Ryan Wakefield

This week on Troubleshooting Agile with @jtf and I discover how the simple yet powerful technique of walking the board can reshape executive accountability, improve focus on strategic initiatives, and foster team cohesion. buff.ly/Er8LktJ

@WyzeCam I just wish your plugs were right-angle plugs where the cord doesn't point straight out, but up/down. I keep having to buy third-party ones so I can plug them in behind furniture and all.

Have access to a window? Then you can use Wyze Window Cam. Easiest installation yet.

@burkeholland I've also been learning to start building my prompts as ordered or unordered lists (depending on the situation). LLMs do great with structure, so why not leverage that with lists.

When you prompt, stop cramming so much information in. This confuses the model. The better way is to use plan mode and be intentionally vague - “I want to port this app to JavaScript”. Let plan mode prompt you for the details. It will assemble a coherent plan. Then for the love of god review the plan. I know that takes time. I know it’s boring. I know you are cooking at 9k miles per hour and you need the model to just do the thing. But there is no beating a well thought out, comprehensive plan. Period. You will spend more time typing “it still doesn’t work please fix it” in the long run and burn WAY more tokens. Slow down. Plan. Review. Execute. It has never been more important for you to move deliberately than it is right now. I cannot stress this enough. This will be the difference between success and “it almost works”.

@HKilset @JamesMontemagno @pierceboggan As well, making it so that you can still view all of the files in a solution. For example, a repo I have doesn't have a .sln file. So because I am missing that, I can't view all of the files in the file explorer (that I have found a way to do so yet).

@JamesMontemagno @pierceboggan A refresh of the Explorer in particular would be great. Coming from Visual Studio it's difficult to read and parse especially large code bases in the VS Code Explorer. Goes for the Solution Explorer with the c# extension too.

@pierceboggan can we ship this?

Finally decided to start a YouTube channel! Would greatly appreciate if you went and checked it out 🫶 First video: how to fully customize Windows Terminal with Oh My Posh (and more!) Also let me know if you want to see any specific videos from me! 💖 youtu.be/VFnr0lwwGuQ?si…

@JamesMontemagno Does GitHub mobile chats sync anywhere? I would love to be able to start working on something on my mobile, and then finish when I get back to my desktop.

New favorite thing to do is plan out features in the GitHub mobile app and immediately have it delegate to the GitHub Copilot coding agent.

Why is it called daily standup if you're not supposed to start it by telling jokes?

@intellectronica @code @GitHubCopilot Good to know. Just wish the good Codex ones could be 0x, but at least I now know I can use GPT-5-mini. That will be nice.

@rlwakefield2 @code @GitHubCopilot you can use gpt-5-mini for many simple actions. I use smaller models like gpt-5-mini (0x) or gemini flash / claude haiku (0.33x) for much of my work, saving the big and expensive models for when it really matters.

Finally got GPT-5.3-codex in @code / @GitHubCopilot and it is 🔥

@pierceboggan @code Oh, I love that Context Window view. Never knew that was there before. I will have to check this out next time I have time. Lol

New in @code Insiders: Manually compact a Chat conversation with /compact. For more control over what is retained after compaction, add it to the end of /compact. Alternatively, compact directly from the context window widget.

Oh this looks like fun! tinytooltown.com cc @shanselman

🔥 @shanselman just created tinytooltown.com (Textream is there too, because it is tiny too)

Have you vibed or made a tiny tool? Something stupid and delightful just for yourself? Join us at tinytooltown.com

@shanselman I saw this original post and this has me so excited. Super cool idea!

Geocities for vibed apps tinytooltown.com

@AhmedHamdy29189 @pierceboggan @code @IsidorN Yeah, that could be really cool. Make it so that you can leverage each model for their strengths.

@rlwakefield2 @pierceboggan @code @IsidorN Or a setting per codebase or vscode workspace.

The model picker experience in @code is long overdue for an overhaul. ... so we're totally refreshing the model picker! - Search - Pin/favorite models - Hide dated models - See model details, like context window - Control reasoning effort What else would you like to see?

@pierceboggan @code I know this isn't with the picker directly, but Iwould really like one of the GPT5 models to be available in the 0x token section. Is there a reason that the old GPT4 and GPT4.1 models haven't been retired so that the newer and better models can be used?

@AhmedHamdy29189 @pierceboggan @code @IsidorN Almost like associating what models you want to use with what file formats? Might be hard to do, but possible.

Examples given current available model lineup in copilot (a) If i am working in a codebase that does analytics on blockchain data and utilizes binary combinatorics (i.e. code is heavy in binary representation) then I disable Sonnet. (B) If i am working in a codebase that is UX heavy then I enable Opus/Sonnet/GPT 5.2 (high) and disable Codex. (C) IF am going from Whitepaper math equations to code i enbale Opus , GPT 5.2 (High) disable Sonnet and Codex.

A small experiment for visualizing @GitHubCopilot agent trajectory in @code Would u find this useful? youtu.be/qj1j-93YZrA?si…

@AhmedHamdy29189 @OrenMe @GitHubCopilot @code Not only just what went wrong, but figuring out how exactly the agent came to the conclusion it did. And knowing where to tweak instructions to make your agent as accurate as it can be.

@OrenMe @GitHubCopilot @code Woww this has many uses for exampl in post-mortum debugging of what went wrong during a chat thread .. great job Oren