Quick view of personalized AI identity platform Honcho: How to enable a super personalized experience for LLM applications?

Original author: Daniel Barabander , Variant General Consultant & Investment Partner

Compiled by: Zen, PANews

On April 11, Beijing time, AI startup Plastic Labs announced that it had completed a US$5.35 million Pre-Seed round of financing, led by Variant, White Star Capital and Betaworks, with Mozilla Ventures, Seed Club Ventures, Greycroft and Differential Ventures participating. Angel investors include Scott Moore, NiMA Asghari and Thomas Howell. At the same time, its personalized AI identity platform "Honcho" has officially opened early access.

Since the project is still in its early stages, the entire crypto community knows little about Plastic Labs. While Plastic released the above financing and product updates through X, Daniel Barabander, general consultant and investment partner of its main investor, also gave an in-depth interpretation of the project and its Honcho platform. The following is the original content:

With the rise of large-scale language model (LLM) applications, the demand for personalization in software has grown unprecedentedly. Such applications rely on natural language, which varies according to the person you are talking about—just like the wording you explain mathematical concepts to your grandparents, which is very different from when you explain them to your parents or children. You will instinctively adjust your expressions to your audience, and LLM applications must also "understand" who they are talking to in order to provide a more effective and more appropriate experience. Whether it is a healing assistant, legal assistant, or shopping companion, these applications require a real understanding of users in order to achieve their value.

However, although personalization is crucial, there are currently no ready-made solutions available for LLM applications to call. Developers often have to build various fragmented systems themselves, store user data (usually in the form of session logs) and retrieve them when needed. The result is that every team has to re-create the wheel and build its own user status management infrastructure. What's worse is that methods like storing user interactions into vector databases and doing retrieval enhancement (RAG) can only recall past conversations, but cannot truly grasp the deep characteristics of the user's own interests, communication preferences, and tone sensitivity.

Plastic Labs brings Honcho, a plug-and-play platform that allows developers to easily personalize any LLM application. Developers no longer need to build user modeling from scratch. By integrating Honcho, they can immediately obtain rich and lasting user portraits. These portraits are more delicate than traditional methods, thanks to the team's advanced technology in cognitive science; and they support natural language querying, allowing LLM to flexibly adjust its behavior based on user portraits.

By abstracting the complexity of user state management, Honcho opens a new level of super personalized experience for LLM applications. But it has much more than that: the rich abstract user portraits generated by Honcho also paves the way for the long-standing “shared user data layer”.

Historically, there are two main reasons why the sharing user data layer failed:

1. Lack of interoperability : Traditional user data is often highly dependent on specific application scenarios and is difficult to migrate across Apps. For example, social platform X may model based on the people you follow, but this set of data will not help you in your career network on LinkedIn. Honcho captures higher-level and more versatile user traits and can seamlessly serve any LLM application. For example, if a tutoring app finds you best to learn with analogies, your therapy assistant can also use this insight to communicate with you more effectively, although the two scenarios are very different.
2. Lack of instant value : Previous sharing layers were difficult to attract application access in the early stages because they did not bring substantial benefits to pioneers, which were the key to generating valuable user data. Honcho is different: it first solves the "first-level problem" of user status management of a single application. When enough applications are connected, the network effect will naturally lead to the solution of "second-level problem" - new applications are not only connected for personalization, but also use existing shared user portraits from the beginning to completely eliminate the pain points of cold start.

At present, Honcho has hundreds of applications on the closed-beta waiting list, covering a variety of scenarios such as addiction quit coaches, educational partners, reading assistants and e-commerce tools. The team 's strategy is: first focus on solving the core problem of user status management of the application, and then gradually launch a shared data layer for applications that are willing to participate. This layer will adopt cryptographic incentives: early access applications will gain ownership share of the layer, thereby sharing its growth dividends; at the same time, the blockchain mechanism can also ensure the system is decentralized and credible, eliminating the concerns of centralized institutions to extract value or develop competitors.

Variant believes that the Plastic Labs team has the ability to overcome user modeling challenges in LLM-driven software. When developing Bloom, a personalized chat tutoring app, the team experienced firsthand the problem that the app couldn’t really understand students and how they learn. Honcho was born based on this insight and is solving the pain points that every LLM application developer will face.