3 minute read

There’s a video making waves this week in the AI rabbit hole community. If you haven’t yet taken the time to fully understand what’s become possible in the last year or so, maybe this video will be the nudge you need. I’ll summarize here, but I think it’s worth watching the video. It’s only 10 minutes.

Please Watch This

  • An 8-year old builds a legit working Harry Potter chatbot app in 45 minutes, from scratch.
  • The app connects to external APIs, and uses LLMs to chat back with you
  • She walks you through how she built it in real-time
  • She incrementally makes it more interesting and feature-rich, iterating on it and tweaking it to align with her vision.
  • Again, this is an 8 year old — building working software

So how is this happening? It’s because Large Language Models have unlocked an entirely new set of building blocks for people to make cool stuff — and these building blocks are accessible to an audience that — 2 years ago — would have experienced too much friction to get started. That friction is entirely gone. It’s never been easier to just get started.

She is using a tool called Cursor.AI, connected to a Language Model called Claude Sonnet 3.5. For about $40/month, you could be doing the exact same experimentation she is doing — and you can even experiment for free. When I say you, I really mean you — you, the one reading this.

I have so many thoughts on this, but overall — it distills down to excitement, awe, and just being incredibly inspired by what’s happening. There’s never been a better time to start building.

Here are some other reasons why I think this is exciting

  • People with problems to solve ‘in the real world’ have often been constrained by the tools accessible to them to solve these problems. Building software is now accessible to many more people. Thus, people with the true domain expertise and ability to deeply understand the problems, can now build tools to solve them. This ‘translation layer’ between domain experts and people who can build software has always caused friction. This puts more tools in the hands of domain experts.
  • The base costs for building a thing have been drastically reduced. In the past, some problems weren’t even worth solving with software, because getting started was too costly. These new technologies reduce the base costs, thus they change the cost equation related to when and why software can / should be built.
  • This opens up an entirely new way to learn. Many people are turned off from trying to build software, because ‘starting with code’ is just too overwhelming. Everyone’s brain works differently. Being able to learn in an applied way with no background knowledge necessary is incredible. You just need to be able to write clearly - that’s the first step. And this first step is accessible / achievable to so many more people.
  • It changes the skills that are valuable in the workforce. I’ve been in software for 20 years, but I’ll be the first to admit that there is a lot of unintentional (and some intentional) gate-keeping through complexity and obfuscation. I’ve always really valued communication skills. Being able to think and write clearly give you a huge advantage. Now, people who work on and build these skills can directly apply the skill themselves, instead of having to work through others.
  • The world is changing, and jobs are changing – and the major failure mode is not being willing to understand how.

I could go on and on here, but I hope you get the main point. A whole world of opportunity has been opened up for people to build cool stuff. Maybe this is you. If it is, then my only question is — did you watch the video? It’s 10 minutes, just get started.

I’m eager to hear how you or your company is trying to use these new tools. Need help getting started? Want some advice? I’d love to hear from you: mattstockton@gmail.com

Updated: