Neuralink said in an Oct. 29 post that its engineers had spent 72 hours with a participant in its VOICE speech-restoration study. In its account of the sprint, the company says the work took place at his home, where engineers built speech-interface features and tested them with him.
One change targets a button press. Neuralink says speech decoding previously required users to manually start and stop recording. Its new system detects when a user is trying to speak so the app can start and stop on its own. The company also says it prototyped a Grok agent trigger: a user thinks of saying “Hey Neuralink” followed by a task, which the agent then carries out on a computer.
For phone calls, Neuralink says it connected Bloom, its mobile speech brain-computer interface app, to iPhone calling. That integration sends BCI-generated speech into the call’s microphone stream, the company says. It also tested an optional greeting to tell the person answering that the caller uses an assistive communication device. Cyborg21 wrote in a quote-post that engineers had spent 72 hours in his living room and that he joked on the phone again.
Neuralink also reports an iOS keyboard that routes BCI dictation into apps, an editing flow and a Bloom lock-screen Live Activity. Its captions say the interfaces shown for those three features underwent further iteration in the weeks after the hackathon. In a trial disclaimer, Neuralink says its devices are investigational and have not been approved by the FDA or other regulatory authorities. It cautions that participants’ experiences may not reflect those of other participants or future outcomes.
Comments