Exploring Personal AI Agents with Div Garg of MULTI-ON
AI that can book your travel and buy you a coffee
The Deload Podcast interviews CEOs, founders, and builders of frontier technology companies that are transforming how we live. The mission of the podcast expands that of The Deload: To educate you about where the world is going to make smarter growth investments. Now with conversation.
If you haven’t subscribed, join more than 1,500 rational optimists:
Personal AI Agents
Div Garg is the CEO and founder of MULTI-ON, a company that builds Personal AI Agents.
Here are my notes from my conversation with Div:
An AI agent is a software program that can interact with an environment (real world, the digital world, other agents, etc.) given some human instruction. An example might be a person who asks an agent to book a vacation where the agent would be expected to find flights that fit into the person’s schedule, book preferred hotels, and find suitable activities.
As these AI Agents develop, we’ll move from today’s Level 2 or 3 style of autonomy where humans are consulted by default to Level 5 where agents are able to act with complete independence. Div mentioned that starting with a human-in-the-loop of an agent’s process both helps humans develop trust with the agent by being part of the process as well as improves the agent’s abilities via feedback from the human.
MULTI-ON marries capable AI Agents with personalization as its key differentiator. We’ve talked about the trillion dollar opportunity in Personal AI in prior posts. For Personal AI to reach that potential, the AI needs to be able to act on your behalf. As these systems learn about users, they will be more efficient in understanding requests with less specificity. “Get me my coffee” will turn into a very specific idea for your Personal AI based on your typical habits.
Div mentioned that users seem to be willing to pay $40-100/month, in-line with the survey work I’ve done.
Privacy is a critical element to trust with an AI Agent. MULTI-ON handles sensitive information by storing it encrypted on the client side and giving control of what’s remembered by the AI to the user.
AI Agents are the right stepping stone to artificial general intelligence. MULTI-ON wants to build something like Jarvis in their three-year vision. A secure, reliable, personal friend.
MULTI-ON is launching an API for developers to build on top of their agent platform.
Learn more about MULTI-ON and sign up for the beta:
Episode Time Stamps
2:38 Div’sAI Agent definition
4:50 The intersection of AI Agents and Personal AI
6:40 Agent use cases
9:00 How MULTI-ON works with a travel example
10:10 Level 2 to Level 5 AI Agents
11:58 How MULTI-ON handles security and privacy
15:55 Enterprise vs individual use cases
19:32 AI Agents on the road to AGI
Disclaimer: My views here do not constitute investment advice. They are for educational purposes only. My firm, Deepwater Asset Management, may hold positions in securities I write about. See our full disclaimer.
Intelligent Indices Update: ChatGPT vs the S&P 500
I’m tracking a series of AI-powered stock indices vs the S&P 500 and Nasdaq 100. They’re Intelligent Indices.
The flagship Intelligent SP Select (AISPS) is a custom weighted large cap index built by a committee of ChatGPT, Bard, and Claude. The aim is to provide a superior market benchmark to the S&P 500 over the long term and the definitive benchmark for AI-informed investment strategies.
The AISPS gained ground vs the S&P 500 over the past week, particularly as stocks corrected on Wednesday with the US credit downgrade by Fitch.
To the extent markets continue to revert, the AISPS is well positioned relative to the S&P 500 with heavier exposure to healthcare and consumer staples names. The AISPS is particularly underweight higher-multiple tech like AAPL, TSLA, and NVDA.
AI will disrupt all information industries, stock indices included. The next great stock index will be AI-powered, like the AISPS.
You can follow all of the Intelligent Indices on Thematic.