At what network size does manual scrolling stop working?

There’s no hard cutoff. Research suggests people actively keep up around 150 relationships, so once your network is several hundred connections, most of it is outside what you track day to day. That’s well within range for anyone who’s been on LinkedIn a few years.

Is AI network search only useful for sales and recruiting?

No. Job seekers, founders, and anyone doing outreach benefit — the core problem it solves, finding the right person in a network too large to remember, applies broadly.

Does AI network search replace LinkedIn's own search?

It complements it. LinkedIn's native search is good for filtering by attribute; AI network search adds ranking, reasoning, and the ability to answer compound plain-English questions.

AI Network Search vs. Manually Scrolling LinkedIn: What's the Difference?

Manually scrolling LinkedIn connections relies on remembering who's in your network and why they matter, which breaks down past a few hundred connections. AI network search instead treats your connections as a searchable, ranked database that answers plain-English questions.

The limits of scrolling and memory

People can only keep up so many relationships. Research by Robin Dunbar and colleagues puts the number people actively maintain at around 150 (a study of Christmas card networks found networks averaging about 150 people), and in a 2016 study of Facebook networks only about 28% of people’s online friends were ones they considered genuine friends. Most of a network of 500, 1,000 or 5,000 connections sits outside that, effectively invisible unless something jogs your memory. That's not a discipline problem, and no amount of scrolling fixes it.

What changes with AI network search

Query instead of recall. You ask "who works in fintech" instead of trying to remember who works in fintech — the system checks every connection, not just the ones top of mind.
Ranking instead of a flat list. Results come back ordered by relevance and relationship strength, not in whatever order LinkedIn happens to display connections.
Reasoning attached to each result. A good network search tool tells you why someone matched — shared company, recent activity, mutual connections — so you know how to reference it in outreach.
Signals you'd otherwise miss. Job changes, promotions, and recent posts across your entire network are easy to miss scrolling, but straightforward to surface with a standing search.

What doesn't change

AI network search doesn't replace judgment about who to actually message or how to phrase the ask — it just removes the bottleneck of finding the right person in the first place. The final outreach decision, tone, and relationship-building is still on you.

When manual scrolling is still fine

For very small networks (under a couple hundred connections) or when you already know exactly who you want to reach, manual search is often faster than setting up any tool. AI network search earns its keep once your network is large enough that memory alone can't cover it — which, for most professionals with several years on LinkedIn, is sooner than they'd guess.

Trying it on your own network

People by Sniff is built specifically for this: connect your LinkedIn account, ask a question in plain English, and get back a ranked list instead of scrolling.

Sources

Hill & Dunbar, Social network size in humans (Human Nature, 2003)
Dunbar, Do online social media cut through the constraints that limit the size of offline social networks? (Royal Society Open Science, 2016)