Back in 1908, a Columbia University psychiatry professor named Frederick Peterson informed a newspaper that he had built a machine capable of reading a person’s thoughts. The apparatus consisted of a mirror and a lamp.
“It is almost like sorcery to observe the behavior of the instrument with a series of words, some purposely chosen to be emotionally indifferent and others to strike some emotional complex deep in the soul of the individual experimented upon,” Peterson said of his device in comments to the Seattle Star.
His psychometer had more in common with a Ouija board than with real science. The sales pitch, though, has aged remarkably well — because a San Francisco startup is currently handing over $50 to $55 to anyone willing to spend two hours in its office wearing a four-pound helmet studded with electrodes while talking to an AI.
By the company’s own account of the setup, the helmet looks like something House Harkonnen might wear.
What you actually do for the $50
The startup is Conduit, launched in 2024, and the sessions it labels “freeform” live up to the name. Participants show up, strap on the hardware, and then talk or type for two straight hours.
“Sessions vary,” the company wrote about its data-collection process in a December blog post. “Some are listening and speaking with an LLM, and some are reading and typing. The goal is to maximize the amount that subjects type or say during the two-hour period, without constraining the topics they discuss.”
Only one prerequisite applies, and it’s oddly specific: applicants must be able to type without glancing down at their fingers. That starts to make sense the moment you consider what eye movement does to a brain scan. A cap of 32 sessions per person is also in place.
Run the numbers at the top end and you arrive at $1,760 for 64 hours of your life — a commitment sitting somewhere between a sleep study and an unusually odd part-time gig.
The reason the data matters more than the hardware
Today’s AI systems run on vast quantities of training data, and reading neural signals doesn’t alter that requirement one bit. The volunteer recruitment isn’t some side experiment. It is the product.
On its website, Conduit frames its ambition as “telepathy at scale” — AI that reads minds purely by noninvasive means. No surgery, no implant, nothing resembling the brain-computer interface Elon Musk is developing at Neuralink.
Further out, the plan calls for a wearable that pipes commands to AI assistants directly from your thoughts rather than through voice or text. Anybody who has wrestled with Siri, or watched ChatGPT confidently misinterpret a straightforward request, can appreciate the draw of bypassing the prompt altogether.
Conduit isn’t alone here
Meta rolled out the newest iteration of its Brain2Qwerty system in June, a project designed to convert brain scans into text. The company frames it as an early move toward restoring communication for people with ALS and other neurodegenerative conditions.
That framing does real work. Medical applications are what open doors to research access and public goodwill. The consumer application — dictating to your coding assistant without touching a keyboard — is what earns a valuation.
The people betting their careers on it
Conduit’s founders are Rio Popper, an Oxford graduate who lost her sight at age five, and Clem von Stengel, an AI researcher. Last week, Naomi Bashkansky, a former alignment researcher at OpenAI, used a blog post to announce that she was departing to join them.
“I like that this is literally a human-in-the-loop vision of the future, where AI directly empowers humans rather than replacing us,” Bashkansky wrote of her decision, before calling the work “frighteningly ambitious.”
Should Conduit deliver on its goals, she anticipates the company will eventually be worth north of $1 trillion. That figure isn’t a typo, and it deserves a moment’s reflection: a trillion-dollar business whose data pipeline right now consists of people sitting in a San Francisco room in a helmet for beer money.
The 2030 version, in her own words
By 2030, Bashkansky pictures Conduit’s “neural headband” letting users issue commands to tools such as OpenAI’s Codex through thought alone. Writing in the voice of her future self, she sketches out how working without one would feel.
“Without it, I feel like I’m talking with a superpowered alien who’s trying its best to be helpful but isn’t sure what I want and is scared I’ll get mad if it does something I don’t want,” she writes. “Ugh.”
The headband could turn into an essential piece of her workday as early as next year, she believes, allowing her to shoot off instructions to an AI assistant while mapping out the day over coffee.
“It feels like magic,” she writes.
That is nearly the exact language Peterson used about his mirror and his lamp in 1908. He turned out to be wrong. Conduit may not be. What separates the two claims isn’t how confident the speaker sounds, nor how elegant the pitch is — it’s whether 32 sessions multiplied by however many volunteers yields enough signal to build on. The $50 checks will settle that question well before any headband reaches the market.