Gokul

@0xSero Inference Engineering by @philipkiely @baseten

What am I missing here? I’ve read tons of technical books, many classics, manga

@github @acolombiadev How does licensing work when using the Copilot SDK for integration into an application or CI/CD pipeline? Since the subscription is tied to an individual GitHub account, is it possible to share that subscription across a team when using the Copilot SDK?

With the GitHub Copilot SDK, you can add the same AI that powers Copilot Chat to your own applications. To test this out, @acolombiadev integrated the Copilot SDK into a React Native app to generate AI-powered issue summaries, with production patterns for graceful degradation and caching. Here's how she did it. 👇 github.blog/ai-and-ml/gith…

I’ve been seeing a lot of people on my timeline mention they received $10K in @cursor_ai credits, any idea what the eligibility criteria is? I’d love to get some and try out the latest Agents Window and new features in v3.

@digitarald @pierceboggan @code In VS Code, I get a summarized plan and need extra requests to refine it. In CLI, I can view the full plan, accept it, run autopilot, or modify it, all within the same request. Much more efficient on premium usage since latest models come with a 7.5x multiplier.

@gkannane @pierceboggan @code Any specifics of what you would like to see in VS Code? You can iterate on the plan before sending it off with Autopilot as well. I'm curious what specific step or interaction is missing in VS Code.

@pierceboggan Plan mode in Copilot CLI is awesome, I like being able to iterate on a plan before handing it off to autopilot and see all the todos. Any way @code could bring this into the editor so we can tweak plans without multiple requests?

@abhijitwt You can set up a cron-triggered fetch or ping using Cloudflare Workers, and it doesn’t cost anything

I bought a $20 VPS just to ping my free Render backend every 10 minutes so it doesn’t go to sleep. Money well spent 👍

@pierceboggan @code @digitarald rubber duck feature is awesome too, great at catching issues and critiquing the plan

@gkannane @code Great feedback - plan mode in Copilot CLI is awesome @digitarald

Hi @_Evan_Boyle, in Copilot CLI Autopilot mode, after a task completes it says “Continuing autonomously (3 premium requests)” and generates a summary, which still consumes premium credits (e.g., Opus 4.7 at 7.5×) just to show the summary. Any way to disable this or force the summary to use a 1× model?

@DeepakNesss That’s expected, since it costs twice as much as GPT-5.4

The new GPT-5.5 consumes much more tokens than GPT-5.4! Earlier, I would rarely finish till 50% on my $100 Codex plan, but today I was able to finish 93% in the 5-hour limit.

@joffteyvl @OrenMe @GitHubCopilot Yeah, agentic workloads would get really expensive for them as long-running tasks could consume tens or even hundreds of millions of tokens, while users would only end up paying a few cents per request.

@OrenMe @GitHubCopilot It looks like their is a lot of abuse of the current pricing system. This was expected I guess, it was fun while it lasted.

Brace, brace, brace Major changes and finally clarity on usage limits, session and weekly, for @GitHubCopilot I think this reflects how much Copilot is popular and being used across but nevertheless this is a hard change for anyone who grown accustomed the the unlimited PRUs and sub agents usage freedom Hope we gonna see more news going forward for anyone wanting to further utilize their account and harness github.blog/news-insights/…

@kanavtwt GLM 5.1 and Kimi K2.5

what are some good LLMs for coding other than OpenAI and Anthropic ones? have leftover credits in OpenRouter so want to explore

Hi @thsottiaux, is there a public beta channel for the Codex app to try out new features early?

@boristane It's for product experiments, similar to Google Labs

since when is it Anthropic Labs ?

We're teaming up with @AnthropicAI to launch Opus 4.7! Open to all builders, all backgrounds. Join our global virtual hackathon powered by Claude Code. One week to push the boundaries of what's possible with Opus 4.7. April 21–28 500 builders Win up to $100k in Claude API credits Applications close Sunday, April 19th Apply Here👇

@gajesh I’d like to request an invite code, sent you a DM.

going to share invite codes here periodically. come and play around with the system. INV-C42EB823 INV-27039D15 INV-2D635711 INV-C8781E71 INV-890F6129 reply to get more codes

we've moved opencode desktop to electron. it's faster, more reliable, and will replace our tauri build soon. try it out in beta via the link below.

@plannotator @opencode Is GitHub Copilot CLI supported?

@opencode users - a new release is coming that makes life better. 1. Relaxation on the planning tool 2. Emphasis on the manual methods 3. You get full control/configurability over both

@pierceboggan @code Why is it being charged at 7.5x when token pricing is same as Opus 4.6?

Claude Opus 4.7 is now rolling out for @code and GitHub Copilot developers!

@gajesh I'd like to give it a try!

Wake the world's sleeping compute. Look at the Mac nearest to you. What's it doing? Probably nothing. There are 100M+ Macs with Apple Silicon out there. Apple quietly made them *really* good at inference. A $3k Mac runs a 60B model at 30 watts. Most sit idle most of the day. Meanwhile every AI API call passes through three layers of margin before reaching the hardware. We call this the Inference Tax. We got curious: what happens if you connect idle Macs directly to inference demand? This is Darkbloom. Private inference network for idle Macs. darkbloom [dot] dev -- paper + code open. Reply for invite + free credits ↓

@itsjoaki GLM 5.1

GLM 5.1 or Qwen 3.5?