I've written (with Claude doing all the work) a client application to send prompts to LM Studio. The cool thing about this example application is that it renders the LM Studio response (in markdown) as it arrives. I have to say this is pretty cool. Code available here.  Watch here.

This is a follow-up to my previous video, in which we used Claude to write a Delphi IDE plugin. I've now tested the same scenario with Gemini and ChatGPT, and this is a discussion of the results.  Watch here.

Ever accidentally hit the Insert key in the Delphi IDE and started overwriting your code? In this video, I use Claude AI to live-code a Delphi IDE plugin using the Open Tools API (OTA) that intercepts the Insert key and replace it with the context menu instead. Watch as we work through compile errors, fix a crash-on-uninstall bug, and iterate on different approaches to remapping the key — all driven by AI pair programming in a browser. Watch Here

Topics covered:

• IOTAKeyboardBinding and the OTA keyboard services
• Registering and cleanly unregistering design-time packages
• Debugging an access violation on package unload
• Using WM_CONTEXTMENU as a key remap target

I had Claude write this description, although we never fixed the "exception after uninstall" bug. Still it's quite a fun watch. I think I'll vacuum my keyboard next time before making a thumbnail.

I have no idea if anyone will be interested in seeing me fix legacy code, but let me know in the comments if you would like to see more (or never see me code again).  Watch here.

We will run Claude Code against the same legacy code we've been using for our static code analysis in previous videos and see what it identifies to fix.  Watch it on YouTube.

Summary:

• I don't think Claude is a replacement for static code analysis
• Run a code formatter and static code analysis first
• Large legacy files eat through tokens quickly
• Run Claude over the file many times with different prompts

I've uploaded the two books I've authored to NotebookLM and asked it to create a video titled "Why Static Code Analysis Matters?" I think the output is quite informative, but perhaps exaggerates the utility of code analysis. What do you think? Am I going to be replaced?  Watch on YouTube.

Continuing our deep dive into static code analysis for Delphi - this time exploring Pascal Analyzer, a powerful standalone analysis tool with decades of proven results.

In this episode, we're stepping outside the IDE to examine external static analysis tools. Pascal Analyzer has been helping developers identify code quality issues since the early days of Delphi, and it's still actively maintained and updated today. Again, we will be running Pascal Analyser against a large legacy codebase to see what it finds. We will (try to) understand the types of issues the tool detects and walk through real-world findings, and discuss practical remediation strategies.  Watch on YouTube.

We continue our exploration of Static Code Analysis plugins for the Delphi IDE. This time, we look at Pascal Analyser from Paganza https://www.peganza.com/. We will put this tool to the test on a large legacy project to see what we can find. We cover a few of the insights provided, and I discuss why this enhances the code (or not).  Watch on YouTube.

Continuing our series on static code analysis IDE plugins, this time we explore TMS FixInsight. We'll run it against a large legacy Delphi project and walk through the issues it uncovers — from minor code smells to potential bugs you'd likely miss in a manual review.  Watch on YouTube.

Delphi's "Method Toxicity Metrics" feature ranks every method in your project from most toxic to least — but what do you actually do with that information? In this video, I dig into one of my long-running projects to examine some real-world toxic code, discuss whether these metrics are genuinely useful, and explore practical ways to improve problematic methods.  Watch on YouTube.