Two stories today, same theme: AI is getting less chatty and more useful. Bluesky shipped a tool that turns prompts into shareable mini-apps, and the "model that can't write a sentence" just posted the fastest adoption numbers Vercel has ever logged. If you're building with Claude Code or Cursor, both are worth five minutes.
The lead: Bluesky's Attie turns prompts into shareable mini-apps
Bluesky announced at TechCrunch Disrupt that Attie, its AI feed-builder from earlier this year, now lets anyone build websites, games, polls, and dashboards from a prompt and share them with other people. CEO Toni Schneider's pitch: "social vibe coding or vibe coding for your community." Translation: instead of solo vibe coding a thing only you'll ever open, you build something your group actually uses.
The mechanics are simple. You type a prompt, Attie generates the app, and it gets a shareable link on the attie.site domain with your attribution at the bottom. The Verge's reporter made a Pokémon poll in a few minutes, and the showcase page has a Flappy Bird clone, a drum machine, a quiz, and an idle clicker. These are lightweight novelties, not production apps, but that's the point: distribution is baked in. Apps run on ATProto, the same protocol under Bluesky, and COO Rose Wang's framing is the real hook: "Who's using the thing you're building, because we haven't solved that problem with distribution." Attie comes with Bluesky's audience attached instead of you having to go find one. Worth noting: Bluesky's own active user base has reportedly been shrinking even as the company expands into tools like this, so "distribution" here means a network that's not currently growing, not a guaranteed audience.
[YOUR TAKE: worth trying Attie this week, or is "vibe coding for your Discord" already solved by other tools you use?]
Quick hits
GitLab's MCP server goes GA, and it covers the whole lifecycle
GitLab 19.5 ships a generally-available MCP server that lets Claude Code, Cursor, Copilot, Codex, Gemini, and other MCP clients work across code, merge requests, work items, and pipelines, under your existing GitLab permissions. New tools include semantic_search (find code by meaning, not filename), get_pipeline (see why a build is red without digging through logs), and the ability to kick off full Duo Agent Platform flows without leaving your editor. Rate limits on GitLab.com: 60 calls/min on Free, 600 on Premium/Ultimate. If your agent's been guessing at GitLab's REST API and burning context doing it, this is the fix. Jev: fastest model adoption Vercel has ever measured, by Vercel's own numbers
Vercel's AI Gateway data shows Jev, the decision-only model from TypeSafe AI, hit 35% of paid teams within two weeks of its September 15 launch, almost double the reach of the next model, Sonnet 5.5. In its first 24 hours it reportedly hit about 13% of paid teams, more than twice any previous launch's day-one number. This is Vercel's own gateway data, so it's a partial picture; it likely undercounts OpenAI and Anthropic usage since both run their own APIs outside the gateway. Still, Vercel used Jev itself, swapping it in for GPT-6 Luna on harmful-command classification. The same report shows open-source models now processing over 70% of gateway tokens, with frontier-lab spending up tenfold even as its share shrinks, which reads like frontier and open models splitting different jobs rather than fighting for the same one.
The bigger idea if you haven't looked at decision models yet: instead of asking an LLM to write a paragraph and then regex-parsing a verdict out of it, you send structured state in and get a typed answer back, a Boolean, Score, or Choice with a probability, no JSON validation, no "sure, here's your answer" wrapper to strip. That's the whole pitch, and it's apparently resonating with the unglamorous plumbing jobs inside agent pipelines, ticket routing, tool selection, risk checks, that never get a demo video but eat a lot of API calls.
[YOUR TAKE: anyone replaced a "classify this with a chatbot" step with a decision model yet, and did it actually simplify your pipeline?]
What this means for you
- If you've got a classification or routing step duct-taped onto an LLM call (parsing JSON, catching malformed output, hedging), that's exactly the job decision models like Jev were built for. Worth a prototype before you build another fallback parser.
- On GitLab: if your agent is already using the MCP server, turn on semantic_search and get_pipeline first. They cut the back-and-forth that usually eats your context window.
- Attie is low-stakes enough to try this week: prompt a poll or a small game for your Discord and see how it holds up outside a demo.
- Remember the Vercel numbers are gateway-only. If you're running Claude or GPT through their native APIs, you're invisible to this data, so don't treat "35% adoption" as "35% of all developers."
What's the first thing you'd wire a decision model up to replace: a classifier, a router, or a moderation check? Drop it in the thread.
Sources
- Bluesky expands its AI-powered Attie app to let anyone build websites, games, and more – TechCrunch, Oct 15, 2026
- Bluesky's AI tool now lets you make games and apps to share on social – The Verge, Oct 15, 2026
- Connect your agents to your software lifecycle with the GitLab MCP server – GitLab, Oct 15, 2026
- Jev is Vercel's fastest-adopted AI model ever – TechCrunch, Oct 16, 2026
- Jev is Vercel's fastest-adopted AI model ever – Mezha.net, Oct 16, 2026
- Jev becomes fastest-spreading AI model on Vercel developer cloud – 디지털투데이, Oct 16, 2026
- TypeSafe's Jev becomes Vercel AI Gateway's fastest-adopted model ever – Startup Fortune, Oct 17, 2026
- Jev Reached 35% of Paid Teams on Vercel's AI Gateway in Two Weeks – TokenPost, Oct 17, 2026