guide you based on single command. But the only reason we have humans doing it is because AI is constrained to a box. If robots could walk and pick things up, the job would be replaced.

Unstructured data is harder and thus didn’t scale well for technology. Things like medical care are incredibly complex and require a very thorough understanding of the patient’s problems and available solutions.

But today, you can largely enter in your blood work and what hurts before AI gives you a solid response. I actually did this with my father-in-law and correctly diagnosed him. His doctor, who is a professor at Harvard Medical School, didn’t and he wasted months of his life!

Back to Your Business

To sum this all up, it’s imperative that you start structuring your workloads. Everything needs systems because systems are reproducible. As the puppet master of these systems, your goal should be to tackle hard things once and then see what percent isn’t within a low cost human or tech solution.

In my final words, I’ll leave you with this quote from Anduril’s Palmer Luckey. Having fun is not where the money is - go towards the suck and you shall prosper.