Programming with LLMs (Interview)

The Changelog: Software Development, Open Source

For the past year, David Crawshaw has intentionally sought ways to use LLMs while programming, in order to learn about them. He now regularly use LLMs while working and considers their benefits a net-positive on his productivity. David wrote down his experience, which we found both practical and insightful. Hopefully you will too!

Join the discussion

Changelog++ members get a bonus 11 minutes at the end of this episode and zero ads. Join today!

Sponsors:

• Retool – The low-code platform for developers to build internal tools — Some of the best teams out there trust Retool…Brex, Coinbase, Plaid, Doordash, LegalGenius, Amazon, Allbirds, Peloton, and so many more – the developers at these teams trust Retool as the platform to build their internal tools. Try it free at retool.com/changelog
• Augment Code – Developer AI that uses deep understanding of your large codebase and how you build software to deliver personalized code suggestions and insights. Augment provides relevant, contextualized code right in your IDE or Slack. It transforms scattered knowledge into code or answers, eliminating time spent searching docs or interrupting teammates.
• Temporal – Build invincible applications. Manage failures, network outages, flaky endpoints, long-running processes and more, ensuring your workflows never fail. Register for Replay in London, March 3-5 to break free from the status quo.

Featuring:

• David Crawshaw – Website, GitHub, Bluesky, X
• Adam Stacoviak – Website, GitHub, LinkedIn, Mastodon, X
• Jerod Santo – GitHub, LinkedIn, Mastodon, X

Show Notes:

• How I program with LLMs
• sketch.dev
• merde.ai: ai-fixed merge conflicts
• How Tailscale’s free plan stays free

Something missing or broken? PRs welcome!

Chapters

1. This week on The Changelog (00:00:00)

2. Sponsor: Retool (00:01:01)

https://retool.com/changelog

3. Start the show! (00:03:47)

4. No Tailscale AI (00:05:06)

5. !Exciting proposals (00:08:11)

6. Not a sponsor! (00:08:40)

7. Tailscale's free plan (00:10:15)

8. Boring software (00:11:08)

9. Adam's Tailscale AI (00:13:15)

10. How many models should there be? (00:16:13)

11. More on Adam's Tailscale AI (00:19:18)

12. Sponsor: Augment Code (00:25:26)

https://www.augmentcode.com

13. David's LLM journey (00:28:55)

14. Reasoning models (00:36:51)

15. Code completion (00:38:09)

16. The right model for the job (00:39:13)

17. The running shoe analogy (00:40:05)

18. Putting the shoes on (00:42:48)

19. Early days (00:43:36)

20. Building non-chat things (00:46:35)

21. Sketch.dev granularity (00:49:11)

22. How Sketch works with LLMs (00:55:38)

23. Let it ask (01:00:28)

24. Sponsor: Temporal (01:03:40)

https://temporal.io

25. Well-suited programming languages (01:05:43)

26. Swapping out Go (01:09:31)

27. LLM Engine Optimization (01:11:35)

28. LLM ads coming soon (01:16:44)

29. Getting started advice (01:20:07)

30. Any good guides? (01:22:55)

31. Massaging your AI (01:24:22)

32. Being nice to your AI? (01:25:00)

33. Closing thoughts and stuff (01:29:17)