The article argues Bluesky’s app is losing users even as the company broadens its pitch beyond a Twitter-like app toward AT Protocol, the underlying network and tooling stack. A lot of the pushback focused on measurement first. The cited metric is mobile app MAU, which misses browser use and alternate clients. That did not really rescue the growth story. People pointed to public activity trackers that count posting and likes and show the same basic pattern: activity is down sharply from the post-election and post-inauguration spikes.
Where people landed was harsher than the article’s framing. The problem is not that Bluesky picked the wrong product roadmap. It is that social networks live or die on network effects, and Bluesky still does not have enough of the people, topics, and live activity that make a text network feel essential. Several users said the product is pleasant enough, sometimes better than X, but still lacks critical mass in their interests. That gap shows up most in sports, niche communities, and professional circles where a few bursts of energy are not enough to create daily habit.
The strongest product complaint was feed quality. Many said Bluesky’s default discovery experience over-indexes on US politics and quickly overfits to stray interactions. Power users said the best experience is to ignore discovery, stay in Following, or swap in custom feeds. That is a nice feature for technical users, but it also undercuts Bluesky’s chances with normal users who expect the default app to work without hand-tuning.
The bigger split was over what counts as success. Some treated Bluesky like a venture-backed Twitter challenger, in which case shrinking usage is a flashing warning light because a VC-backed company needs growth, not just a satisfied niche. Others said the app is only one client on top of a more important open protocol, and that AT Protocol could outlast the company if the ecosystem diversifies. Even that optimistic view ran into a hard reality check. Today, usage is still heavily concentrated on Bluesky’s own infrastructure, so protocol openness is more aspiration than operational fact.
Mastodon kept surfacing as the comparison that clarifies this. Mastodon does not need to “win” in the same way because its communities can survive at smaller scale and without a single company justifying a runway. Bluesky is trying to straddle both worlds: a protocol story that sounds like public infrastructure and a funding model that still demands app-level growth. The mood was that this is the core tension, and declining activity makes it harder to keep pretending it is not.
If you care about Bluesky as a business, user decline and venture funding are the real issues, not protocol elegance. If you care about it as infrastructure or a niche communications layer, watch whether AT Protocol escapes Bluesky’s own hosting and moderation bottlenecks.
Mostly skeptical. People broadly believed Bluesky activity is falling, and the dominant reasons given were weak network effects, poor discovery, political homogeneity, and the mismatch between a VC-backed growth story and a niche community reality.
Key insights
01
Public activity data also points down
Alternative measurements do not save the app from the decline story. Public trackers based on posting and likes show a large year-over-year drop, and one commenter with social media experience argued that if app retention is falling, web usage usually falls too rather than acting as a hidden cushion.
Do not anchor on one flattering metric debate. If several independent activity measures slope the same way, treat retention and habit loss as the real problem.
The protocol may be independently implementable, but that is not the same as being operationally decentralized today. Several commenters said the hard part is not raw infrastructure cost but moderation, compliance, and the fact that almost all real usage still sits on Bluesky-run services, including key sources of truth that make the ecosystem dependent on the company in practice.
If you are betting on AT Protocol, track independent hosting, identity, and moderation capacity, not just open specs and GitHub repos. A protocol is not resilient until meaningful traffic can move without one company.
Bluesky’s strongest differentiator is that users can choose or even build feeds, but people described that as work they should not have to do. The usable path is often to hide Discover, stick to Following, or install a better community feed. That helps expert users. It also signals that the default product is failing the mainstream job of immediate relevance.
For consumer social products, extensibility does not compensate for a weak default. If onboarding requires feed surgery, you are selecting for enthusiasts and filtering out everyone else.
The sharpest framing was that Mastodon can remain small because nobody needs it to become a venture-scale company, while Bluesky cannot. Calling the app a modest, healthy community does not answer the funding question. If the company took venture money, shrinking into a niche is not a stable equilibrium.
Separate product satisfaction from business viability. A social app can be good enough for users and still be in trouble if its cost structure and cap table require growth.
Overfitted recommendations train users to click less
People described recommendation systems that latch onto tiny signals and flood the feed with a narrow interpretation of one stray interest. That does more than annoy. It changes behavior by making users afraid to interact, which is poison for a product that depends on lightweight exploration and habit formation.
If your recommendation loop punishes curiosity, engagement metrics may mask long-term trust loss. Watch for users becoming more passive because they do not trust the feed to recover from one accidental click.
For existing communities the platform feels healthy
Some users said the decline narrative does not match their lived experience at all. In the circles they follow, discourse quality and activity have improved, which suggests Bluesky may be stabilizing into a denser network for certain domains even if aggregate growth is fading.
Do not confuse global growth curves with usefulness for a target cohort. A smaller network can still matter if your customers or community already cluster there.
As protocol infrastructure Bluesky may already have succeeded
A few commenters argued that judging Bluesky only as a Twitter replacement misses the company’s stated goal. If the point is to move social networking onto an open protocol with portable identity, custom feeds, and independently run services, then the app does not need X-sized adoption to count as progress.
If you evaluate Bluesky, decide first whether you are scoring a consumer app or a protocol transition. Those are different bets with different success metrics.
Several replies pushed back on the idea that Bluesky’s main problem is left-wing politics. They argued that incumbent platforms survive far worse product and moderation failures because celebrities, governments, and niche interest groups are already there, and that inertia explains more than culture-war framing does.
When assessing a social platform, weight distribution and incumbency before you overfit to politics. A weaker culture fit can be overcome more easily than a missing social graph.
techcrunch.com