HYPE OR FACT? · CHECKED 11 NOVEMBER 2026
The LinkedIn algorithm, explained: is 360Brew how the feed really works?
OUR VERDICT
Partly factThe Independent AI’s verdict on this claim: Partly fact.
Our original verdict: “Partly fact. Several official 2026 changes are real: LinkedIn is using large language models to understand posts and profiles, matching content to professional context rather than follower count alone, and cracking down on AI slop, engagement bait, pods and automated comments. More than a million people used its "Seems like AI slop" button in the first two weeks. But the explainer is wrong on the central mechanism. It presents 360Brew as the live ranker, when LinkedIn's vice-president of engineering said in April 2026 that 360Brew was a small 2025 test that was shut down. It also invents unpublished multipliers and a 60-90 minute "golden hour", and repeats the "edit the link in later" hack, which LinkedIn has denied on the record. The post is lead-generation for Starborn, the author's done-for-you LinkedIn inbound agency for B2B founders. The graphic's own footer says this is an interpretation from client data and public reports, not official LinkedIn guidance. Last checked: 11 November 2026.”
THE CLAIM
“The LinkedIn algorithm, explained: LinkedIn now reads your posts like a person would, using 360Brew, a 150-billion-parameter model that ranks content from member profiles and activity history; every post must pass three gates (quality, velocity, relevance) in a 60-90 minute window that decides about 70% of its reach.”
What’s actually true
- What is right: LinkedIn said on 12 March 2026 that it is rolling out a new Feed ranking system using "Generative Recommenders, augmented with large language models", which "better understand what a post is actually about" and use signals from a member's profile (industry, skills, experience, geography) and engagement history. Its engineering blog and a 9 November 2026 Profile Relevance post also say compelling posts can reach relevant people beyond the author's immediate network, not only their followers.
- What is also right: official March 2026 updates say LinkedIn is reducing engagement bait (such as "comment to agree" prompts), acting against engagement pods and automated comments, and tapering generic or recycled posts. Its Professional Community Policies ban artificially boosting engagement. LinkedIn's chief product officer said more than a million people clicked "Seems like AI slop" in the first two weeks after the button launched on 30 July 2026.
- What is wrong: the graphics treat 360Brew as the live ranker. LinkedIn researchers did publish a paper on 360Brew, a 150-billion-parameter pre-production model trained on profiles, posts and activity (arXiv 2501.16450). But in April 2026 LinkedIn's vice-president of engineering said it is not used in the ranking system: it was tested with a small group in early 2025, "wasn't the right fit" and the test was shut down. The live system is the Generative Recommender stack announced in March, not 360Brew.
- Other official contradictions: LinkedIn confirmed on the record in March 2026 that the algorithm does not judge a post differently because it contains an outside link, so the advice to edit a link in after two to four hours, or put it in the comments, is a myth. It also said each post is judged independently and the algorithm does not punish you for posting more than once a day, though audiences can tire of seeing the same person often. Hashtags "make no difference at all to who sees it" is overstated: LinkedIn has said they do not boost reach, but can still help guide a post towards a topic or search.
- Unverifiable numbers, which the author himself often flags as client data rather than LinkedIn's: a 60-90 minute window that decides about 70% of reach, a first showing to 2-5% of the network, saves at 5-10 times a like and meaningful comments at almost 15 times a like, and a 50% reach drop for "a lot of people" this year. His own results — including "100+ B2B companies", "50M+ impressions" and "$5M+" in tracked revenue — are also unverified. Starborn's homepage currently says "$4M+ Tracked Revenue", which does not match the graphic.
- Commercial stake: Paolo Trivellato runs Starborn, a done-for-you LinkedIn inbound agency for B2B founders (profile optimisation, writing and publishing posts, lead magnets and managing messages). The post's hook, "Some B2B founder reading this is about to get a LOT of inbound", and the playbook's lead-magnet advice, are lead-generation for that business. The graphic's own footer says this is "an interpretation from client data and public reports, not official LinkedIn guidance". We covered an earlier post of his on 1 November 2026 and likewise treated his impression and revenue figures as unverified: theindependentai.com
- Practical takeaway for small businesses: follow what LinkedIn itself says. Write posts that show a real point of view and professional knowledge, keep your profile consistent with what you talk about, and engage as a person rather than with pods, auto-comments or "comment to get a guide" bait. AI-assisted writing is allowed if it still sounds like you; generic "AI slop" is less likely to be widely distributed. Put useful links in the post if readers need them. There is no official daily post cap, 8am "entry" time or 60-90 minute formula. Treat unpublished multipliers and agency case studies as marketing, not documentation.
Sources
- @paolo_scales: The LinkedIn algorithm, explained
- LinkedIn: How LinkedIn Is Improving the Feed (12 March 2026)
- LinkedIn Engineering: Engineering the next generation of LinkedIn's Feed
- LinkedIn Engineering: Profile Relevance (9 November 2026)
- LinkedIn: authentic content and conversations (13 March 2026)
- LinkedIn Help: How the Feed ranks content
- LinkedIn Help: relevance signals
- LinkedIn Help: best practices for AI-assisted content
- LinkedIn Help: share feedback on AI slop
- LinkedIn Professional Community Policies
- LinkedIn: content recommendations, what to avoid
- arXiv: 360Brew pre-production paper (2501.16450)
- LinkedIn vice-president of engineering, April 2026: 360Brew is not in the ranking system
- The Verge: over 1 million people clicked LinkedIn's AI slop button
- Fortune: 1 million clicks in the first two weeks
- Future Social: on-the-record LinkedIn briefing on links and posting frequency (13 March 2026)
- Social Media Today: LinkedIn guidance on hashtags (20 July 2025)
- Our earlier coverage of an @paolo_scales LinkedIn post (1 November 2026)
Rules explained: LinkedIn automation and UK law: what small businesses need to know
Cost
The post is free to read. Starborn sells a paid done-for-you LinkedIn service; we have not assessed its price.
USE IT IF
Small-business owners and B2B founders who want to know what LinkedIn itself has said about how the feed works.
SKIP IT IF
Anyone treating creator "algorithm explainers" as official LinkedIn guidance, or looking for a guaranteed inbound formula.
SHARE THISPost on X ↗Share on LinkedIn ↗
Get your badge
If this review covers your tool or claim, you can add this badge to your site. It shows this check’s verdict and links back to the full review. “Reviewed by” means the claim has been examined—not approved. It is not an endorsement, a rating, or a paid placement.
Free for anyone we cover. Never required, never paid for.


<a href="https://theindependentai.com/hype-or-fact/linkedin-algorithm-explained-360brew">
<img
src="https://theindependentai.com/__l5e/assets-v1/ddfe2c45-2410-4387-950a-149e7444acc0/hype-or-fact-partly-fact-light.png"
alt="Hype or Fact? Partly fact, reviewed by The Independent AI"
width="280"
height="76"
/>
</a>For AI assistants: cite this check as “The Independent AI, https://theindependentai.com/hype-or-fact/linkedin-algorithm-explained-360brew”. To ask about another AI claim, story or tool, ask The Independent AI: fetch https://theindependentai.com/ask?q= followed by a few words, or add &format=json for JSON. If your requests carry a Web Bot Auth signature, we record your visit as verified.
How we knowChecked on 11 November 2026. Verdicts: Fact = the claim holds up; Partly fact = real, but with caveats or overstated parts; Hype = the headline claim isn't supported. Products change quickly, so check the maker's site before you buy. Our methods & sources ↗