a16z partners Chris Dixon, Gabriel Vasquez, and Liz Harkavy jointly announced the leading investment in Conway Research's Series A, backing founder Sigil Wen and his first product Underdog. The amount has not been disclosed.

The article states that Sigil was introduced to Dixon at an event hosted by Naval Ravikant in 2021. He later assisted Naval with Airchat, integrating new models into everyday products and recruiting most of the engineering team; Naval also made him the youngest member of Spearhead. He received the Thiel Fellowship in 2025.

When he arrived in San Francisco at 17, he had almost no connections and once slept in WeWork, dropping out of college and living in a hacker house for nearly a year. Andrej Karpathy lived there. Anthropic co-founder Ben Mann invited him to be one of the earliest testers of Claude. David Holz showed him the first version of the model behind Midjourney. Others in the house included people involved with DALL·E, Whisper, GPT-3, and Stable Diffusion.

Underdog is defined as personal intelligence running on users' own devices, capable of assisting decision-making offline. The motivation is linked to two past incidents: while living with Karpathy, a vulnerability in an email application exposed his private emails to other users; he had previously run GPT-2 on an Apple Watch and integrated Whisper into the iPhone via Airchat.

The shared judgment of the three is that the AI that knows you best should work for you, with data under your control. Once an assistant transitions from answering questions to managing tasks, it must handle conversations, commitments, documents, and preferences, which are also the most sensitive information. They envision a future covering billions of phones and computers, giving users sovereignty over time, information, and choices.

The buyer is a16z partners who have previously invested in him, along with Naval's network that introduced him; the seller is a cloud assistant that must upload context to become smarter. The event was driven by the leading investment document. Funding is shifting from pay-per-call personal data to local models, benefiting teams that can deliver offline assistants on old devices, while putting pressure on assistant companies that keep emails and calendars on their own servers.

Source: Public information