Arun T S

@ethanhunt2026 @poserarcher Twitter (X) is adding dislike feature just for tweets like this.

@poserarcher T20 cricket is meant for high scores. If we want testing conditions let them play test match and survive.

The ball moves a little and look at these flat track bullies batters crumble.

@paraschopra Completely agree man. It's tiring to read at this point. Either a bot or a human not willing to spend a ounce of his energy to think

More than half of the replies on my tweets is from bots who simply rephrase what I just tweeted. What's the point of it? It's really sad.

@rcarmo @badlogicgames Out of curiosity before I trying this out. In agents md i have mentioned use powershell commands as you in Windows environment to pi agent. Why is an extension needed over the above? Apologies for my question before trying out rhe extension

People of pi, I beg your mercy, but I got pi to run under bun in Windows and speak PowerShell, because it's Friday, and... no good reason, really /cc @badlogicgames gist.github.com/rcarmo/369b26f…

@skill_evolve @mgechev Why do we have this kind AI slop tweet replies everywhere? @nikitabier

@mgechev This. The difference between a skill that works once and a skill that works at scale. Regression testing = the moat that separates production systems from demos.

skillgrade allows you to evaluate your agent skills and keep them from regressing over time Just shipped a couple of more versions over the past few days. Give it a try! github.com/mgechev/skillg…

@martinwoodward Wow. I thought co-pilot cli and chat uses different harness. When do you cli and when do you chat? Just picking your brain

Being able to seamlessly pick up GitHub Copilot CLI sessions in VS Code and bounce between the two just feels like magic. One of those 'well of course it works like that' experiences that are so incredibly hard to pull off in how invisibly it all just works.

@vivekgalatage Got it. Thanks for the info

@actuallyTSA These are not really text books or a course work to finish start to end. These rather are references, where you can come back to them anytime you face some issues across your implementations. There are some great case studies covered that makes these resources valuable.

The Little Book of Semaphores by Allen B. Downey greenteapress.com/wp/semaphores/

@AkashSharm44677 @arpit_bhayani Are you a bot? Or are you just trying to contribute something but don't want to spend your brain cycles and using AI to do the thing? Why?

@arpit_bhayani the insight is: correlated logs > scattered logs when every line from one request shares a trace ID, debugging goes from "grep and pray" to surgical precision Stripe-level observability is what separates good infra from great infra

Fun Fact: Stripe does logging differently. They take a super interesting approach to log processing, something that helps them debug production issues faster, cheaper, and with far less complexity. Logging data belonging to one request is typically dumped across multiple lines. Stitching together these fragmented logs became a pain as they scaled. So what they do is emit a single, comprehensive log line at the end of every request, and they call it the Canonical Log Line. I just published a video dissecting their entire process and the exact steps they took to implement canonical log lines without breaking their existing log processing systems. Give it a watch. It is a pretty short video, but interesting and fun. Something you can implement at your workplace right away.

@martinwoodward Does it work without cloud agents enabled? '/fleet' uses local background sub agents?

Folks are sleeping on /fleet. I regularly use it, sometimes getting different models to work on different parts in parallel. No need to serialize certain things when you can run in parallel, sprinkle some work tree magic and get the results back quicker.

@nicopreme Hi Nico, have you gone through this LCM stuff - github.com/Martian-Engine… ? It would be great if we can add a pi extension for this. Without breaking token caching.

pi-mcp-adapter lets you install 100+ MCP servers without any token or resource bloat. Just added MCP UI support with bidirectional messaging and streaming. pi install npm:pi-mcp-adapter github.com/nicobailon/pi-…

@DOMZE @davidfowl I am working in enterprise. I don't have permission to create another organisation

@actuallyTSA @davidfowl So create another organization that you are assigned to for Copilot and let that organization decide the features it can enable. Like this you can have an innovation org where you test and valide your governance while you keep the other org as is

I worked on the copilot cli today and they move so fast my PR had merge conflicts every hour for the last couple of hours 😅

@davidfowl Our GitHub enterprise admin has not enabled it for us since the GitHub copilot cli says `Preview` on the copilot settings page.

@actuallyTSA Aren't there stable builds you can use?

@_Evan_Boyle This is awesome. I badly want to use copilot cli. Can you help? x.com/i/status/20258…

An engineer on the team forwarded this to me. Copilot using Chronicle to lookup some colloquial terms it doesn't recognize. This is true memory.

@burkeholland @awakecoding @_Evan_Boyle Why am I not getting a response to this genuine request? x.com/i/status/20257…

@awakecoding Hmmm I wonder if autopilot in the CLI is what you are looking for here. @_Evan_Boyle

I wish GitHub Copilot in VSCode had a way to automatically try again when hitting errors, or to switch to a different model from a list of preselected models when hitting rate limits. Cursor doesn't mind, but charges $$$ for running long hours. I just need VSCode to not stop

@davidfowl I want to use GitHub copilot Cli badly

@davidfowl We are in enterprise setup. Our admin only enables stuff that is in GA. Copilot CLI is in preview for a long time. I am really excited to use it - ability to invoke multiple models in a single prompt. Can GitHub copilot team share an ETA of when we will have GA of co-pilot cli?

Since December my primary coding development has shifted back to the CLI. I primarily use copilot CLI for my day to day development, I still have a claude max plan and I occasionally use claude code as well. I use VS code to look at diffs of the code but haven't written much by hand in ~1.5 months. I used big VS the other day to do some advanced multi threaded debugging, but I rarely open it now. PS: I have unlimited tokens.

@github We are in enterprise setup. Our admin only enables stuff that is in GA. Copilot CLI is in preview for a long time. I am really excited to use it - ability to invoke multiple models in a single prompt. Can GitHub copilot team share an ETA of when we will have GA of co-pilot cli?

Cool: Building a 60fps animated ASCII banner for GitHub Copilot CLI. 😎 Even cooler: Rendering high-res animation in a terminal without breaking performance. 😎😎 Here's how we did it. 🛠️ github.blog/engineering/fr…

@github We are in enterprise setup. Our admin only enables stuff that is in GA. Copilot CLI is in preview for a long time. I am really excited to use it - ability to invoke multiple models in a single prompt. Can GitHub copilot team share an ETA of when we will have GA of co-pilot cli?

Copilot now has memory! 🧠 Any of your favorite agents on the GitHub platform (Copilot, Claude, Codex, c-c-c-combo breaker) can learn across your repo. Now in VS Code Insiders, Copilot CLI, Copilot coding agent, and Copilot code review.

@_Evan_Boyle Any update on this? Cc: @pierceboggan @code

@actuallyTSA Coming soon

Afternoon hack: all sessions embedded and available in a search index and exposed as a tool. Do you want Copilot CLI to give you reports about how you are working, patterns, what to improve, and most importantly wins?