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

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

Compiled by: Zen, PANews

On April 11, Beijing time, AI startup Plastic Labs announced the completion of a $5.35 million Pre-Seed funding round, led by Variant, White Star Capital, and Betaworks, with participation from Mozilla Ventures, Seed Club Ventures, Greycroft, and Differential Ventures. Angel investors include Scott Moore, NiMA Asghari, and Thomas Howell. Meanwhile, its personalized AI identity platform “Honcho” has officially opened for early access.

Due to the project’s early stage, the entire crypto community knows very little about Plastic Labs. Meanwhile, as Plastic announced the above financing and product updates through X, its main investor Variant’s chief advisor and investment partner Daniel Barabander also provided an in-depth interpretation of the project and its Honcho platform. The following is the original text:

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 depending on the conversation partner – the wording when explaining mathematical concepts to your grandparents is entirely different from that used when explaining to your parents or children. You instinctively adjust your expression based on your audience, and LLM applications must similarly “understand” who they are conversing with in order to provide a more effective and tailored experience. Whether it’s a therapeutic assistant, a legal aide, or a shopping companion, these applications need to truly understand the user to deliver value.

However, despite the critical importance of personalization, there are currently no ready-made solutions available on the market for LLM applications to call upon. Developers often have to build various fragmented systems themselves, storing user data (usually in the form of conversation logs) and retrieving it when needed. The result is that each team has to reinvent the wheel, building their own user state management infrastructure. Worse still, methods such as storing user interactions in vector databases and performing retrieval-augmented generation (RAG) can only recall past conversations and cannot truly grasp deeper characteristics of the user, such as their interests, communication preferences, and tone sensitivity.

Plastic Labs presents Honcho, a plug-and-play platform that enables developers to easily personalize any LLM application. Developers no longer need to build user modeling from scratch; by integrating Honcho, they can instantly obtain rich and enduring user profiles. These profiles are more nuanced than traditional methods, thanks to the team’s adoption of advanced techniques from cognitive science; moreover, they support natural language queries, allowing LLMs to flexibly adjust their behavior based on user profiles.

By abstracting away the complexity of user state management, Honcho opens up new heights of hyper-personalized experiences for LLM applications. But its significance goes far beyond that: the rich abstract user profiles generated by Honcho also pave the way for the long-sought “shared user data layer.”

Historically, the failure of shared user data layers can be attributed to two main reasons:

1. Lack of Interoperability: Traditional user data is often highly dependent on specific application scenarios, making it difficult to migrate across apps. For example, social platform X may model based on the people you follow, but this data is of no use for your professional network on LinkedIn. Honcho, on the other hand, captures higher-level, more universal user traits that can seamlessly serve any LLM application. For instance, if a tutoring application discovers that you are best suited for learning through analogy, then your therapy assistant can also leverage this insight to communicate with you more effectively, even though the scenarios are completely different.
2. Lack of Immediate Value:* Historically, shared layers struggled to attract applications early on because they didn’t deliver tangible benefits to the first movers, who were the key to generating valuable user data. Honcho is different: it solves the “first-level problem” of managing the user state of a single application, and when enough applications are connected, the network effect naturally solves the “second-level problem” - the new application is not only connected for personalization, but also leverages existing shared user personas from the start, completely eliminating the pain point of cold starts.

Currently, Honcho has hundreds of applications on the closed beta waiting list, covering various scenarios such as addiction coaching, educational companions, reading assistants, and e-commerce tools. The team’s strategy is to first focus on solving the core challenge of user status management for applications, and then gradually launch a shared data layer for willing applications. This layer will adopt encrypted incentives: early-access applications will receive ownership shares of this layer, allowing them to share in its growth dividends; at the same time, the blockchain mechanism will ensure that the system is decentralized and trustworthy, alleviating concerns about centralized entities extracting value or developing competing products.

Variant believes that the Plastic Labs team has the capability to tackle the user modeling challenges in LLM-driven software. The team experienced firsthand the issue of applications not truly understanding students and their learning styles while developing the personalized chat tutoring application Bloom. Honcho was born out of this insight and is addressing the pain points that every LLM application developer will face.