Building a Knowledge-Based Network: How Experts Scale Without Replacing Themselves

LinkedIn solved one problem: connecting people by what they’ve done. You search for someone by the schools they attended, the companies they worked for, the credentials they’ve accumulated. It’s a professional identity card, a connection layer for people who share similar pedigrees.

But credentials don’t tell you how someone thinks.

That’s where knowledge-based networks differ fundamentally. Instead of connecting nodes by shared work history or mutual contacts, they connect people by similarity of thought, approach to problems, or how well one person’s knowledge fills a gap in another’s.

Dara Ladjevardian sees this as the next layer in how experts scale. Not by delegating to assistants or hiring a sales team, but by making their thinking accessible at scale—and building a network where people find experts not by reputation but by compatibility.

“LinkedIn is a professional network where the nodes are connected by who you know and who you’ve shared work experiences with,” Dara explains. “Where Delphi is a knowledge-based network where the nodes are actually connected by similarity of thought or similarity of knowledge, or someone’s knowledge actually filling a gap in your own mind.”

The Limits of Credential-Based Networks

LinkedIn’s model works if you’re hiring for a job. You search for someone with the right title, company, and skills, and you assume they can do the work. But it’s a poor model if you’re trying to learn or get advice.

A CEO might have the exact credentials you’re looking for. But does she think about company culture the way you do? When you ask her about scaling from 10 to 100 people, will her answer feel like a template she’s used a hundred times, or will it reflect something unique about how she thinks? You have no way to know until you talk to her. And if she’s well-known, she might not have time.

A knowledge-based network flips this. You’re not searching by credentials. You’re searching by what you need to know and then finding people whose thinking is aligned with your approach or whose perspective complements your own.

Scaling Expertise Through Personalization

The scaling challenge for experts is real. An investor with unique insights could spend eight hours a day on calls with founders seeking advice. A founder could answer the same questions about hiring and fundraising endlessly. An academic could give the same lecture explanation 100 times a semester instead of creating it once.

“The product works the way it is, either you’re happy with it or you’re not,” Dara says about Delphi’s mission. The always-on availability of an expert—their thinking, their framework, their perspective—available 24/7 without them having to be present.

This is different from a chatbot. A chatbot can answer FAQ questions about a company or a topic. But Delphi represents the person. When you talk to an expert’s Delphi, you’re getting their perspective personalized to your specific situation, grounded in their actual experience and reasoning patterns.

For someone with high inbound demand—a founder getting investor questions, an investor evaluating founders, a consultant fielding client requests, a domain expert being asked the same thing repeatedly—this is transformative. You’re not hiring a customer service team. You’re scaling your own thinking.

The Network Effect of Knowledge Similarity

As Delphi scales, the magic happens at the network layer.

Imagine millions of people creating knowledge representations of themselves. Your own Delphi learns who you are and how you think. The platform can then recommend people for you to learn from based on similarity: “Here are five people who think like you do.” Or based on complementarity: “Here are five people whose thinking is different from yours in ways that might stretch you.” Or based on gap-filling: “You’re interested in X, but have no expertise in Y. Here are five people who are strong in Y.”

This recommendation system doesn’t exist in LinkedIn. LinkedIn recommends connections based on proximity and mutual connections. A knowledge network can recommend based on actual cognitive similarity or gap-filling potential.

“That’s the magic of the knowledge-based network where you have a Delphi, you have a digital mind, and based on that, we can say, here are five people who think similarly to you. Here are five people who can help you achieve your goals. Here are five people whose opinions actually contradict your own, and we think it would be good for you to explore the other viewpoint,” Dara explains.

The Competitive Dynamics

One question naturally arises: if Delphi gets very good at this, won’t it become a market for expertise that competes with direct consulting, coaching, or mentorship?

Yes. But Dara doesn’t see Delphi as replacing those things. Instead, it’s a layer underneath them.

The flow would work like this: You explore someone’s Delphi. You get value from conversing with their knowledge representation. If you want deeper engagement, you can reach out to them directly on LinkedIn. Your Delphi has already done the heavy lifting of showing you how they think. The subsequent 1-on-1 is a premium upgrade on a foundation of compatibility you’ve already established.

From the expert’s perspective, Delphi is not a replacement for their advisory business. It’s a funnel. It handles the repetitive questions and the low-touch interactions. It surfaces the people who are genuinely aligned with them (or who their thinking can help). Those are the people they choose to spend time with.

“LinkedIn, if anything, could be complementary to Delphi,” Dara suggests. LinkedIn stops being the place where you signal your expertise and becomes the place where you deepen relationships that started on Delphi.

Why Building This Now Matters

The network effect takes time to manifest. You need enough experts with high-quality knowledge representations that the recommendation engine has signal. You need enough seekers that exploring the network is rewarding. Right now, Delphi has celebrities and well-known investors and founders. The network is starting to show this benefit at the top of the curve.

But “we’re not going to focus too much on the network until probably like March or April of next year,” Dara says. First, the product needs to work for individual experts. The network scales after you’ve proven the core use case.

This sequencing is wise. Many platforms launch with network features before the single-player experience is solid. But a knowledge graph is only valuable if it accurately represents the person. Getting that right matters before you invest in discovery and recommendation.

Once that foundation is solid, the network effects multiply. An expert’s knowledge becomes more discoverable. Seekers find better matches. Experts get better insights into what questions they’re getting and where the gaps are in how they communicate their thinking.

FAQ

Is Delphi building a replacement for LinkedIn?

No. LinkedIn is a professional network. Delphi is a knowledge network. They serve different functions. LinkedIn is where you signal your work history and credentials. Delphi is where you share your thinking and reasoning. They’re complementary.

How is a knowledge network different from a community built around an expert?

Communities are often centered on one expert or brand. A knowledge network connects many people’s expertise directly. Instead of a community member asking the expert a question in a forum, they ask that expert’s Delphi, and they might also discover other experts whose thinking is relevant to their interest.

Won’t AI recommendation engines get biased? What if they recommend people who just sound like me instead of people who challenge me?

Dara’s point is that Delphi can do both. It can recommend people who think like you, and it can also recommend people whose thinking contradicts yours. The choice is in the algorithm design. The knowledge graph gives you the data to do either.

How does discovery work if there are millions of Delphis?

That’s the hard problem Delphi hasn’t fully solved yet. Right now, people find Delphis through the explore page or direct links. As the network grows, recommendation engines will become critical. This is why they’re explicitly not focusing on the network layer until the individual product is solid.

Can you game the recommendation system by crafting a specific knowledge graph?

Theoretically, yes. But Delphi’s data comes from real content: your tweets, your podcasts, your essays, your interviews. You can’t fake your thinking over years of public content. Short-term gaming is possible, but long-term authenticity is harder to fake.

What about people who want privacy? Do you have to make your thinking public?

No. You choose what goes into your knowledge graph. You can be selective about which podcasts, tweets, and essays you include. But if you want the network to recommend people to learn from you, you need enough signal. Total privacy and total visibility are both valid choices.

Can knowledge networks work for non-experts?

Yes. Dara’s definition of expertise is broad: “founders, investors, consultants, domain experts, content creators.” Someone with niche knowledge about a very specific field can create a Delphi. The question is whether there’s enough inbound demand to make the always-on availability useful.

Will knowledge networks eventually replace expert marketplaces like Maven or education platforms?

Unlikely fully, but they might consume a portion. A knowledge network is better for discovery and asynchronous learning. Marketplaces are better for curation and trust-building in the early days. Education platforms are better for structured curriculum. Different functions for different use cases.

How do you ensure quality in a knowledge-based network?

Delphi’s human verification and data curation help. You can only create a Delphi if you’re a verified human, and your knowledge graph is built from data you choose. This is different from platforms that scrape the internet. Higher barrier to entry = higher quality network.

Can experts use Delphi as a sales channel?

Absolutely. If your Delphi surfaces you to people who need what you offer, those conversations become warm leads for your services. But you have to provide genuine value in the always-on representation for this to work. People can tell if you’re using it as a funnel versus actually sharing your thinking.