HappycapyGuide

By Connie · This article contains affiliate links. We may earn a commission at no extra cost to you if you sign up through our links.

AI & Social Media

Bluesky's AI Tool Attie Gets 125,000 Blocks in Days — What It Tells Us About AI Trust

By Connie  ·  April 1, 2026  ·  7 min read

TL;DR

Bluesky launched Attie — a standalone AI app powered by Anthropic's Claude — at its ATmosphere conference on March 28, 2026. Within days, over 125,000 users had blocked it, making Attie the most-blocked account on the platform besides JD Vance. The backlash reveals a fundamental split in how users want AI to work: invited and controlled, not ambient and algorithmic.

125K+users blocked Attie
83:1block-to-follow ratio
#2most-blocked account on Bluesky
3 daysfrom launch to backlash peak

What Happened: Bluesky Builds an AI, Users Rebel

On March 28, 2026, Bluesky's former CEO Jay Graber and CTO Paul Frazee took the stage at the ATmosphere developer conference to unveil Attie: a standalone AI app powered by Anthropic's Claude that lets users build custom social media feeds using plain-language descriptions. No code required. Just type "show me posts about sustainable fashion from artists I don't follow yet" and Attie builds the feed.

By the time TechCrunch covered the story on March 30, the backlash was already historic. More than 125,000 Bluesky users had blocked @attie.ai, against roughly 1,500 who followed it — an 83-to-1 rejection ratio. Only US Vice President JD Vance has a higher block count on the platform.

"For many Bluesky users, the platform serves as a reprieve from the more mainstream social internet, where AI search, AI chatbots, and even AI-generated video feeds are omnipresent — which makes the launch of Attie feel like a betrayal."
— TechCrunch, March 30, 2026

Graber acknowledged the response. In a post on the platform she wrote that Bluesky would "look into ways to take into account the preferences expressed by people who've blocked @attie.ai." Attie remains in a closed, invite-only beta.

What Attie Actually Does

Attie is a separate app — not embedded directly in Bluesky — built on the AT Protocol (atproto), the open, decentralized protocol that underlies Bluesky. It is powered by Anthropic's Claude and has three stated functions:

  1. Custom feed creation — describe in plain English the kind of content you want, and Attie builds an algorithm for it. "Show me science posts that are optimistic" or "find emerging artists posting about generative AI."
  2. Cross-network content surfacing — processes posts from Bluesky and other atproto-compatible networks, letting you build feeds that span the broader decentralized social web.
  3. "Vibe coding" your own app — a future roadmap item to let users build custom Bluesky-compatible apps through conversation with Attie, without writing code.

On paper, this is a genuinely useful idea. Bluesky's custom feeds are already one of its most powerful features — allowing users to build interest-specific timelines rather than following only accounts. Attie was meant to democratize that capability. The problem was not what Attie does. It was how Bluesky introduced it.

AI that works the way you want it to

Happycapy is opt-in by design — you choose when and how to use it. No ambient AI, no uninvited algorithms. Just a powerful assistant for your work, when you need it.

Try Happycapy Free

Why Bluesky's Users Are Different

To understand the backlash, you have to understand who is on Bluesky. The platform grew rapidly in 2024-2025 primarily by attracting users who left X after Elon Musk's acquisition, journalists, academics, and creators who were specifically fleeing algorithmic manipulation and AI-generated content floods.

This is not a typical early-adopter tech crowd eager to embrace every AI product. Bluesky's core user base actively values:

What Bluesky users wantWhat Attie represents to them
Chronological or user-controlled feedsAlgorithmic curation by AI
Human-generated contentAI processing and reshaping content
Transparent, open protocolBlack-box AI recommendations
Escape from corporate AI surveillanceAn AI account observing their behavior
Community-built normsTop-down AI feature imposition

Users also pointed to a practical grievance: Bluesky still lacks basic features that have been requested for months, like the ability to send images via direct messages. Shipping an AI product before fixing core functionality felt like misplaced priorities.

The Numbers in Context

The block-to-follow ratio is striking even against other controversial accounts:

Bluesky accountApproximate blocksContext
@jdvance.bsky.social (JD Vance)~130,000US Vice President
@attie.ai (Bluesky's AI tool)125,000+Launched March 28 — 3 days old
@whitehouse.bsky.social~122,000Official White House account
@ice.gov.bsky.social~112,000US Immigration enforcement

Attie reached its block count in three days. The accounts above took months or years to accumulate theirs. It is one of the fastest collective rejection events in social media history.

The Real Lesson: AI Consent Is Everything

The Attie backlash is not fundamentally about Claude or Anthropic's technology. Attie is powered by one of the most respected AI models in the industry. The backlash is about context and consent.

AI adoption follows a consistent pattern in 2026: tools that users invite — a writing assistant they open intentionally, a coding tool they activate deliberately — are embraced enthusiastically. Tools that appear uninvited, ambient, or observational generate immediate resistance, regardless of their underlying quality.

The opt-in vs. ambient AI divide

Futurum Group's 1H 2026 CIO Insights Survey (n=695) found that 67.1% of CIOs cite data security and privacy risks as their leading AI concern. The Attie backlash reflects this at the consumer level: even technically capable AI is rejected when users feel they haven't consented to its presence.

What Bluesky should have done differently

The core product concept is sound. AI-assisted custom feed creation could be genuinely valuable. But the launch pattern — new account, conference announcement, closed beta with no opt-in pathway for the broader community — felt like something being done to users rather than for them. An opt-in toggle in the main Bluesky app, with the ability to hide Attie entirely, would have changed the dynamic significantly.

Implications for AI Products in 2026

The Attie incident joins a pattern. Bluesky's case is an extreme, but similar dynamics played out at smaller scales with Meta's AI stickers on Instagram, Microsoft Copilot's deep integration into Windows, and Amazon's Rufus shopping AI appearing in search results without opt-in. In each case, user pushback forced the company to add visibility controls or opt-out mechanisms.

The lesson for AI product builders is direct: the same technology receives dramatically different responses based on whether users feel in control. Happycapy, for example, succeeds because users open it intentionally — they come to it when they need it, rather than encountering it unexpectedly in a space they considered AI-free.

Bluesky built a genuinely interesting product. Its mistake was assuming that transparency about the technology (it's on a separate app, it's powered by Claude, it's in closed beta) would translate into trust. For a community that values autonomy, transparency is necessary but not sufficient — the right to say no before something starts matters more than the ability to block after it arrives.

AI on your terms

Happycapy is never ambient. You open it when you need it — for writing, research, coding, or analysis. Full control, no surprises. Try it free with your first workspace.

Start Free with Happycapy

Frequently Asked Questions

What is Bluesky's Attie AI tool?

Attie is a standalone AI app built by Bluesky and powered by Anthropic's Claude. It lets users create custom social media feeds using plain-language descriptions — no coding required. It was unveiled at the ATmosphere conference in late March 2026.

Why did Bluesky users block Attie?

Bluesky's user base skews strongly anti-algorithmic-AI. Many joined the platform to escape AI-driven manipulation on X, Instagram, and Facebook. Attie felt like a betrayal of the platform's promise of a more human, chronological social experience. Over 125,000 users blocked it within days — an 83:1 block-to-follow ratio.

How did Bluesky respond to the backlash?

Jay Graber, Bluesky's Chief Innovation Officer, acknowledged the backlash and stated the team would 'look into ways to take into account the preferences expressed by people who've blocked @attie.ai.' Attie remains in closed beta.

What does the Attie backlash tell us about AI adoption?

The Attie incident reveals a clear AI trust deficit: users are willing to adopt AI when they choose it (opt-in), but reject it when it appears uninvited or feels manipulative. Consent and context determine whether AI is welcomed or blocked — even when the underlying technology is the same.

Sources
  • TechCrunch — "Bluesky's new AI tool Attie is already the most blocked account other than J.D. Vance" (March 30, 2026)
  • TechCrunch — "Bluesky leans into AI with Attie, an app for building custom feeds" (March 28, 2026)
  • The Verge — "Bluesky's new app is an AI for customizing your feed" (March 29, 2026)
  • Futurism — "Bluesky Users Respond With Overwhelming Disgust to Platform's New AI" (March 31, 2026)
  • Futurum Group — 1H 2026 CIO Insights Survey
  • Happycapy AI — opt-in AI workspace for knowledge workers
SharePost on XLinkedIn
Was this helpful?

Get the best AI tools tips — weekly

Honest reviews, tutorials, and Happycapy tips. No spam.

Comments