So, you’re thinking about using LLMs to help you code?

Maybe you’re unsure about the negative impacts - on the planet, on the industry, on education, on your ability to actually code. Honestly, apart from a little fence-sitting handwringing I can’t help you much with that, but recently I decided to sell my soul to the devil and start experimenting with using LLMs to help me build software.

And you know what? It’s fascinating, and it’s fun. And it’s frustrating, and infuriating, and will NOT necessarily save you any time. But it could. Maybe. If you pay attention to what you’re doing.

I’m still on this journey myself, and there’s no end in sight, any more than there ever is when you learn new code-related things. But LLMs in particular are changing so fast that you should probably give up right now on the idea of keeping up with it all.

Some recommendations

Still, you can start playing. Here are my recommendations for how to get started.

1. Windsurf is a great tool. It seems to be free (for now). It’s a pretty rich IDE, which integrates with source control and your file system, all the while giving you a chat window to interact with a choice of four LLMs out of the box (GPT, Claude, Gemini and SWE).
2. One of the first things I did was to watch the video of Llewellyn Falco’s talk about using process files as blueprints for Agentic AI, from Craft Conference 2025. I learnt a lot from this, and it was a great starting point to get me going with LLMs.
3. Don’t trust LLMs. Don’t give them access you wouldn’t give to a stranger. Don’t let them push to source control or any external systems without your permission. Don’t assume they know what they’re doing or are following good practices. Use tests and process files and source control to keep an eye on what they’re doing.
4. Treat everything LLMs do as disposable. Throw things away frequently. Start new chat contexts. Use process files and source control to have persistence between LLMs and chat contexts (Llewellyn explains all this in his coding demo).
5. This conversation between Kent Beck and Jessica Kerr is really interesting, and if you follow Kent you’ll see a lot more from him about his experiences with augmented coding (which I believe is a term coined by him, in preference over “vibe coding”).
6. Read Co-Intelligence, by Ethan Mollick. But then read his follow-up article here.
7. Keep an open mind. Pay attention to what people are saying. I follow these people (among others). Some are cheerleaders, some are sceptical:
   1. ​Kent Beck​
   2. ​Emily Bache​
   3. ​Birgitta Boeckler​
   4. ​Gergely Orosz​
   5. ​Jessica Kerr​
   6. ​Ron Jeffries​
   7. ​Charity Majors​
   8. ​Rob Bowley​
   9. ​Liz Fong-Jones​
   10. ​Gregor Riegler​
   11. ​Dave Farley​
   12. ​Simon Wardley​
   13. ​Noah Sussmann

Come on this journey with me

I plan to keep writing about this. I’ve already got a raft of draft posts in pocket. I love to learn, and I love to teach (and I’m really bloody good at it). I use teaching as a way of deepening my own knowledge and pushing me to learn things more effectively.

If you want to know more, you can do the following:

• Read some advice on getting started with LLM-augmented coding​
• ​Subscribe to this newsletter​
• Sign up for a workshop in building robust maintainable software using genAI:
  • 25th Nov 2025: A one-day workshop at Software Architecture Gathering in Berlin
• Keep an eye on my socials (LinkedIn, BlueSky, Mastodon)
  • …and my events page…
  • …for new workshops and training sessions
• ​Read the testimonials​
  • …about the above workshops and others

🔗 Want to share this article or save it for later? Here’s a handy link for you!​