This startup wants to optimize your entire life with its new ‘proactive’ AI

Daniel George, the founder and CEO of AI company TwinMind, has quite the resume.

He was a part of a Nobel Prize-winning team that worked on using artificial intelligence (AI) to detect gravitational waves and black holes. He also worked on AI projects at companies like Wolfram and Google X. Perhaps most notably, he created an AI tool that automated his own job as a VP at JPMorgan, allowing him to spend a few years traveling the globe.

And now, George counts himself as a startup founder, and he’s bringing his latest project to the masses: TwinMind, an AI platform that “listens to everything you do, say, and browse, and it’s all stored locally and encrypted,” he says.

TwinMind can be downloaded as an app on an iPhone and functions like “Jarvis in your pocket.” 

‘Proactive’ AI that runs your life?

Effectively, users turn it on and can leave it running all day (if they choose to). The app listens to and digests the user’s surroundings, and produces end-of-day reports or wrap-ups, capturing things that you may have forgotten or missed. Everything is processed as it’s ingested and transcribed on the fly.

George claims that this helps with security issues, as audio isn’t actually stored or saved anywhere. “All that is saved are the final outputs of the model, for privacy considerations,” he says.

For iPhone users, the mic icon will always display when the app is listening and transcripts don’t actually identify “who said what words exactly.”

Given how the app functions, however, it does seem possible that people could be caught in the background as the app is running without knowing it. TwinMind says it encourages users to ask for consent according to local laws.

TwinMind runs as an app on existing hardware and stores information locally, then uses what it learns to be “proactive” and surface relevant information for users when and as needed.

Attracting users and Silicon Valley investors

The app launched in late March and has already attracted thousands of users through word-of-mouth, George says.

It has also attracted the attention of Silicon Valley investors, including Anand Rajaraman, Dan Roth, and Michael Liou, who collectively were some of the earliest investors in companies like Facebook, Robinhood, and others. Data from PitchBook clocked TwinMind’s valuation at $30 million last year, while George says its latest valuation is $50 million.

George and his two cofounders, CTO Sunny Tang and chief scientist Mahi Karim, all met working at Google X. He says they’re now living together in a Bay Area house while they build and scale TwinMind, “working 100-hour weeks for the past year-and-a-half.”

The app works in more than 100 languages and, according to George, can run for more than 12 hours without sapping a smartphone’s battery, as it uses an LLM to process speech directly and produce a daily “memory,” or rundown, that resembles a bulleted memo comprising each session. 

“It really understands everything you’ve gone through, everyone you know, and your values,” George says.

TwinMind’s app is available for download and is free to use. As for what’s next, George says TwinMind is now planning a Series A funding round and for the first time is speaking with VC firms while the company grows. He says it’s amassed roughly 10,000 users so far.

“Once people use it, they won’t stop,” George predicts. “I can’t live without it.”

View the full article