In 1908, a Columbia University psychiatry professor named Frederick Peterson told a newspaper he’d built a machine that could read your thoughts. It involved 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 told the Seattle Star about his device.
The psychometer was probably closer to a Ouija board than to science. But the pitch has aged remarkably well, because a San Francisco startup is now paying people $50 to $55 to sit in its office for two hours wearing a four-pound helmet covered in electrodes, talking to an AI.
The helmet, by the company’s own description of the setup, looks like something House Harkonnen might wear.
What you actually do for the $50
The company is Conduit, founded in 2024, and the sessions it calls “freeform” are exactly that. You show up, you put on the hardware, and you talk or type for two 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.”
There’s one requirement, and it’s a specific one: you have to be able to type without looking at your fingers. Makes sense once you think about what eye movement does to a brain scan. There’s also a 32-session cap per person.
Do the math on the upper end and that’s $1,760 and 64 hours of your life, which puts this somewhere between a sleep study and a very strange part-time job.
The reason the data matters more than the hardware
Modern AI systems need enormous piles of training data, and nothing about reading neural signals changes that. The recruitment drive isn’t a side project. It’s the product.
Conduit describes its goal on its website as “telepathy at scale” – AI systems that read minds entirely through noninvasive means. No surgery, no implant, nothing like the brain-computer interface Elon Musk is building at Neuralink.
The long-term plan is a wearable that sends commands to AI assistants straight from your thoughts instead of through voice or text. Anyone who has fought with Siri or watched ChatGPT confidently misread a simple request understands the appeal of skipping the prompt entirely.
Conduit isn’t alone here
In June, Meta unveiled the latest version of its Brain2Qwerty system, built to translate brain scans into text. Meta is positioning it as an early step toward helping people with ALS and other neurodegenerative illnesses regain the ability to communicate.
That framing matters. The medical use case is the one that gets research access and public goodwill. The consumer use case, dictating to your coding assistant without typing, is the one that gets a valuation.
The people betting their careers on it
Conduit was founded by Rio Popper, an Oxford graduate who went blind at age five, and an AI researcher named Clem von Stengel. Last week, Naomi Bashkansky, a former alignment researcher at OpenAI, announced in a blog post that she was leaving 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.”
She expects Conduit, if it hits its goals, to eventually be worth more than $1 trillion. That’s not a typo, and it’s worth sitting with for a second: a trillion-dollar company whose current data pipeline is people in a room in San Francisco wearing a helmet for beer money.
The 2030 version, in her own words
Bashkansky imagines that by 2030, Conduit’s “neural headband” will let users pass commands to tools like OpenAI‘s Codex using thought alone. Writing as her future self, she describes what working without it would feel like.
“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.”
As soon as next year, she believes the headband could become a vital part of her workday, letting her fire off commands to an AI assistant while she plans the day over coffee.
“It feels like magic,” she writes.
Which is close to word-for-word what Peterson said about his mirror and his lamp in 1908. He was wrong. Conduit might not be. The difference between the two claims isn’t the confidence of the person making it, and it isn’t the elegance of the pitch – it’s whether 32 sessions times however many volunteers produces enough signal to build on. That’s a question the $50 checks will answer long before any headband ships.
