When self-taught coder Sigil Wen was 17, he moved to Silicon Valley and lived in an AI hacker house with famed AI researcher Andrej Karpathy.<\/p>\n
While there, he hacked and coded alongside other people who would become the biggest names in AI, like Perplexity founder Aravind Srinivas and OpenAI researcher Noam Brown. He tested early versions of AI tools that would later become well known, including a chatbot shared by Anthropic co-founder Ben Mann that would become Claude, an image generator from David Holz that would become Midjourney, and what would become OpenAI\u2019s GPT-3 and the image generator Stable Diffusion.\u00a0Prominent investor and entrepreneur Naval Ravikant hired him for Airchat, Ravikant\u2019s now-defunct rival to the Clubhouse social network. <\/p>\n
For fun, he figured out how to get GPT-2\u00a0running on his Apple Watch. \u201cIt was a magical time,\u201d Wen told TechCrunch.<\/p>\n
Now a Thiel Fellow \u2014 the program from investor Peter Thiel that invites young founders to pursue projects instead of college \u2014 Wen on Monday launched<\/a> an invite-only beta of Underdog, one of the most private AI assistants Silicon Valley has yet to offer. The model runs wholly on-device, meaning the user\u2019s data remains on devices they already own, currently Macs and Windows PCs, with Linux, iPhone, and Android versions coming soon.<\/p>\n Underdog is powered by Husky, the inference engine Wen built to run AI models quickly on a user\u2019s own hardware. Wen says it moves less data between a computer\u2019s main chip and its graphics chip than rival on-device engines.<\/p>\n Underdog has other security features baked in, too, like encrypting the keys to the email and other accounts that users authorize Underdog to access.<\/p>\n Underdog is, however, using much smaller models than today\u2019s state-of-the-art ones hosted in data centers. It currently uses a 27-billion parameter reasoning model fine-tuned from Qwen3.8-27B.<\/p>\n Wen argues that this model compares favorably with Claude Opus 4.6 in some benchmarks<\/a>, or what was considered top performance six months ago. He says that means it can handle the everyday tasks that people want an AI assistant to do, like shopping research or answering math homework questions.<\/p>\n \u201cYou don\u2019t need to sacrifice your privacy for the capability because they\u2019re just as capable,\u201d Wen says. He adds that small on-device models will continue to grow more capable over time.<\/p>\n Perhaps the most interesting thing about Underdog is its early business model. The app will be free at first and never ad-supported. Since the AI runs on users\u2019 machines, Underdog doesn\u2019t have the giant overhead of paying a provider for inference. \u201cI don\u2019t have to charge you a subscription to run this because my costs are so super low,\u201d he said.<\/p>\n Instead, with Stripe co-founder Patrick Collison as one of his angel investors, he\u2019s borrowing a play from the fintech era. Underdog will take a tiny percentage of payment transactions that the AI assistant makes using Stripe\u2019s secure payment rails, something like an interchange fee. In this way, the AI assistant never has to mine your data. It is as aligned with you as your bank or credit card providers.<\/p>\n This runs contrary to the business motivations of many of the other players<\/a> in AI assistants, whose privacy policies allow them to collect data on users<\/a> that they may sell to advertisers or other third parties, or use to train other models.<\/p>\n That kind of data collection could be a particularly treacherous trade-off for users of an AI assistant, which may need access to the most intimate details about you in order to be useful, from medical conditions to financial data to data on your kids.<\/p>\n As Wen wrote in what he calls his AI manifesto<\/a>, \u201cWhy should using AI require surrendering your private information?\u201d<\/p>\n He tells TechCrunch: \u201cI honestly want to build Underdog for myself. I\u2019m building a product that I would be proud for my future children to use.\u201d<\/p>\n The startup behind Underdog is named Conway Research, and Collison isn\u2019t its only big-name investor. Conway is backed by Andreessen Horowitz<\/a> via partner Chris Dixon, as well as Khosla Ventures, Hummingbird, SV Angel, the Anthology Fund (the partnership fund between Menlo Ventures and Anthropic), and a prominent list of angel investors<\/a> that includes Vercel founder Guillermo Rauch, Noam Brown, and Deedy Das, among others.<\/p>\n