Coding is Fun! (Again)
Joe's Nerdy Rants #72 - Weekend reads, podcasts, and other stuff
I’m not sure about you, but I go through waves where I love building things in code. Other times, I’m very burned out, and code is the last thing I want to see. I was definitely in the latter camp for a while. Especially while writing my upcoming book, my mind has been focused on more abstract concepts like data modeling. Ironically, and to the dismay of some practitioners, data modeling involves much more thinking than purely cranking out code. To the extent I’ve written code, it’s more to research and explore the contours of physical data modeling within certain types of databases and data systems.
Mid/late last year, I got the itch to build things, primarily for my new education and media business. The big difference now is I’m using AI code assistants and LLMs as my “rubber ducky.” The impact is massive in the hands of someone who knows what they want to build and understands what they’re doing. But I had a hard time putting a name on what I was feeling.
While at Google Next this week, I chatted with my friend and OG engineer/devrel legend Patrick McFadin. We compared notes on our coding workflows, and he agreed that AI was a game changer for him. What might have taken weeks or months to build can be done in hours or days. Got a crazy idea? Fire up AI and try to create it. He told me that AI’s made coding fun again.
Fun. That’s the word I’ve been looking for. Building things in code is fun again.
My typical flow for building something in the pre-AI days would look like this. Design what I want to develop, usually using pen and paper. Then, choose a tool or framework best suited for the design and requirements, often taking days or weeks to decide (and inevitably reaching for Python since I can code Python in my sleep). I start coding, write tests, fumble around and inevitably get stuck, look at stack traces, debug, look at the source code of the tool I’m using, and peruse StackOverflow for often lousy advice. Rinse and repeat—lots of effort and toil. Eventually, something gets built, but it takes a lot of effort and toil. The benefit is that going through this process teaches you a LOT about the tools you’re using, and you will have a very intimate and deep sense of what you’re building. More on this in a bit.
With AI, I’ll still draw and design things on paper. I’m old school this way. My house is littered with pens and legal pads, all with various levels of notes and scrawlings. The difference now is I’ll chat with one of the AIs (or multiple ones to get different opinions) about my ideas and design outline. Sometimes, the discussions go nowhere, but often, I get some great ideas that I’d probably never thought of. Then, I’ll have it generate some scaffold code to test out. AI isn’t perfect by any means. Sometimes, the code doesn’t work, possibly due to my prompts or maybe because it just likes to make my life miserable. Other times, the code works out of the box. Either way, the iteration cycles are much faster than before. I can test ideas in a rapid fashion, which raises my dopamine for quick wins. As I’m coding, I’ll use a co-pilot (I’m a Boomer and mainly use GitHub Co-Pilot…for now) to help tweak specific lines of code and help debug stack trace errors if I’m stuck. Again, quick wins and getting unstuck quickly means faster cycle times and iteration toward something useful.
It’s not all great, and this is where it pays to know what you’re doing. I occasionally get AI-generated code where I yell at the AI, “WTF did you just give me?!” Again, AI hallucinates, and sometimes it will make some wild suggestions. This is my issue with mindlessly vibe coding, especially in production. If I don’t understand what I’m doing and what the code is supposed to do, this can lead to problems. It puzzles me that this is controversial, but given how crazy the world is, I shouldn’t be shocked. My concern with vibe coding is there’s usually no such thing as a “prototype.” The moment your prototype works in the wild, it’s in production. This is where you need to be explicit about testing things out and exploring versus building something that might have a chance of working in the real world. We will have the equivalent of the Great Pacific Garbage Patch of AI Slopware very soon, as I’ve discussed in a few places (my podcast, Hugo Bowne Anderson’s podcast, etc). This is a talk I’m giving around the world soon, so stay tuned.
Anyway, coding is fun again. I’m writing code every day and loving it. I’m stoked about finally tackling the long list of things I’ve always wanted to build.
I’d love to hear about your experiences using AI for coding. Please post your comments.
Please listen to the audio above or on Spotify (or your podcast platform of choice).
Have a wonderful weekend,
Joe
📈 The Secret to Scalable Analytics? Think Like a Developer. 🧑💻
Collaboration, reusability, and transparency aren’t just for developers. In this white paper for GoodData, industry expert Donald Farmer shows how using Analytics as Code for versioning, modular design, and automated testing brings flexibility and scalability to analytics.
Ready to transform your BI? Download here
Sponsored by GoodData
Cool Weekend Reads & Listens
Roadmap: Data 3.0 in the Lakehouse Era - Bessemer Venture Partners
How are businesses implementing AI into their data architecture in 2025?
Announcing the Agent2Agent Protocol (A2A) - Google Developers Blog
Gemini can now turn your Google Docs into podcasts | The Verge
The evolution of graph learning
Time to Better Know The Time in PostgreSQL
The 500 Million Worker Problem - by Manish Singh
Podcast
Freestyle Fridays - Coding is Fun (Again) (Spotify)
Juhani Vanhatapio - From the Arctic Circle to AI (Spotify)
Freestyle Fridays - Shifting Left AND Right. The Data Engineering Lifecycle in 2025. - (Spotify)
Mark Freeman - Shifting Left in Data, Startup Rocket Ships, and More (Spotify)
Vaibhav Gupta - BAML and AI-First Tools (Spotify)
Freestyle Fridays - Figuring Out Your Next Move (Spotify)
Willis Nana - Navigating Data Engineering Leadership, YouTube, and More (Spotify)
Salma Bakouk - Data Observability, the Balance of Running a Startup, and More (Spotify)
Freestyle Fridays - Public Speaking Tips w/ Jordan Morrow (Spotify)
Simon Späti - The Art of Writing about Data Engineering (Spotify)
Todd Beauchene - The Early Days at Snowflake, Modern Data Platforms, and More (Spotify)
Freestyle Friday - How I Use AI for Writing (Spotify)
Matthew Kelliher-Gibson - The Data Cynic (Spotify)
Carly Taylor - The True Cost of Replacing Engineers with AI (Spotify)
Freestyle Friday - The Cult of Scrum (Spotify)
There are way more episodes over at the Joe Reis Show, available on Spotify, Apple Podcasts, or wherever you get your podcasts. Also available on YouTube.
Upcoming Events
AI-Ready Data, Data Products & Governance w/ Data Galaxy - April 16. Register here
Data Quality Day w/ Monte Carlo - April 24. Register here
Matillion’s Deep Dish Data w/ Carly Taylor and Mark Balkenende - April 29. Register here
SLC Low Key Happy Hour - May 9. Register here
Current London - May 20-21. Register here
Snowflake and/or Databricks - June TBA
Iceland - Global Data Summit, June 23-24. Register here
Australia (Sydney, Melbourne) - Data Eng Bytes, July 24-30. Register here
UK - Big Data London, September 24-25. Register here
Helsinki Data Week - October TBA
More to be announced soon…
Thanks! If you want to support this newsletter
The Data Engineering Professional Certificate is one of the most popular courses on Coursera! Learn practical data engineering with lots of challenging hands-on examples. Shoutout to the fantastic people at Deeplearning.ai and AWS, who helped make this a reality over the last year. Enroll here.
Practical Data Modeling. Great discussions about data modeling with data practitioners. This is also where early drafts of my new data modeling book will be published.
Fundamentals of Data Engineering by Matt Housley and I, available at Amazon, O’Reilly, and wherever you get your books.
The Data Therapy Session calendar is posted here. It’s an incredible group where you can share your experiences with data - good and bad - in a judgment-free place with other data professionals. If you’re interested in regularly attending, add it to your calendar.
My other show is The Joe Reis Show (Spotify and wherever you get your podcasts). I interview guests on it, and it’s unscripted, always fun, and free of shilling.
Would you like me to speak at your event? Please submit a speaking request if you want me to give a workshop or talk at your event.
If you’d like to sponsor my newsletter or podcast, please get in touch with me.
Could you be sure to leave a lovely review if you like the content?
Thanks!
Joe Reis
Been vibe coding a bit this weekend. It’s a ton of fun!
Highly resonate with your thoughts Joe. Most of the fun with vibe coding lies in how well we can prompt the tool to get the desired output.