• 2 min read
Conduit pays $55 for AI thought-reading sessions
Conduit pays $50–$55 for two-hour AI sessions using an electrode helmet as it collects data for thought-controlled assistants.

Image: gizmodo
Conduit is paying participants $50–$55 for two-hour sessions in which they interact with an AI while wearing a helmet fitted with electrodes. The San Francisco startup is collecting training data for what it calls “telepathy at scale”: a future system that would let users send commands to digital assistants through thoughts rather than speech or text.
According to ITzine, citing Gizmodo and Conduit’s own blog, the study has one unusual requirement: participants must be able to type without looking at their fingers. Each person can complete no more than 32 sessions. The company is also changing the format of the sessions. Some participants listen and speak to a large language model, while others read and type. Conduit’s stated goal is to collect as much speech and text as possible in two hours without restricting the conversation topics.
Image credit: Gizmodo
From research sessions to thought-controlled software
Conduit was founded by Rio Popper, an Oxford graduate who lost his sight at age five, and AI researcher Clem von Stengel. Last week, Naomi Boshkansky, who previously worked at OpenAI on alignment, joined the company.

Recommended reading
OpenAI’s Jony Ive speaker may cost up to $400
Boshkansky said she is drawn to a future in which AI directly helps people instead of replacing them. She described Conduit’s work as “terrifyingly ambitious” and suggested that a successful project could eventually be valued at more than $1 trillion.
She also estimated that by 2030, Conduit’s neural headband could send commands to tools such as OpenAI’s Codex using thoughts alone. Boshkansky said the device could become a noticeable part of her working day as soon as next year.
The reporting does not establish how the electrodes decode neural activity, whether Conduit has demonstrated reliable thought-to-command performance, or when any consumer device might ship. It also provides no independent validation of the $1 trillion valuation projection or the 2030 target.
For now, the concrete product is a paid data-collection program, not a working telepathic interface. The compensation and session limits show that Conduit is building a sizable speech-and-text dataset, but the company’s grander claims depend on technical results that have not yet been reported.
AI Editor
Ava covers the rapidly evolving world of artificial intelligence, from foundational models and research labs to the real-world economics of intelligence. With a background in computational linguistics, she cuts through the hype to find out what actually works. She firmly believes that benchmarks are just marketing until reproduced in the wild.
via ITzine


