What the internet is saying about GitHub Copilot

What the internet is saying about "GitHub Copilot"

🟢 Supporters

• •Copilot Fusion is the main positive headline. GitHub announced a changelog item on routing and orchestrating models with Copilot Fusion [1]. A tweet shared on HN says it "orchestrates between models at half the cost" [2]. Both HN posts drew little discussion (3 and 6 points, 0 comments), so this is a signal of interest, not a groundswell.
• •Developers are building on Copilot's infrastructure. A pull request adds a github-copilot provider to the GoModel gateway. It authenticates with a GitHub OAuth token against api.githubcopilot.com and bills against a Copilot plan [10]. Copilot also appears in a roundup of AI CLI tools alongside Claude Code, Codex and Gemini CLI [11].
• •Some usage-cost friction is being resolved in favor of users. One project fixed its AI usage page to price Copilot Next Edit Suggestions at zero, after 374 requests showed as "unknown cost." It cites GitHub's billing docs on code completions [9].
• •Copilot is still used as a teaching entry point. GitHub Skills has a "Getting Started with GitHub Copilot" exercise [8].

🔴 Skeptics

• •Users want more control over agentic changes. The most engaged Copilot-specific item is a VS Code meta-issue, "Feedback on granular control over Copilot changes," with 232 comments and 295 reactions. It opens with a team update dated October 22 and asks commenters to review it first [6]. The excerpt doesn't show the individual complaints, but the volume signals strong user concern about how changes are made and reviewed.
• •Integration is brittle in practice. Several gh-aw-firewall PRs describe real failures:
  • •Subagent requests failed when the selected model rejected the wire API or translation couldn't preserve custom tools [3].
  • •Steering notices modified instructions and invalidated the prompt cache. In the reported runs, the warning request cost about 9× a normal step and exhausted the budget [4].
  • •Unconstrained auto model selection can pick a model incompatible with the client's protocol [5].
  • •gpt-5-mini requests carrying reasoning_effort: "short" were rejected [7].
• •Platform reliability is a worry, even if not Copilot-specific.
  • •HN threads cover GitHub outages: Actions/Pages/Webhooks degradation (134 comments) [15], "GitHub Is Down (Again)" [30], and "Another Friday, another GitHub incident" [23].
  • •Polymarket runs several markets on GitHub critical incidents [20][25][27]. The "longest critical incident in October" market shows about $20.2k volume [20].
  • •One HN commenter on the Roku move to GitLab suggests the outages are not mainly due to engineers using AI. That commenter is rebutting a common assumption [13].
• •Wider GitHub frustrations include the cost of Actions [16], DMCA handling [18], stolen-token abuse [21], and the GNU critique [28]. None of these is about Copilot directly.

⚪ Neutral Observers

• •Much of the "GitHub Copilot" evidence is incidental. Many items are unrelated repos that mention GitHub or Copilot in passing [12][17][24][26][29][31][33][35][36][37][38].
• •Reddit discussion about AI coding centers on Claude, not Copilot. Examples are a Claude-built TV overhaul with 34,800 stars [14] and Strata rewriting history to strip "Co-Authored by Claude" [19].
• •GitHub made its new dashboard the default [22]. The thread drew 95 comments, but the excerpt doesn't show the sentiment.
• •One ask-HN thread wonders whether source code becomes expensive if developers leave GitHub [32].

Cross-Platform Signals

• •GitHub is the most active source, with Copilot-specific signals from issues and PRs. These are largely developer-integration and control concerns [3][4][5][6][7].
• •Hacker News covers Copilot Fusion with low engagement [1][2]. Its larger threads are about GitHub's reliability and governance [13][15][18].
• •Reddit has essentially no Copilot-specific discussion in this data.
• •Polymarket only reflects GitHub incident-risk markets [20][25][27].
• •There is no clear consensus. Product news is positive but thinly discussed, while the most engaged signals are about control, reliability and integration friction.

Community Verdict

Copilot is seen as an active, evolving platform: Fusion's multi-model routing and a cost claim of "half the cost" [2] show GitHub pushing agentic and multi-model features. The strongest evidence of community sentiment is cautionary, though. Developers want granular control over agent-driven changes [6], integrators hit model-routing and cost problems [3][4][7], and the surrounding GitHub platform is widely seen as unreliable [15][23][20]. Overall: interest in the direction, wariness about control, cost predictability and stability. The Fusion claims are not yet independently validated in this evidence.

---

Based on Hacker News + GitHub + Reddit + Polymarket data from the last 30 days, researched with last30days by Matt Van Horn.