Your New App Idea Still May Suck

Happy Wednesday. Few things for today:

1. If you want to learn more about the topic, or the best way to build with AI, just reply to this email. It’s a lot to distill in one piece.

2. Sign up to this if you haven’t already. Thanks.

3. USA play Canada tomorrow night in hockey. Free Bird is now the temporary national anthem.

Alright, carry on and enjoy.

It’s Sunday afternoon and I’m sitting in a wine shop with a friend having a nice German pilsner and a pack of salami while playing with some new AI tools. Before you judge me, we have Monday off so my actions are justified. One app I tried was called bolt.new, a full-stack mobile and web app development AI tool. This same tool has gone from $0-$20m in ARR in 2 months. I know that line is insane to read and may even make you and your work feel insignificant, but that’s how I felt so I wanted to make sure I wasn’t the only one. 

Anyways, Bolt.new is by far the most impressive tool I tried. I told it to build me a twitter-like application called Burner, dedicated to allowing users to post burner-tweets completely anonymously. My prompt was exactly 62 words. Obviously, this app is a hilariously horrible idea and I would never launch it, but with just one prompt, it built this: 

62 lines of code, and it delivered this. No wonder they made $20m in two months. Being able to build an entire front-end and back-end from one specific prompt is a game changer for startups, founders, and venture capital players. 

First, any idea that someone has can be built in a matter of days as long as you have just a tiny bit of technical knowledge and understanding of how to prompt an AI. Second, I think the dynamic of early stage funding rounds has changed. Before AI, one of the most expensive parts of building a company was hiring the people to actually build it. Now, you don’t have that cost anymore.

Early Stage Funding Dynamics

I already know some of you are smashing your keyboard thinking that I just said you don’t need engineers anymore. I am not saying that at all. You think I can understand the entire codebase of a SaaS business? I’d rather take a razor scooter to the ankle. What I’m saying is that instead of hiring 3-5 engineers, you may only need 1-2 now. You need people at the end of the day who understand how to put all the pieces together, but the pieces themselves can now be built by AI.

I have a great example of this. One of my friends recently started a health tech company. He has a team of 4 with around 1-2 engineers. Most of the co-founders come from product management backgrounds and know how to build. Within the first two weeks of founding, every single person in the company was contributing to the codebase by using AI to code with the engineers reviewing the final product. 

This now eliminates a huge spending hurtle for every company at its start. Companies can get to a finished product quicker and cheaper than before. This will translate to small pre-seed rounds, or large seed rounds, but eliminate this middle “Seed” ground.

Pre-seed rounds will be users for an array of costs, and maybe an engineer or two. 

Large seed rounds will occur after a team has successfully shipped a product and is making meaningful revenue. 

The only time the typical seed round will occur is for a unique product that has different starting costs. For any typical B2C, even some B2B businesses, you will be able to build your first product with a small team and a small round. (Also I think smaller teams are the future, but that's a topic for another time).

You can build anything you want

Going back to the first point I made, now anyone can build anything they want. Alright that may be an exaggerated statement, not anyone can build anything. I doubt your grandma is going to build the next Facebook.

In order to build leveraging AI, you do need a little bit of common sense when it comes to building technology, but even as someone who can’t code, I am able to build pretty incredible products (maybe a biased take). Turns out there's a lot of people like me because the companies that are building these coding AI tools are creating some incredible revenue:

I mean 9900% revenue growth in a year is absolutely comical. 

So with virtually anyone being able to build anything, what does this mean? Initially, you are going to see more and more software products come to both the B2C and B2B market. A lot of these products are also likely to serve a really niche market. There are plenty of situations right now where people or businesses are paying for a product that isn’t exactly the right fit. Kind of like wearing shoes that are too big. They get the job done, but it’s not the most efficient and looks pretty weird. 

For example, instead of a golf resort using a software CRM to manage their guests, they can build one that makes the most sense for themselves and their situation. Yes, this is a real example, and now these resorts can build their own system for a relatively low cost if they choose to.   

Looking back toward consumer products, there is going to be a flood of new apps and websites. Some will be good, but some will suck. One of the issues with giving everyone the ability to code is the college kid who thinks he can build the next social media app can now go and build it without an engineer. And yes I was exactly that kid in college except we had to spend $1500 on an engineer who didn’t know how to build IOS apps. He’s working at Blackstone now though so hats off to him.

Anyways, the point is, just because you have an idea doesn’t mean it’s a good one. The good news is you don’t have to spend thousands of dollars to find this out now. Just a few software subscriptions and you can launch in a matter of weeks to find out this harsh reality (something I have felt many times). 

The other good news is you could have an incredible idea. All you have to do is solve a meaningful problem with an incredible product, then charge for it. It only takes 1,666 users paying $5/month to reach $100k in ARR. Surprisingly, this is harder than it looks, but if you build a great product, have some good organic marketing, you can really change your life.

That’s pretty optimistic, but it’s become more of a reality with AI. Speaking of optimism, USA plays Canada in the 4 nations championship on Thursday at 8pm. I’m not going to say it’s un-American to not watch this game, but I am also not not going to say that. 

#PlayFreeBird