Half of it, genuinely. By late 2025, over 50% of long-form LinkedIn posts are likely AI-generated, AI content overall has surged 189% since ChatGPT launched, and average impressions have dropped 63 to 66% since 2023. The waffle is real, it is growing, and the algorithm is increasingly allergic to it.
The other half, specific, first-person, anecdote-rich content from niche experts, still drives genuine pipeline. Below: how to tell them apart, who LinkedIn still works for, and when walking away is the right call.
LinkedIn in 2026 is not a load of waffle, but it hosts a lot of it. The platform's 360Brew algorithm now actively suppresses generic content and rewards posts with genuine dwell time, specific detail and substantive comments. A niche expert with 2,000 engaged followers can outperform a guru with 200,000 in the same vertical. If you produce signal, LinkedIn still works. If you produce waffle, it will bury you.
Not all waffle is the same. Here is the taxonomy, how to spot each type, and why 360Brew rewards them briefly before suppressing them hard.
How to spot it
Perfectly punctuated paragraphs with zero personal detail. Ends with 'Thoughts?' Contains 'in today's fast-paced world', 'leveraging synergies', or a wall of em dashes. No specific number, no story, no name.
Why it spreads (briefly)
It looks credible at first glance. The grammar is clean and the argument is superficially coherent. LinkedIn's 360Brew algorithm only realises it is waffle when dwell time collapses after the first two sentences. By then it has already served 300 impressions.
How to spot it
'I was eating 2-pound noodles in a tiny flat. Then I built a 10M-pound company. Here's the lesson.' The hardship is performative, the lesson is completely generic ('consistency wins') and could have been written without the story at all.
Why it spreads (briefly)
Vulnerability triggers emotional engagement. Likes and 'so inspiring!' comments pour in from people who enjoy the narrative, not the insight. The 360Brew comment-weight multiplier (15x versus a like) briefly rewards it, then the algorithm notices comment quality is shallow and suppresses distribution.
How to spot it
'Leaders listen. Real leaders adapt. The best leaders do both.' Pure tautology stack. No specifics, no story, no receipts, no actual takeaway. Could have been written in 30 seconds by anyone.
Why it spreads (briefly)
People feel smart agreeing with truisms. The post offends nobody and flatters everyone who self-identifies as a leader. Short posts used to benefit from LinkedIn's brevity boost, but the Depth Score introduced in 2026 now penalises posts with zero dwell time and no saves.
How to spot it
'10 productivity hacks I learned at McKinsey.' The hacks are the same ones every productivity blog has recycled for a decade. No personal application, no single number from the author's own experience, no original angle.
Why it spreads (briefly)
Lists are scannable and the format still gets a minor distribution nudge. But when readers open the post and see nothing new, dwell time bottoms out. The Depth Score tanks. 360Brew stops distributing it within six hours.
How to spot it
'Companies that ask for free trial work are exploiting candidates. Period.' Designed to harvest 'I agree!' comments. The framing is maximally binary, nuance is stripped out intentionally to manufacture the appearance of unanimity.
Why it spreads (briefly)
Outrage still drives comments, and comments weigh 15x more than likes in 360Brew. The problem: the algorithm also measures comment sentiment and reply depth. A thread of 'totally agree!' one-liners reads as low-depth engagement and gets suppressed faster than a post with 10 genuine back-and-forth replies.
How bad is the waffle problem, really? These figures give you the honest picture.
AI content surge since ChatGPT launch (Nov 2022)
LinkedIn AI-generated content has grown 189% since ChatGPT launched. The platform now handles an enormous volume of near-identical posts every day.
Long-form posts likely AI-generated by late 2025
Independent analysis suggests more than half of long-form LinkedIn posts published after mid-2025 contain substantial AI generation. Most have been lightly edited, a few have not been edited at all.
More engagement for human-AI hybrid vs pure AI content
Human-AI hybrid posts, where a real person adds a specific anecdote and edits for voice, outperform pure AI posts by 156% on engagement. The algorithm detects the difference through dwell time and comment quality.
Drop in average LinkedIn impressions since 2023
As LinkedIn shifted from social-graph to interest-graph ranking (the 360Brew transition), average impressions per post fell 63 to 66 percent. Waffle suffers most; niche expert content can still punch above its weight.
Engagement gap between low and high dwell-time posts
Posts with 0-3 seconds of dwell time get 1.2% engagement. Posts with 61+ seconds of dwell get 15.6%. That is a 13x gap driven entirely by content quality, not follower count.
Comment weight vs likes in 360Brew's ranking model
LinkedIn's 360Brew (150B-parameter model deployed in late 2024, fully rolled out by fall 2025) weights a single comment 15x more than a like. Waffle that collects only likes gets almost no distribution amplification.
Six dimensions that reliably separate a post worth reading from a post that is taking up space in your feed.
| Dimension | Signal looks like... | Waffle looks like... |
|---|---|---|
| Dwell time | Dwell time: readers pause, scroll back, re-read a line | Dwell time: readers see the first line and scroll past in under 2 seconds |
| Reply quality | Replies that ask follow-up questions or share a contradicting view | Replies that say 'Great point!' or 'So true!' with no elaboration |
| Author credibility | Author has a body of work: consistent niche, specific examples across posts | Author posts about leadership, productivity and motivation interchangeably |
| Specificity | A specific number, date, client name (with permission) or measurable outcome | Vague claims: 'significant growth', 'huge results', 'massive impact' |
| Anecdote vs generality | A scene: 'Last Tuesday I was on a call with a SaaS founder in Edinburgh...' | A generalisation: 'Most founders struggle with X.' |
| Visual content | A chart, screenshot or original image tied to the text's specific claim | A stock image of a lightbulb, handshake or city skyline with no relation to the post |
Lifast turns rough ideas into LinkedIn posts with specifics, anecdote and dwell time built in, so the 360Brew algorithm rewards your work in a feed full of AI noise.
Write a Real Post90 days of consistent posting. No ads.
The platform is not broken for everyone. It is broken for specific use cases. Here is the honest split.
B2B founders with a specific niche
A niche expert with 2,000 engaged followers can outperform a 200,000-follower guru in the same vertical under 360Brew's interest-graph model. Specificity is the moat.
SDRs and AEs doing pipeline via content
Multiple B2B reps report 2x to 3x pipeline from consistent, specific posting over six months. The platform is still the highest signal-to-noise professional network relative to cold email.
Consultants, coaches and service providers
A single inbound enquiry from a well-crafted post can be worth thousands of pounds. The ROI calculation is wildly in favour of staying for anyone who sells a high-ticket service.
Hiring managers and talent teams
LinkedIn is still by far the most effective channel for attracting passive candidates. Employer brand content on LinkedIn converts at rates that other platforms cannot match.
Journalists, researchers and analysts
The platform remains the primary place to find and be found by professional sources. Specific commentary on industry data still attracts genuine inbound from press and conference organisers.
Scroll-only lurkers who never post
If you only read LinkedIn and never post, the feed is optimised to serve you engagement bait and waffle. Without posting, you cannot train the interest graph to show you the signal content. You get the worst of the platform.
Creators who rely on generic inspiration posts
The 360Brew Depth Score now buries posts that lack specificity, anecdote and dwell time. If your entire strategy is motivational quotes and leadership truisms, distribution has already collapsed and will not recover.
B2C brands targeting mass consumers
LinkedIn's demographic skews 25 to 55 working professionals. If your product is a consumer app, fashion brand or mass-market FMCG, your audience is not here in meaningful numbers.
Anyone who finds LinkedIn deeply anxiety-inducing
The performative culture is real. If spending time on LinkedIn makes you feel worse about your career rather than better, walking away is a legitimate and rational choice.
Businesses that already have a full pipeline from other channels
LinkedIn takes six to twelve months of consistent effort to generate reliable pipeline. If your pipeline is already full from SEO, referrals or events, LinkedIn's opportunity cost is genuinely high.
A 7-branch if-then to cut through the meta-noise and give you an honest answer right now.
1Do you have a specific professional niche (not 'business' or 'marketing' in general)?
If yes
Stay. 360Brew rewards niche expertise disproportionately.
If no
Pause. Posting without a niche produces waffle almost automatically.
2Have you posted at least 3 times per week for the past 30 days?
If yes
Check your Depth Score analytics. If dwell and saves are climbing, keep going.
If no
You cannot assess LinkedIn's value without consistent input. Give it 60 days of real effort first.
3Is your content based on personal experience, specific clients or measurable outcomes?
If yes
You are producing signal. Stay and optimise.
If no
You are producing waffle. Fix the content, not the channel.
4Are you getting inbound enquiries, replies or DMs from target buyers?
If yes
LinkedIn is working. Do not walk away from something that is converting.
If no
Wait 90 days of specific, anecdote-rich posting before concluding the platform does not work.
5Do you spend more time scrolling than posting each week?
If yes
Flip the ratio. Producers get the compounding benefit. Scrollers pay the attention tax.
If no
Good. Keep the producer mindset.
6Is the time you spend on LinkedIn displacing a channel that is clearly working better?
If yes
Reduce LinkedIn to one post per week and redirect effort to the better channel.
If no
LinkedIn likely warrants its current share of your time.
7Has it been six months of consistent posting with zero engagement from target buyers?
If yes
Walk away. Six months is a fair sample size. Either the audience is not here or the content needs a root-and-branch rethink.
If no
Keep going. Most meaningful LinkedIn traction appears between months three and six.
The single biggest reason well-intentioned LinkedIn posts turn into waffle is not laziness, it is the blank-editor problem. You know what you want to say but the structure falls apart, the anecdote gets buried, and by the time you have finished editing it sounds like every other post in the feed. Tools like Lifast help you turn a rough, specific idea (not a generic topic) into a post that leads with the anecdote, lands the specific number and ends with something worth a comment. The specificity has to come from you. The structure, hook and dwell-time formatting can be handled by the tool.
You cannot stop waffle being posted. You can stop it reaching you. These tactics work within weeks because they train 360Brew's interest graph directly.
Unfollow rather than disconnect
Unfollowing removes someone from your feed without disconnecting. Use it liberally on anyone who posts waffle. Your feed trains the interest graph to show you more signal over time.
Use 'Not interested' on AI slop aggressively
The three-dot menu on every post includes 'Not interested in this post.' Use it on every piece of obvious AI slop you encounter. 360Brew learns your preference within days.
Follow niche hashtags, not broad ones
Following #Leadership or #Marketing floods your feed with waffle. Following #B2BSaaS or #TechnicalRecruiting surfaces more specific, expert-authored content.
Prioritise 'Comments' from your notifications
Comment threads from people you respect are higher-signal than the main feed. Check your network's comment activity to find the posts worth reading before scrolling.
Search by company or person, not the feed
Proactively searching for posts from specific people you trust cuts out the algorithmic layer entirely and gets you directly to signal.
Save posts, do not just like them
Saving a post is a strong positive signal to 360Brew. Saving good posts also trains your own reading habits: you stop scrolling for novelty and start engaging with depth.
Set a 15-minute posting window, not a scrolling session
Decide before you open the app whether you are posting or consuming. Going in without an intent defaults to scrolling and exposes you to maximum waffle.
Follow the author, not the post
When a post genuinely surprises you, click the author's profile and look at their last 10 posts. If the body of work is consistently specific, follow them. One good post does not always mean a signal account.
What signal and waffle actually look like in practice, told through the people who made the wrong and right calls.
James built a developer tooling startup and spent 18 months posting on LinkedIn. His posts were technically accurate, well-structured and almost universally ignored. He was writing about distributed systems architecture for an audience whose interest graph was trained on sales and marketing content. The platform was not broken. It was mis-targeted.
He quit in early 2025 and moved to a niche Substack and a developer Discord. Within four months his inbound was higher than it had ever been on LinkedIn. His takeaway: LinkedIn is a professional network, not a universal professional network. If your buyers do not use LinkedIn as their primary professional reading, your signal will always drown in someone else's waffle.
The lesson is not that LinkedIn is a load of waffle. The lesson is that waffle is relative to audience. For developer tools targeting CTOs who do not scroll LinkedIn, even excellent signal posts are invisible.
Sophie was a mid-market SDR at a SaaS company in Manchester. She started posting on LinkedIn in January 2025 with a simple brief: one post per week, always tied to a specific conversation she had had with a prospect that week. No AI generation. No templates. Just what she actually saw in the market.
By month four she had 4,200 followers and inbound from three enterprise accounts who had found her through her posts before she had a chance to cold-reach them. Her quarterly pipeline tripled, and her manager noticed that her inbound-led deals closed at a 40% higher rate than her cold-outreach deals.
The key was specificity. She named the industry vertical (never the company name), the actual objection ('every prospect this week said our pricing model was confusing'), and what she learned. Exactly the kind of post that earns 60-plus seconds of dwell time and a comment from someone who was just thinking the same thing. Signal, not waffle.
Ravi was a senior product manager who spent 20 minutes a day reading LinkedIn and leaving thinking it was mostly waffle. He agreed with the diagnosis. He was right about the feed he was seeing, which was heavily weighted toward generic leadership content and recycled list posts because he had never taught the algorithm anything different.
He started posting three times a week in March 2025: short posts about specific product decisions, what failed and why, what the data showed that surprised him. Within six weeks his notifications showed follows from three founders he had been trying to get in front of for months. None of them cold-messaged him. They saw the posts first.
The feed did not change overnight, but the inbound did. His rule: never open LinkedIn without having something to post. The moment he became a producer rather than a consumer, the platform's signal-to-waffle ratio in his specific corner of it improved dramatically. The waffle did not disappear. He just stopped feeding it attention.
The jargon around LinkedIn's algorithm can itself be a form of waffle. Here are five terms used throughout this page, defined plainly.
360Brew
LinkedIn's proprietary large language model (150B parameters, built on Mixtral 8x22B architecture), deployed in late 2024 and fully rolled out by fall 2025. It replaced the previous interest-graph ranking system and is responsible for the shift to dwell-time and comment-depth signals as primary distribution factors.
AI Slop
Industry shorthand for AI-generated content that has not been meaningfully edited for voice, specificity or personal anecdote. Typically identifiable by perfect punctuation, generic structure, absence of any specific number or name, and a 'Thoughts?' call to action.
Dwell time
The amount of time a reader spends on a post before scrolling away. 360Brew treats dwell time as a primary quality signal. Posts with 0-3 seconds of dwell get 1.2% engagement. Posts with 61+ seconds of dwell get 15.6%. A 13x gap that no follower count can compensate for.
Interest graph
The model LinkedIn uses to decide what you see in your feed, based on topics and content types you have engaged with, rather than who you are connected to (social graph). The shift from social to interest graph is why posts from strangers with 2,000 followers can reach you, while posts from your connections with 50,000 followers do not.
Signal-to-noise ratio
Borrowed from engineering: the ratio of meaningful content (signal) to irrelevant or low-quality content (noise) in a given feed or channel. A low signal-to-noise ratio does not mean a channel is worthless, it means the work of filtering is on the reader unless the reader actively trains the algorithm.
LinkedIn is not dead. 89% of B2B marketers use it for lead gen, 93% use it for content distribution, and roughly 80% of B2B social leads still come from it. But the noisy feed has nudged senior operators toward smaller, higher-signal channels. Here is how the landscape stacks up in 2026.
| Platform | Monthly Active Users | Skews | Best For | The Catch |
|---|---|---|---|---|
| ~1B accounts / ~310M MAU | Professional, 25-54 | B2B lead gen, recruiter outreach, decision-maker reach | Feed noisy with AI slop, important features paywalled | |
| X (formerly Twitter) | ~580M MAU, 128.8M DAU (Oct 2025) | Mixed, leans male | Founders building in public, niche tech communities, breaking takes | Quality control low, ad-driven feed, B2B reach unreliable |
| Threads (Meta) | 400M MAU (Aug 2025), 141.5M daily mobile (Jan 2026) | 58% male, 42% female, broadening | Lifestyle-adjacent B2C, casual founder voice | Limited B2B audience yet, search is weak |
| Bluesky | 42.3M registered, 27.5M MAU, 3.68M DAU (Feb 2026) | Under 35, ~60% male, tech-forward | Tech, indie creators, journalists | Tiny relative scale, B2B presence thin |
| Substack | 30M+ subscriptions | Mixed, knowledge-work skew | Owned-audience newsletters, paid subs | Slow to grow without external traffic |
| Slack / Discord | Niche, by community | Industry-specific | Higher-ROI niche networking | Discovery is hard, not search-indexable |
For all the justified criticism, the data that keeps B2B operators on the platform is hard to argue with.
of B2B marketers use LinkedIn for lead gen in 2026
No other social platform comes close for B2B pipeline. Nine in ten teams with a B2B mandate are still actively using LinkedIn as a lead generation channel.
of B2B marketers use it for content distribution
Content teams default to LinkedIn first for B2B distribution. The professional audience concentration makes it the channel of record for thought leadership regardless of feed quality.
of B2B social leads still come from LinkedIn
Four in five B2B leads that originate from social media come from LinkedIn specifically. X, Threads and Bluesky account for the remaining 20% combined.
monthly active users, still the dominant professional graph
Despite the waffle, LinkedIn's addressable professional audience remains an order of magnitude larger than any competitor platform targeting the same demographic in 2026.
Six conditions. Find yours, read the decision. No hedge, no "it depends" non-answer.
1If you have a clear B2B offer and post 3-5x per week
Stay on LinkedIn. 80% of B2B social leads still come through here and no competitor platform matches the addressable professional audience for decision-maker reach.
2If you are jobseeking
Stay. LinkedIn Recruiter search is the largest active candidate database on earth. Your profile and activity posts are live CV signals that passive headhunters and hiring managers check before anyone else.
3If your audience is consumer / B2C
Mostly leave. LinkedIn audiences buy a fraction of what Threads or Instagram audiences buy in consumer categories. Redirect your content budget to where your buyers actually spend leisure time.
4If you have an engaged community of 500+ people elsewhere (Slack, Substack, Discord)
Make that the main channel. Use LinkedIn for top-of-funnel posts only. Owned communities compound faster than rented algorithmic reach, and a 500-person engaged list outperforms 50,000 LinkedIn followers in conversion.
5If you keep scrolling without posting
Cut your LinkedIn time to 15 minutes per day. Posting compounds over months. Scrolling pays a pure attention tax with no asset being built. The feed serves you more waffle the more you scroll passively.
6If you post but get under 5 comments per post for 30 days
The format is the problem, not the platform. Revisit hook style, post length and dwell-time formatting before quitting. A content diagnosis almost always surfaces a fixable structural issue before you conclude the audience is simply not there.
LinkedIn's 360Brew model fundamentally changed distribution in a way that most creators have not fully reckoned with. The old system favoured connection-graph proximity and early engagement velocity. If your first 100 connections liked a post quickly, it spread. The new system runs on interest-graph relevance and content depth. A post with zero personal anecdote, no specific number and no dwell time simply does not propagate, regardless of how many followers the author has.
The practical effect is a feed that is increasingly bimodal. Genuine expert content with specific, first-person detail gets more reach than ever, because the interest graph can route it to exactly the right niche audience, even if that audience is spread across connections-of-connections. Generic waffle gets less reach than ever, because its dwell time collapses the moment the algorithm begins distributing it and 360Brew reads the signal as a negative vote.
360Brew also surfaces posts two to three weeks old when a topic becomes relevant to a reader's interest graph. The chronological firehose is genuinely gone. A well-crafted post from a fortnight ago can arrive in someone's feed today if they just started engaging with content in your niche. That changes the calculus on consistency: one excellent post per week outperforms seven mediocre ones because the excellent post has a longer effective shelf life.
LinkedIn introduced a composite 'Depth Score' metric in 2026 that combines dwell time, comment depth, saves and private shares into a single quality signal per post. This is distinct from the engagement metrics visible to creators and is used internally by 360Brew to decide whether to expand or suppress distribution beyond the initial serving.
The practical implications are significant. A post that collects 200 likes but zero saves and five one-word comments will score poorly on Depth. A post with 40 likes, 12 substantive comments and 30 saves will score well and receive continued distribution for days rather than hours. This explains why many high-follower creators have seen reach collapse even as their engagement rates appeared stable: likes without depth do not move the Depth Score needle.
Private shares, the 'Send to a connection' behaviour, are particularly heavily weighted. A post that people forward privately to a colleague is almost definitionally signal rather than waffle. If you are tracking your Depth Score proxies, watch your save-to-impression ratio and your comment-to-like ratio alongside raw impression numbers.
There is a real answer to the question this page is named after. LinkedIn is worth your time if you have a specific professional audience, a point of view based on real experience, and the patience to post consistently for at least 90 days before judging the results. It is not worth your time if you are using it as a place to amplify generic content, reach a consumer audience, or generate output without putting genuine experience into the input.
The platform's waffle problem is real and it is getting worse as AI generation costs approach zero. But the waffle problem is also self-selecting: the creators who are producing waffle are finding their distribution collapse in 2026, which means the signal-to-noise ratio in the feeds of people who actively curate is actually improving. The question is whether you are curating your consumption or passively accepting what the algorithm serves before you have trained it.
Walking away from LinkedIn is also a legitimate decision. The opportunity cost of 30 to 60 minutes per week of genuine content creation is real. If that time would be better spent on a channel that already works for your specific business (email, SEO, events, referrals), redirect it. LinkedIn is not the only professional channel and it is certainly not the best one for every business. The founders who do best on LinkedIn are the ones who chose it deliberately because their audience is there, not the ones who felt obligated to post because everyone else was.
Six honest answers about AI saturation, 360Brew, and whether the platform is worth your time in 2026.
Partly, yes. Independent analysis suggests over 50% of long-form LinkedIn posts published by late 2025 contain substantial AI generation. AI content on the platform has surged 189% since ChatGPT launched in November 2022. However, LinkedIn's 360Brew algorithm actively downranks posts with zero personal anecdote, clichéd templates and near-zero dwell time, which means heavily AI-generated posts are increasingly invisible rather than dominant in well-curated feeds.
360Brew is LinkedIn's proprietary large language model (150 billion parameters, built on Mixtral 8x22B), deployed in late 2024 and fully rolled out by fall 2025. It powers LinkedIn's content ranking and shifted the platform from social-graph distribution (what your connections liked) to interest-graph distribution (what matches your topic interests and engagement history). It weights dwell time and comment quality far more heavily than likes, and surfaces relevant posts two to three weeks old rather than following strict chronological order.
Not yet, unless you have genuinely given it 90 days of consistent, specific, anecdote-rich posting. Most LinkedIn engagement trajectories are flat for the first four to six weeks and then grow non-linearly once 360Brew establishes your interest-graph placement. If after 90 days of three-per-week posting with personal specifics you are still getting zero engagement from target buyers, reassess. But low engagement from generic or AI-generated posts is not a platform verdict, it is a content verdict.
Unfollow (not disconnect) people who consistently post waffle, use 'Not interested in this post' on AI slop aggressively, follow niche hashtags rather than broad ones, and save posts from authors whose body of work is consistently specific. Liking and saving good content trains 360Brew's interest graph faster than any other action. Most people notice a meaningful feed quality improvement within two to three weeks of active curation.
Yes significantly. Posts with external links in the post body lose approximately 60% of their reach. LinkedIn wants users to stay on the platform. Placing links in the first comment was a workaround that the platform began penalising in 2025 and 2026. The recommended approach is to put the link in the first comment while ensuring the post itself is substantive enough to earn dwell time and saves that compensate for the distribution penalty.
For most B2B founders selling to decision-makers who are active on LinkedIn, yes. The platform remains the highest signal-to-noise professional network relative to cold email for reaching senior buyers. The key shift in 2026 is that generic content no longer works at all. Founders who post specific, experience-based content with real numbers and named contexts consistently report meaningful inbound pipeline within three to six months. Founders who post AI-generated thought leadership are invisible.