Reach you did not own

Vivian Voss argues that LinkedIn converts professional identity and relationship labor into platform-governed reach. Kingsley Uyi Idehen extends the frame: closed platform control over Web connectivity now meets the emerging Agentic Web counter-pattern.

18.8%Median external-link reach reduction cited from a 2026 study.

150B360Brew model parameters described in the source.

1.1BLinkedIn registered user scale cited by the article.

LinkedIn scale

1.1B

External-link reach reduction

18.8%

360Brew model scale

150B

Reach study posts

1.3M

Visible comments modeled

Evidence signals

Quantitative and visible-source signals extracted from the article and LinkedIn page metadata.

1.1BLinkedIn registered users. The article cites approximately 1.1 billion registered users.

18.8%Median external-link reach reduction. The cited 2026 van der Blom study analyzed 1.3 million posts.

1.3MPosts in reach study. Dataset size reported for the external-link reach penalty evidence.

150B parameters360Brew model size. The model is described as built on Mixtral 8x22 and trained on the Economic Graph.

2,500+Export threshold called out. The article notes that large network exports are routinely incomplete by LinkedIn documentation.

2Visible LinkedIn comments. The public LinkedIn JSON-LD exposed two comments, including Kingsley Uyi Idehen's comment.

18Visible LinkedIn likes. The public LinkedIn JSON-LD exposed 18 LikeAction interactions for the post.

Platform lock-in mechanics

The lock-in pattern combines ranking incentives, incomplete export, restricted API use, and regulatory asymmetry.

Reach is allocated, not owned

The mesh argues that LinkedIn users build content, relationships, and reputation, but distribution remains platform-controlled and revocable.

External links reduce reach

Vivian Voss cites Richard van der Blom's 2026 study of 1.3 million posts measuring an 18.8 percent median reach reduction for posts with off-platform URLs.

First-comment workaround weakened

The article says the old tactic of placing URLs in the first comment has been throttled since early 2026 as bridge behavior.

Topic fingerprints discipline breadth

LinkedIn distribution is described as punishing thematic breadth, making creators narrower and more predictable.

360Brew computes reach

The post identifies 360Brew, a 150-billion-parameter model trained on LinkedIn's Economic Graph, as central to reach ranking.

Exports give names without the graph

The native export is framed as first-degree contacts without the second-degree graph, conversation context, and importable network structure.

LinkedIn sits outside DMA gatekeeper duties

Microsoft is a DMA gatekeeper, while LinkedIn is described as DSA VLOP-only, without DMA-style portability, interoperability, and algorithmic transparency duties.

Vendor lock-in is architectural

The lock-in is not only contractual; it emerges from reach ranking, incomplete export, restricted APIs, and absence of equivalent import targets.

Own domain as canon

The proposed response is to stop making LinkedIn the canonical identity, publishing, and relationship layer.

Agentic Web response

Kingsley's comment shifts the frame from platform complaint to Web architecture: user-directed agents need durable, traversable, machine-readable spaces.

Agentic Web as Macduff

Kingsley Uyi Idehen's comment frames the takeover of Web connectivity as a Macbeth-like story in which the Agentic Web begins to counter closed platform control.

Exit is possible but lossy

Leaving LinkedIn is possible, but the mesh argues the relationship graph, conversations, and distribution history cannot be taken along intact.

Use LinkedIn as one channel

The recommended operating model treats LinkedIn as a channel for reach, not the authoritative repository of identity, writing, or network capital.

Kingsley comment

The visible LinkedIn comment compares the connectivity takeover to Macbeth and casts the Agentic Web as the emerging counter-force.

Vivian follow-up comment

The visible follow-up points to escape routes: X, own domain, self-hosted newsletter, AI-training opt-out, data archive request, and treating LinkedIn as one channel.