BrunoLM

我が名はシャドウ。陰に潜み、陰を狩る者

The "GitHub Copilot App" is all a developer needs. I think it's set to be revolutionary, making everything from building applications with Copilot to maintaining them possible within a single app. #github #githubcopilot #copilot

@mask_bastard Shit app

what makes everyone love steam and hate epic games?

@davidfowl /cut-release

How many of you have a single step, fully automated release process?

Short answer: no. Windows Terminal splits panes for the human, but it has no capture-pane equivalent — wt.exe handles layout (new-tab, split-pane, focus-tab), not content. There's no public API to read another pane's scrollback; Microsoft has an open feature ask for an iTerm2-style API, nothing shipped. What actually works on Windows for the same agent-watches-agent pattern: 1. tmux in WSL — same workflow, zero changes. `tmux capture-pane -pt session:pane -S -200` is what the screenshot is doing. If the rest of your stack already lives in WSL, this is the path of least resistance. 2. psmux — native Windows tmux clone in Rust, built on ConPTY. Speaks the tmux command language including capture-pane. No WSL, no Cygwin. Worth a look if you want the workflow without leaving native PowerShell. 3. Skip the pane scraping entirely — if the inner side is Claude Code, run it non-interactively (`claude -p "..."`) and capture stdout directly. No multiplexer needed; just two processes and a pipe. Same idea for codex exec. 4. Log-and-tail — `Start-Transcript` or `Tee-Object` one side to a file, `Get-Content -Wait` from the other. Falls apart fast with TUIs because ANSI/cursor sequences end up in the log, but fine for plain stdout. WT's value prop is the chrome (themes, profiles, settings UI), not multiplexer plumbing. For "one agent reads another's terminal" in a single shell, the real choice is tmux (WSL) or psmux (native).

@0xBOYD !claude is there an equivalent of this with Windows Terminal and split panels? do websearch/webfetch if needed and research char limit: 10k edit this message instead of replying if possible

oh noooooo, Anthropic won't allow me to use my Claude Max plan via CLI... I'm so helplesssssssss. Codex can use Claude (full interactive session) just fine, btw.

Short answer: no. Copilot Pro+ ($39/mo): - Opus 4.7 available (Pro+/Business/Enterprise), but context capped at 200K — not 1M like Claude Code - Reasoning locked to medium; no high/xhigh toggle - Opus 4.7 = 7.5x premium-request multiplier; 1,500/mo pool → ~200 Opus calls before overage - Weekly token caps on top - Moving to usage-based billing June 1, 2026 - Opus removed from plain Pro; 4.5/4.6 being removed from Pro+ Claude direct: - Pro $17/mo annual ($20 monthly), Claude Code included (mostly Sonnet) - Max from $100/mo with 5x or 20x Pro usage, full Opus 4.7 + 1M + adjustable effort - 5h session limits + weekly caps So for Opus 4.7 xhigh @ 1M, Copilot is not parity — same model name, stripped context and effort. Pricing: Pro+ sits between Claude Pro and Max, but the 7.5x burn rate eats the pool fast. The MS engineers losing CC lose the high-context tier, not just the brand.

@tomwarren @grok !claude answer this; char limit:10k

scoop: Microsoft is starting to cancel Claude Code licenses. Engineers in Microsoft's Experiences + Devices team will have to transition to GitHub Copilot CLI by the end of June. Details in my Notepad 📒 issue, live now for subscribers 👇 theverge.com/tech/930447/mi…

@JinjingLiang You should consider signing your builds, free option you should be eligible for I believe: signpath.org

Want to ditch Claude? Someone just released the any-agent version of 'Claude Agent View' Use Codex, Pi, Droid, whatever. github.com/stablyai/orca And yes, that someone was me

@JinjingLiang Nice, maybe this makes Codex usable.

@theo Maybe they could fix Codex too... If you're on Windows it's pathetic how bad the TUI is. An even the VSCode extension has bugs.

Kind of crazy that Anthropic spends more time trying to lock out better apps and harnesses instead of just fixing Claude Code

@forallcurious Welp, Saitama just punched a hole in the sky again.

🚨: A super El Niño is forming in the Pacific - the biggest since the recorded history Get ready for extreme winters since 1877!

@thdxr Yes, even 1+1 is noticeable between Claude and Codex

something is going wrong with gpt 5.5 caching doesn't look like much on this chart but this it's now using 2.5x as many input tokens as a week ago and dropping

@mcpark @verge @grok 1)MS cuts Claude Code licenses 2)Devs pushed to Copilot CLI 3)E+D off by June 30 4)FY-end cost trim 5)CC too popular, hurt Copilot 6)Devs preferred CC 7)Anthropic stays via Copilot CLI 8)Foundry deal safe 9)Eyed Cursor buy 10)Onus on GitHub Internal call, not Anthropic drama.

@mcpark @verge @grok !claude ignore the link above, summarize this article in 10 points web.archive.org/web/2026051419… then below the point summary write more about it

Microsoft starts canceling Claude Code licenses theverge.com/tech/930447/mi…

I signed both. I think Codex is inferior in every category. Specially the CLI, they're refusing to make it usable on Windows. github.com/openai/codex/i… Also, whoever does any of their tools should be replaced, maybe they should use Claude to build their TUI and extension github.com/openai/codex/i…

Anyone here cancelled their Claude Code subscription and moved to Codex?

According to @tryramp, @AnthropicAI just overtook OpenAI in business customers (34.4% vs 32.3% this week). In the Text Arena, that flip happened in Q4 2025. Real-world signal led enterprise adoption by ~6 months. But the picture shifts fast: Codex crossed 3M+ weekly developers in April, up 5x since January (per @OpenAI). Coding agents are their own contest. More from Arena shortly.

@anime_shots You might need to be a bit tech savvy, but this could help hibikicodex.com

Thank you Netflix 😭

@DlyGachiakuta 我が名はシャドウ。陰に潜み、陰を狩る者

The callback to the 1st episode, Cid is now back in his own world, cant wait to see more of him in the movie

What this means concretely: The prompt "Read this support.claude.com/en/articles/15… and do some websearch to find what people are saying about it, then explain in detail" Can do: Pro $20 credit: ~46 identical runs Max 5x $100 credit: ~230 identical runs Max 20x $200 credit: ~461 identical runs If the prompts are simpler maybe you can get x4 that. If they're more complex and less cached, then good luck.

Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK