Claims

Kernel Flow uses near-infrared spectroscopy with 52 photon modules (each containing one laser and six sensors firing in ~100 picoseconds) to measure blood oxygenation changes in the cortex, with approximately 1,000+ channels sampling brain activity in real-time without requiring laboratory settings or surgical implants.

Kernel was initially built with invasive brain interface technology; approximately one year after a meeting with Elon Musk (in which they discussed both pursuing brain interfaces), Johnson was presented with a neuroscientist who suggested that 'every 15 years or so a new technology comes along that changes everything...it's probably already here, you just can't see it yet'; this prompted Johnson to pivot Kernel's entire technology from invasive to non-invasive.

Kernel Flow provides 'a full screen on the picture of information but the image is going to be blurred and the audio is going to be muffled' (lower resolution coverage) compared to Neuralink which provides 'a circle on the screen of 4k' (high resolution but narrow coverage); both have pros and cons, and Johnson believes the next 7 years will see innovation in hardware that improves non-invasive approaches.

Internally at Kernel, when discussions veer toward intuitions about what applications should be pursued, the team flags it as a 'human intuition alert' and stops the discussion, believing that organic discovery through users playing with devices in expected and unexpected ways is superior to exercising informed intuitions.

News consumption, social media use, and how individuals allocate attention online are compelling domains for brain interface application because the coupling of brain data with digital behavior patterns could reveal the effects of digital world behaviors on physical body effects.

Brain data could provide gentle nudges that encourage self-investigation without imposing interpretation, helping individuals discover which behaviors and activities make them feel good or bad through concrete data rather than relying on fallible memory and emotional state.

Most technologies enabling mainstream adoption languish in academia for long periods before reaching market, and the difficulty of bringing transformative brain technologies to market is why most brain interface efforts remain academic rather than commercial, which Kernel hopes to overcome in coming years.