The hiring model that can't be vibe coded
AI didn't start with ChatGPT. Anti-fraud used graph theory and behavioral signals to detect fraudsters with over 90% accuracy — even when those fraudsters were actively lying. Quant trading used statistical models, reinforcement learning, and time series forecasting. Self-driving cars used sensor fusion — combining computer vision, LiDAR, and radar — alongside path planning algorithms. Each field built specialized tools for specialized problems.
Nowadays when you say you use AI, people assume LLMs.
You would be loathe to use an LLM for any of those use cases. In fact, you shouldn't. A self-driving car that hallucinates a pedestrian is a different category of problem than a chatbot that hallucinates a fact. Higher stakes demand the right tool.
And yet we get asked: "Why can't I just vibe code this?"
The answer is that we've built a specialized model from the ground up. That can't be vibe coded. It's deep research.
The problem with language
Companies are drowning in applicants: thousands per role, half of them AI-generated and indistinguishable. That's the problem Tied is built to solve — cutting through the noise to surface the five that actually matter.
The tool for that job isn't an LLM. Because candidates are incentivized to perform. To say what interviewers want to hear. That's why so many YC founders pivot the minute they get in: they said whatever it took to get through the door. LLMs are trained to process language. They're extraordinary at it. But that's also their weakness. A well-coached answer or a lie looks great in text.
The problem is that most of communication isn't verbal. It's behavioral. A slight condescension in tone can flip the meaning of a sentence. Anti-fraud learned this early: the question was never purely what did they say? — as the fraudster attempts to persuade you to reinstate their access. It was how do they behave? Those two things are very different data sources.
As time goes on and incentives warp, we believe VCs and the tech industry will experience more and more fraud. 4 in 10 candidates now use AI in their job applications — to generate resumes (54%) and cover letters (50%). The Financial Times reports that nearly half have used AI to embellish or outright falsify their credentials. So how do you know who you're actually getting?
Why graph theory
We draw from anti-fraud and behavioral science — except instead of scoring whether someone is a fraudster, we're scoring whether they're the type of candidate a given client has historically succeeded with. Same class of problem, same methodology, different domain.
A person is not a data point. They're a dynamic system of behaviors over time. Graph theory lets us encode not just individual attributes but relational patterns — how clusters of behavior form, how those clusters map to outcomes, who someone actually is versus who they say they are on a resume.
Domain knowledge as a competitive moat
LLMs are generalists. Human judgment — in talent, in investment — is domain-specific. A good engineer might not make a great salesperson. We encode that knowledge deliberately rather than hoping the model figures it out. Every variable earns its place, every prediction is traceable, and the model sharpens with targeted feedback.
And because we work from behavioral signals, not resumes or self-reported demographics, we don't use race, gender, or age. Excellence surfaces on its own terms.
Crunchbase tells you what founders have done. We tell you who they are. Tied is the people intelligence layer for hiring in the era of AI.
Sources
- Cristina Criddle and Delphine Strauss, Financial Times — Jobhunters flood recruiters with AI-generated CVs (Aug 12, 2024)