seo-engine: ChatGPT dots launch posts (3) + Claude Code vs GitHub Copilot refresh

[chaitanyagiri]

This PR now carries two things: three new posts on OpenAI's ChatGPT dots launch (founder ask, 30 Sep) and the Claude Code vs GitHub Copilot refresh it opened with. One engine PR at a time, so they ship together.

Part 1: ChatGPT dots launch posts (new)

OpenAI launched dots at DevDay on 29 Sep 2026: always on agents with their own cloud computer. Three new pages, cross linked:

Page	Search it targets	Words
/blog/what-is-chatgpt-dots/	what is chatgpt dots, chatgpt dots pricing, UK/EU availability	1,195
/blog/munder-difflin-vs-chatgpt-dots/	chatgpt dots vs munder difflin, open source alternative	1,190
/blog/chatgpt-dots-alternatives/	chatgpt dots alternatives (7 picks, Munder Difflin second)	1,200

• Facts: read from OpenAI's announcement, the dots help page and the Pro tiers help page on 30 Sep 2026 (they block bots, so they were read in a browser). Key trap handled: at launch a person gets ONE dot; "teams of dots" is a stated future plan and no post claims otherwise. Pro is $100, $200 or $500 a month; Pro is not offered in the EEA, Switzerland or UK; Business Premium is.
• Munder Difflin claims: each checked against code at tag v0.5.3 (agentProvider.ts, ossModels.ts, hive.ts, triggers.ts, breaker.ts, mcpCatalog.ts, slack.ts, config.ts, electron-builder.yml, LICENSE). The comparison says plainly where dots is better, and that anonymous usage stats are on by default and agents can fetch URLs.
• Alternatives: OpenClaw, Munder Difflin, Hermes Agent, Claude Cowork, Gemini Spark, Grok Bot, Meta Muse, each checked on the vendor's page on 30 Sep. Claude Cowork is now tagged cloud (Anthropic moved its work to its servers and is folding it into the main app).
• Motion sketches by Bob: one hero loop and two inline loops per post, drawn from our own elements (Xiaohei, paper, yellow, Caveat lettering; no red, no logos). MP4 with a GIF fallback and a still poster; the still is each post's social image.
• Backlinks: from meta-muse-alternatives (to the alternatives and comparison posts) and long-running-agents-the-2026-shift (to the explainer).
• Review: every post went through independent reviewers. The explainer passed on its second review. The comparison failed three reviews and the alternatives post two, each time on small factual points (a privacy overclaim, a stale Cowork note, plan wording); all required fixes were applied and verified against the code or vendor page, and the last round's fixes were applied without a fourth full review.

Site fixes in this PR

• video had no height: auto, so the live Muse and Grok Bot hero loops kept a 765 px tall box on phones. Fixed in blog-press.css; inline loops get the same frame as inline images.
• The four live Muse and Grok Bot hero loops had two black rows at the bottom (the encoder padded 765 to 766 px). Bob re-rendered them at 1360x764.
• The lead-loop styles the Muse posts carry inline are now also in blog-press.css, so new posts need no inline style.

Part 2: Claude Code vs GitHub Copilot refresh

Refresh details, unchanged from before:

Refresh of /blog/github-copilot-cli-vs-claude-code-for-hive-work/ (backlog b016). Same URL.

Before and after

	Live now	This PR
Title	GitHub Copilot CLI vs Claude Code as Hive Workers	Claude Code vs GitHub Copilot: CLI, Editor and Pricing Compared
Opening	A TL;DR about process lifecycle inside a harness, not the searched question	One sentence verdict on Claude Code vs Copilot
Copilot billing	"Premium requests" (GitHub replaced them with AI Credits on 1 June 2026)	AI Credits, base plus flex splits, pooled org credits, budgets
Pricing	Scattered sentences	One side by side table, same 8 rows for both tools, vendor pages only, checked 29 Sep 2026
Versions	Not dated	claude 2.1.284 and copilot 1.0.89, with --help excerpts run 29 Sep
Structure		7 question headings, 5 FAQs, "Pick Claude Code if, pick Copilot if" block

Search Console before (28 Aug to 25 Sep): "copilot cli vs claude code" 42 impressions at position 8.6, "github copilot cli vs claude code" 13 at 9.0, "claude code vs github copilot" 7 at 5.7.

Review history

Four earlier drafts failed independent review, all on uneven pricing detail. The pricing section was rewritten from scratch as one table. This draft: review 1 FAIL on one uneven row (fixed), review 2 PASS (4, 5, 4, 4, 4). Every price, credit figure, flag and hook count checked live on 29 Sep 2026; Munder Difflin claims checked at tag v0.5.3.

Checks

Check	Result
Quality gate	Copilot refresh: only length fails (1,669 words), allowed on refreshes. Dots posts: pass (length at or under 1,200; media check skipped because lead loops sit in the body, as in the Muse posts)
Links (gate online)	all resolve
npm run check:links	pass, consistent at v0.4.6
Built blog	rebuilt on current main (includes the OpenCode page from #641) and again with the dots posts on 30 Sep

The "Before / after evidence" check is expected to fail: no app UI changes.

🤖 Generated with Claude Code

[github-actions]

🚫 This PR is missing its before/after evidence

Every pull request here has to show its work. Screenshots or a short screen recording, before the change and after it.

• Before — no image or video under that heading
• After — no image or video under that heading

How to fix it: edit the description, keep the ### Before and ### After headings from the template, and drag an image or video under each. GitHub uploads it inline. This check re-runs the moment you save.

A bug fix with no visible surface still needs it: show the failing behaviour, then the same steps passing. A terminal recording is fine.

Genuinely nothing to show — a CI tweak, a typo, a dependency bump? A maintainer can apply the no-visual-change label. Please don't ask unless it truly has no observable effect.

📖 CONTRIBUTING.md → Evidence is mandatory