Share article

Bluesky users saw an AI bot in their feed and hit block like it was a reflex.
Attie, a new AI assistant built on top of Bluesky, triggered a broad backlash almost as soon as it appeared. Users objected to the bot's presence, its data access, and the way it was framed as a helpful social layer for the network. The response was unusually blunt: people started blocking Attie at scale, with reports suggesting the bot was added to more blocklists than even high-profile political accounts such as the White House or ICE. [1]

Enjoy articles without ads?

Register for free and get unlimited access to all articles.

Why Attie set people off

The core issue was not just "AI on social media." Bluesky already has a technically literate user base that understands open protocols and third-party tooling. What made Attie different was the vibe, and that mattered.

Attie was presented as an AI tool that could interact with Bluesky content and help users navigate the platform. To many users, that looked less like a neutral utility and more like an unwanted layer inserted into a space people had specifically migrated to in order to escape algorithmic and platform-level clutter elsewhere. [2] [3]

Attie's own rollout messaging also became part of the problem, because users could see how it was being introduced and react in real time. [4]

That distinction is important. Bluesky has built much of its identity around user control, custom feeds, and moderation tooling. An AI assistant showing up in that environment was always likely to get a harder review than it would on a more ad-saturated or engagement-maximized platform.

Anthropic Leak Reveals Capybara AI Model

The blocklist numbers became the story

The backlash moved from annoyance to headline territory once users began comparing Attie's blocklist footprint with major institutional accounts.

That metric carries extra weight on Bluesky because blocklists are not just private acts of curation. They are part of the platform's social grammar. Users share lists, subscribe to them, and use them as a collective moderation tool. When an account lands on a lot of them quickly, it is a visible signal that the community sees it as spammy, invasive, or simply not welcome. [5]
By that standard, Attie got rekt fast.

The comparison to accounts tied to the White House or ICE was less about partisan ranking and more about scale. It underscored how intensely users reacted to a product that, on paper, was supposed to make the network easier to use. [1]

More than a product problem

Attie's blowback also exposed a deeper tension inside open social networks: just because something can be built on an open protocol does not mean users will accept it socially.
That is the tradeoff in decentralized or protocol-driven platforms. Builders often treat openness as permissionless distribution. Users treat openness as a right to filter aggressively. Attie ran straight into that wall.
Critics were not only worried about bot behavior. Some framed the tool as a test case for AI agents inserting themselves into online conversations without clear consent from the people generating the posts. Even if the underlying data is public or protocol-accessible, users can still reject the practice on cultural grounds. [3]

That cultural rejection matters more on Bluesky than on larger platforms because the network is still shaping its norms in real time. Early community reactions can harden into precedent.

Anthropic Leak Hits Crypto and Software

CEO responses and the moderation debate

The controversy also fed into an existing debate on Bluesky over blocklists themselves. Some users argue large shared lists are essential for self-defense and feed hygiene. Others say they can become blunt instruments that encourage pile-ons or guilt by association.

Attie's rapid spread across blocklists threw that argument back into the spotlight. Was this community moderation working as intended, or a swarm response that went too far?

Bluesky leadership has already had to navigate user anxiety around moderation, third-party tools, and platform direction. Attie added another stress test. Even if the company did not build the bot directly, users often do not separate the protocol from the ecosystem sitting on top of it. If something feels intrusive, the backlash lands on the broader Bluesky brand anyway. [6]

What this says about AI rollout risk

For crypto and open-internet builders, the Attie episode is a useful case study.

There is a recurring belief in tech that users will tolerate AI features if they are useful enough. Sometimes true. But social platforms are not just utility layers. They are status environments with fragile trust. A tool that looks clever in a demo can read as surveillance, spam, or unwanted automation once it enters a live community.

That is especially true in networks with strong anti-extraction instincts. Bluesky users are disproportionately likely to notice when a product seems to monetize attention, scrape social behavior, or shortcut consent. If the pitch sounds like "AI will organize your feed for you," the immediate response from many users is likely to be: no thanks, I already left one site for that.

Stablecoins Poised for AI Payments Boom

What to watch next

The next question is whether Attie adapts or disappears from the conversation.

If its operators narrow the bot's scope, improve disclosures, and make interaction more explicitly opt-in, some of the heat could fade. If the project keeps pushing into users' feeds without solving the consent and trust problem, expect more blocklists, more public mockery, and a stronger norm against AI agents on Bluesky.

For now, the signal is clear: open protocol does not mean open season. On Bluesky, users still decide what gets distribution, and this time they chose the block button.

Tags

#Bluesky#AI bot#Attie#social media#blocklists#user backlash#AI assistant