AsiaTechDaily – Asia's Leading Tech and Startup Media Platform
Recently, Mindverse, an AI startup based in Singapore, successfully secured $5 million in seed funding from Square Peg, a venture capital firm. The capital infusion comes at a crucial stage for the company as it plans to advance its large AI personalization model when the current market is driven by one-size-fits-all GenAI applications.
Founded in 2022 by former executives from Meta and leading AI researchers, Mindverse specializes in AI technology. Their flagship products include MindOS Studio, facilitating the creation of AI-powered websites with personalized chat interfaces, and Mebot, a productivity tool adept at learning and recalling user preferences.
In an exclusive interview with AsiaTechDaily, Kisson Lin, the Co-Founder and COO of Mindverse AI sheds light on the inception of Mindverse and the motivation behind its founding.
With a background as the Monetization Strategy Director for TikTok and previous roles at Meta as a Corporate Strategist and Partner Manager, Lin has a deep understanding of the tech industry’s nuances and dynamics.
In the interview, Lin delves into the company’s flagship products, MindOS Studio and Mebot, offering insights into how these offerings contribute to Mindverse’s mission of providing diversified and personalized services to users worldwide. She also discusses the strategic partnership with Square Peg and outlines the company’s immediate priorities following the recent funding round.
How did the idea for Mindverse come about, and what motivated you to co-found the company?
I’m always intrigued by ‘paradigm shift’. While working at Meta, I became fascinated about AI, and believed that would be the next paradigm shift. Then when I met my Co-founder Dr. Tao, I knew it’s time.
As one of the most visionary tech leaders I personally know, he had been advocating for AGI, and developing ‘Agent framework’ on top of LLMs in 2020, at a lab he founded. When we discussed how AI agents could be deployed to game characters and robots, an idea suddenly came to our mind: what about apps?
Agents could be a more intuitive way of interaction. By combining graphic interface and language interface (chats), using apps will be as easy as talking to a person who can also project images and videos in real time.
For example, a fitness studio can launch its own agent. When users want to exercise, they no longer need to click to filter videos, but directly interact with the fitness agent, which can guide users on a 1:1 basis. We envisioned that in an AI-native world everyone would have their own personal AIs who understands them well. In the meantime, there’d be many professional agents in the agent store, like today’s app store, to serve specific needs. With that vision, we named our first product MindOS, thinking that such an agent creation platform would be an AI-native operating system.
ADI, or Artificial Diversified Intelligence, stands in contrast to the more traditional one-size-fits-all AI solutions. How does Mindverse’s ADI approach enhance user experiences compared to these standardized models, and what specific benefits does it bring in terms of AI personalization and adaptability?
In July last year we launched MindOS globally, with more than a hundred user generated AI agents on the platform. We heard a lot of great feedback and feature requests. The most popular ones are: ‘Can my agent get smarter through conversations?’ ‘I want to ensure that my personal data is stored locally.’ ‘I want my agent to remember what customers say, and analyze it for me.’ We realized that a one-size-fits-all AI solution wouldn’t work.
The definite trend is to build an AI personalization model on top of LLMs, so the former can store personal profiles and long term memories, and provide customized services to every user.
LLMs are like the frontal cortex of our brains, and to make us human, we have another memory section that remembers our friends’ profiles, ourselves, etc. Memory is one of the most important factors that differentiate us from each other. That’s why we started building the Large Personalization Model, and advocating for Artificial Diverse Intelligence. The benefits are substantial: Our model understands each user as a person, and provides more personalized services to each one. It can also continue to learn through the conversations. Moreover, in the future, we plan to bring this AI personalization model on-device.
Here’s an example of how ‘personalized services’ can be in the AI era. If you ask AI to remind you of today’s schedule, LLM alone won’t be able to do that. With API, LLM may say, here’s your schedule, blah blah blah. But with the personalized model, it doesn’t only tell you the schedule, but also prepares you for the meetings, with all the relevant context from your email threads, from your previous meeting notes, from research data, etc. This sounds like a real assistant, and it’s exactly what AI can do with AI personalized models.
Mindverse aims to pioneer Artificial Diversified Intelligence (ADI) with its Large Personalization Model. Can you elaborate on what ADI entails and how it differs from other AI approaches?
We build on top of LLMs. LLMs alone do not have memory of each user. To add memory to LLMs, an external database is required. But that’s still not enough. Our brains are so efficient in storing and retrieving information because we compress information and connect the dots. Our model mimics the brain’s approach to abstract and rank information, so our memories are more accurate and efficient, and it’s also small, so everyone can have their own personalized model, which is compatible with multiple LLMs. Think about storing all your files in your own USB, and you can plug in your USB to any device.
Could you give us an overview of Mindverse’s flagship products, MindOS Studio and Mebot, and how they contribute to the company’s mission?
MindOS Studio enables businesses to build professional AI agents and deploy to their websites as chatbots. But they are more than just chatbots. They can craft dynamic interfaces on demand, with the most relevant information the user requested. For instance, a travel agency may resemble travel plans, hotel packages and a map for the user on demand, rather than letting the users click through several sections on the website to gather the plan themselves. Such personalized service does not only require the business information, but also requires the deep understanding of user preferences through the chats. And this is how our personalization model can empower.
On the individual front, Mebot is a note-taking app. It offers a deeply personalized experience, learning and memorizing user preferences and thoughts. Think about you sending messages to yourself when you come across interesting thoughts or links and don’t yet have time to put them into organized notes, now AI can organize, remember, and retrieve the notes for you. In the future, Mebot will not only be a second brain, but also a personal assistant.
Providing diversified and AI personalized services to everyone is our mission. And we believe that through MindOS Studio and Mebot, whenever people use personal AI or access a business website, their experience will be redefined.
How do you envision collaborating with Square Peg and tapping into their expertise and network to support Mindverse’s goals?
Square Peg is an early investor of many great companies like Canva and Fiverr. They are very experienced in helping SaaS companies grow. They have provided great support to us in terms of product insights and resources. Their team had been sharing with us product feedback, and also feedback using our competitors’ products; they constantly connect us with other investors for the next fundraising, and with other founders for business partnerships. We are so grateful to work with them.
(Also) Square Peg is one of the most founder-friendly investors we’ve ever met. They looked to back the real game-changers, and they invested for long term success. They value the founders, the technology, and the value we bring to people, more than whether we can make profit immediately upon investing.
What are some immediate priorities for Mindverse following the recent funding round, and how do they align with the company’s long-term goals?
We’ve now built great technology, and validated the value of our products. With this recent funding round, we aim to aggressively grow our user base this and the next year.
Can you share any product developments or initiatives that we can expect from Mindverse in the near future?
I’m most excited about the development of Mebot. We will introduce a memory palace for the Mebot library, so you can see how your notes are connected with each other, and how your memory grows day by day. We are also building the browser extension and multiple API connections, so Mebot can be accessed easily through different channels.
Whenever you come across articles or webpages, you can share with Mebot, and Mebot will return to you daily reading comprehensions. There are many other exciting features which will be launched in the next months. We aim to make Mebot the most intuitive, easiest app to record your precious memories and thoughts, in all formats.
How do you foresee the adoption and reception of ADI technologies evolving in the Asian market, particularly given the region’s diverse cultural landscape and rapidly growing digital economy?
That’s a very good question. We are confident that ADI is the future, especially for our diverse cultural landscape here. One interesting aspect, among many others, is that Asia is more mobile native, and with many different languages. So ‘speaking’ or taking photos are easier for many people in this region, than typing long messages.
Our model is great at understanding voice memos and images, and building personal profiles of user preferences and cultural nuances based on the multimodal input. We can’t wait to share our product with more people in the Asian market.
This recent funding injection will enable Mindverse to further enhance its research and product development endeavors, particularly in advancing its large personalization model. With a focus on Artificial Diversified Intelligence (ADI), Mindverse aims to personalize AI experiences to accommodate the diverse range of human thoughts and values.
Also Read: