August 5, 2026
Mind-reading or mind-melting?
Why I'm leaving OpenAI to build telepathy
OpenAI researcher quits to build mind-typing—and commenters are spiraling
TLDR: An OpenAI researcher left to help build a headband that turns thoughts into computer text, pitching it as a near-future everyday tool. Commenters were split between awe and alarm, with the loudest debate focused on privacy, bad-thought paranoia, and whether “telepathy” should do more than speed up desk jobs.
An OpenAI researcher has dramatically exited the AI giant and immediately joined Conduit to chase a sci-fi dream: turning thoughts into text with a headband. The pitch is pure future-shock—by 2027, you’re supposedly glancing at code, thinking vague instructions, and your computer just gets it. By 2030, the writer claims going without the brain band could feel as annoying as using a phone without a touchscreen. Naturally, the comment section did what comment sections do best: turn one bold announcement into a full-blown public debate about privacy, mind-reading, and whether this is genius or deeply cursed.
The biggest split? One camp is thrilled. Some readers called it exciting and immediately started imagining huge benefits, like using brain signals to teach AI what people really want instead of making them type it out. Another crowd slammed the brakes, asking the question everyone was thinking: how exactly do you stop your private thoughts from leaking out? One commenter went straight to dystopia, worrying about people getting in trouble for thinking the “wrong” thing. That, unsurprisingly, lit up the drama.
And then came the jokes. The funniest running bit was basically: if a microphone can be both a speaker and a listener, can this thing also do text-to-thought? In other words: today it’s “mind to keyboard,” tomorrow it’s “ads directly in your brain,” and everyone is nervously laughing. Others weren’t sold on the vision at all, arguing that real telepathy should transform human life—not just make office work faster. So yes, the founder-style announcement landed, but the real show was the crowd asking whether this is the next great leap… or the moment we accidentally invented Bluetooth for intrusive thoughts.
Key Points
- •The author says he left OpenAI on July 23, 2026 and joined Conduit on July 24 as a Founding Researcher.
- •Conduit is described as developing thought-to-text models trained on non-invasive neural data.
- •The article predicts that thought-based interaction could become a main interface for working with AI systems within a few years.
- •A 2027 scenario describes a neural headband connected by Bluetooth to a laptop, where a Conduit model decodes thoughts into requests sent to Codex and a GPT-7 agent.
- •A 2030 scenario describes AI systems interfacing directly with Conduit’s latent representations, enabling communication of information that is hard to express in text, such as mental images.