Convergence raises $12M for AI agents with long-term memory

Post Thumbnail

Startup Convergence has announced raising $12M in investment for developing an innovative platform of universal AI agents with long-term memory. The pre-seed round was led by venture fund Balderton Capital with participation from corporate investors Salesforce Ventures and Shopify Ventures.

The company’s founders, CEO Marvin Pourtorab and CTO Andy Thulis, have impressive experience in artificial intelligence. Before creating Convergence, they worked on developing recommendation systems and AI assistants at Shopify, and then joined the Cohere team, specializing in corporate AI solutions.

“Honestly, we left [Cohere] because we had been experimenting with agents for several years already, about three to four years,” explains Pourtorab, noting that the market wasn’t ready for such technologies before.

The company’s key product is the universal agent Proxy, using Large Meta Learning Models (LMLM) technology. Unlike existing solutions focused on highly specialized tasks, Proxy can adapt to various workflows thanks to a “long-term memory” system that allows it to acquire skills similar to humans.

“There are many companies in the market now creating ‘narrow’ agents: for sales, HR, or financial operations. We chose a different approach, developing the foundation for the first class of universal agents that can adapt to specific user needs. We don’t see a future where everyone will have a thousand different specialized tools. Over time, consolidation will occur, and this is where we’ll find our niche,” Pourtorab comments on the company’s strategy.

To implement ambitious plans, Convergence is actively recruiting talented specialists from leading technology companies, including Google DeepMind, Meta, OpenAI, and PolyAI. James Wise, partner at Balderton Capital, emphasizes the founders’ unique competencies: “Few possess Marvin and Andy’s experience and skills, making them ideal candidates for solving complex technical challenges in creating a product at Proxy’s level.”

The raised investments will be directed towards developing innovative models underlying Proxy assistants. The company plans to create technology that will allow artificial intelligence to effectively perform routine tasks, freeing up employees’ time for strategic decision-making.

AIvengo > Reviews > Convergence raises $12M for AI agents with long-term memory
Почитать из последнего
UBTech will send Walker S2 robots to serve on China's border for $37 million
Chinese company UBTech won a contract for $37 million. And will send humanoid robots Walker S2 to serve on China's border with Vietnam. South China Morning Post reports that the robots will interact with tourists and staff, perform logistics operations, inspect cargo and patrol the area. And characteristically — they can independently change their battery.
Anthropic accidentally revealed an internal document about Claude's "soul"
Anthropic accidentally revealed the "soul" of artificial intelligence to a user. And this is not a metaphor. This is a quite specific internal document.
Jensen Huang ordered Nvidia employees to use AI everywhere
Jensen Huang announced total mobilization under the banner of artificial intelligence inside Nvidia. And this is no longer a recommendation. This is a requirement.
AI chatbots generate content that exacerbates eating disorders
A joint study by Stanford University and the Center for Democracy and Technology showed a disturbing picture. Chatbots with artificial intelligence pose a serious risk to people with eating disorders. Scientists warn that neural networks hand out harmful advice about diets. They suggest ways to hide the disorder and generate "inspiring weight loss content" that worsens the problem.
OpenAGI released the Lux model that overtakes Google and OpenAI
Startup OpenAGI released the Lux model for computer control and claims this is a breakthrough. According to benchmarks, the model overtakes analogues from Google, OpenAI and Anthropic by a whole generation. Moreover, it works faster. About 1 second per step instead of 3 seconds for competitors. And 10 times cheaper in cost per processing 1 token.