A new startup, Supermemory, has raised $3 million in seed funding. Dhravya Shah, a 20-year-old from Mumbai, founded the startup. The funding round was led by Susa Ventures, Browder Capital, and SF1.vc. Several well-known investors including Jeff Dean from Google AI and Logan Kilpatrick from DeepMind took part in the funding.
Building AI with Long-Term Memory
Supermemory is developing a system to provide AI with long-term memory. Most of today’s AI models are good at responding to questions. However, they tend to forget previous conversations. Supermemory is building a system that gives AI a long-term memory. This $3 million funding marks a big moment for Dhravya and his team.
The Technology Behind Supermemory
Supermemory develops a system that stores information from previous interactions with the user and utilizes it when required. The primary objective of supermemory is to create AI tools that are capable of remembering previous conversations, user preferences, and contextual information from earlier sessions.
The technology behind Supermemory operates in conjunction with large language models such as GPT and Claude. It provides a way to connect to these models, allowing them to save and retrieve information in a secure manner. This makes the technology inherently smarter and more reliable.
The Memory Engine
Supermemory is working towards incorporating an external memory engine for AI models. The memory engine runs outside the main model, thereby ensuring the safety of all data in a highly regulated structure. When the model needs to recall information, it sends a query to Supermemory. The memory engine retrieves relevant information and returns it to the conversation. This setup keeps the memory flexible and easy to expand. This compatibility makes it easier for developers to integrate with different AI models.
Data Safety and Efficiency
One of the key features of Supermemory is data safety. Since memory involves storing personal user information, Supermemory has developed strict access controls and encryption techniques. However, it also enables users to have control over the information that gets stored and what can be forgotten. Supermemory’s intent is also to have its technology use the least possible energy. Long-term storage and recall can be resource-heavy. The team is building systems that optimize cost and computing power.
Supermemory has also been building a developer API. This allows other AI startups to plug into its system easily. By opening access, the company is hoping to become an infrastructure for AI memory management by default.
The Future of Human-Like AI
Supermemory is part of a larger trend wherein AI companies are focusing on intelligence that feels more human. Memory, reasoning, and learning are the emerging frontiers. The company’s work may lead to more personalized and stable AI applications in everyday life. Whether it’s digital tutors, virtual assistants, or health coaches, memory is what will make AI useful, but also something we can empathize with.
This startup marks the beginning of a new wave of a young founders from India building global AI products that push the limits of technology.
Also read | Anthropic Expands Worldwide, but All Eyes Are on India
Stay tuned with BharatLinkr for the latest updates from India’s startup ecosystem.
