Kenny Pflug

Forget Vibe Coding, do Guided Coding! 1) Plan iteratively with the Coding Agent 2) Let the Coding Agent implement the plan 3) Review and iteratively guide the Coding Agent towards better design, security, performance, etc. #AI #CodingAgents #GuidedCoding #VibeCoding

The performance benefits are pretty good: for simple flat DTOs, the execution time and memory allocations are roughly 25% less than FluentValidation. For complex DTOs, both numbers go up to over 32%.

6800+ lines of new code, including hundres of new tests and dozens of additional benchmarks in about 14 hours of work in my spare time. This is the benefit of AI — if I had written this all by myself, it would have taken me up to five times as long. github.com/feO2x/Light.Po…

In the past days, I used GPT-5.4 to add the foundations for a major new feature to my new library Light.PortableResults. With #GuidedCoding, I implemented validation support with one plan and then optimized the code with four separate plans. #AI #CodingAgents #csharp #dotnet

@mark_heath I'm very happy with the ones from RME Audio. Personally, I have been using a Babyface Pro for 7 years, but if you need more analog inputs, you need to switch to a larger interface.

Am finally going to abandon my Focusrite Scarlett 18i8 after years of battling audio driver issues on Windows. What 8-input USB audio interface has reliable Windows drivers? Considering Audient / UA Volt etc

Long time, no see, posting here several projects I worked on in the past months, as they got zero (re)post here: 😅 First, a new terminal UI framework xenoatom.github.io/terminal/

@RogerAlsing Nice! Do you have a link to a GitHub repo?

I´ve built my own web PDF renderer. Now trying to make white PDFs look decent in dark-mode. Is my "Night mode" yay or nay? does it look too weird?

📊 Around 45.000 lines of code were created, more than 90% of them written by Agentic Coding tools like GitHub Copilot, Windsurf, and Codex App, using a mixture of the models Codex-5.2, Codex-5.3, Opus 4.5, and Opus 4.6. It was a great learning experience.

🏁 I started the project at the beginning of January and used Guided Coding extensively. If you want to know what Guided Coding is: join Johannes Rabauer an me tomorrow evening at 8 PM CET: linkedin.com/events/7431253… #AI #CodingAgents #Java #dotnet #csharp

🚀 The first version of Light.PortableResults is out now! It implements the Result Pattern, but with a twist: every result instance is serializable and deserializable. You can easily map results to any protocol format. #dotnet #csharp github.com/feO2x/Light.Po…

@wieslawsoltes Give this man a medal! I've been following your progress over several years and you never cease to amaze me. What you have built with Avalonia is genius! Please keep up the good work, you are an inspiration to every software dev.

This is just beautiful progress from my new Xaml source generator based compiler for Avalonia. Entire Fluent Theme and sample app are now compiled using source generator compiler and it includes full Hot Reload support.

Live session: Guided Coding in Java with @feO2x Not vibe coding. Not blind prompting. We explore a structured way to collaborate with AI while keeping architectural control and code quality high. youtube.com/live/vopBYXp9Y…

Have used it a little bit more. Man, this model is bad at agentic coding tasks. It takes forever and it uses tools weirdly (duplicate tool calls, so many wrong tool calls, and after that, it simply writing python code to circumvent). Anyone else having these issues?

What's your opinion on #Gemini 3.1 Pro? I think its ultra-slow in GitHub Copilot - not happy at all. What's your experience with the model? #AI #GitHub #Copilot #Google #dotnet

How to secure AI agents with Zero Trust! A really helpful overview video from the guys over at IBM. #AI #AiAgents #IBM #ZeroTrust youtu.be/d8d9EZHU7fw?si…

You own the code that your LLM produces! Invest in your knowledge, you guide the LLM. Be clear in your communication, know the limits of the Coding Agent (especially its Context Window size). Improve on the way: adapt your agent rules when you face issues.

3) Review the code and iterate on the result, together with the Coding Agent. Question design choices, don't be afraid to change things. Document deviations in another md file and commit it to the repo. This typically takes 45 minutes to 6 hours in my experince.

Forget Vibe Coding, do Guided Coding! 1) Plan iteratively with the Coding Agent 2) Let the Coding Agent implement the plan 3) Review and iteratively guide the Coding Agent towards better design, security, performance, etc. #AI #CodingAgents #GuidedCoding #VibeCoding