Ryan Castillo: Jordan, quick question before we get into it — if I told you 'six to twelve months' twice, fourteen months apart, would you believe me the second time?
Jordan Hale: I mean... once, sure. Twice I'm starting to wonder what 'months' means to you.
Ryan Castillo: Elon Musk has said it about Neuralink's Blindsight device — the visual cortex implant — three times now. April 2025 at a Wisconsin town hall, end of 2025. Missed. June 2025 at Y Combinator, six to twelve months. August 2026, six to twelve months again.
Jordan Hale: Wait, the April one was just, what, two months before the Y Combinator one?
Ryan Castillo: Two months. Which is why this is — look, this isn't a single missed deadline, this is a pattern. Three resets, sixteen months, and as of October 2026 there's no Blindsight trial on ClinicalTrials.gov. No public FDA authorization for a human study.
Jordan Hale: So the question is whether the six-to-twelve month window is a real engineering target or just... the thing you say when you don't have a firmer answer.
Ryan Castillo: And whether the FDA's Breakthrough Device Designation — which is real, September 2024 — is doing credibility work it isn't actually designed to do.
Jordan Hale: That's the thing I want to pull apart, because I think people hear 'FDA designation' and assume the finish line is right there.
Ryan Castillo: Right — but the designation question actually leads straight into what the device does, because once you understand the mechanism, the regulatory gap gets sharper. Blindsight doesn't touch the eye at all. Your eye is a broken camera — Blindsight just plugs a cable directly into the TV. The visual cortex. Skips the camera entirely.
Jordan Hale: So the eye doesn't even need to be involved.
Ryan Castillo: Doesn't need to exist, functionally. What Neuralink is doing with Blindsight is stimulating the visual cortex directly with electrodes. The brain then produces what are called phosphenes — perceived patterns of light. Not a crisp image, but something the brain interprets as light and shape.
Jordan Hale: Okay, so picture someone who went blind at eighteen from retinal disease. Her retinas are gone, her optic nerve is damaged — but her visual cortex is just... sitting there, intact, waiting for a signal that never arrives.
Ryan Castillo: That's exactly the target patient. And that's also the eligibility condition — the cortex has to be functional. Which is a meaningful but narrow window. You can't be blind from, say, cortical damage and expect this to work.
Jordan Hale: Wait, so someone blind from birth — could they even be a candidate? Like, you know, if the cortex never developed the visual pathways in the first place...
Ryan Castillo: Neuralink says yes — designed to work even in people blind from birth, provided the cortex is intact. Whether that holds up in practice, I mean — animal trials have happened, but what data actually reached the FDA? Nobody's published that. That's the part of the paper trail that's genuinely thin.
Jordan Hale: And Blindsight is targeting the visual cortex — which is a completely different brain region than what PRIME does for motor function. So the two years of human trial data they're building with PRIME doesn't just... transfer.
Ryan Castillo: Right, but that gap between PRIME and Blindsight is actually where the Breakthrough Device Designation becomes interesting — and a little misleading. The FDA granted it in September 2024. People read that and think green light. It's not. It means the FDA will talk to Neuralink more. More frequent check-ins, faster feedback loops. That's it. It doesn't authorize a single human implant.
Jordan Hale: Wait, so it's like... getting a meeting with the principal, not getting permission to leave school early.
Ryan Castillo: That's the whole thing. Closer interaction, maybe accelerated review down the line — but not permission to test in humans. And the gate that actually matters is ClinicalTrials.gov. When Neuralink filed to run PRIME trials in 2024, it showed up there. That's how you know the FDA authorized it. You can look it up.
Jordan Hale: And for Blindsight — nothing.
Ryan Castillo: As of October 2026, no entry. Which means either the IND — the investigational new device application — hasn't been filed, or it was filed and something happened quietly. Neither option supports a six-to-twelve month window for a first human implant.
Jordan Hale: You know what this reminds me of? Like, picture a woman — say she's been blind since her early twenties from retinal disease, cortex totally intact, she reads about the Breakthrough Designation in September 2024 and thinks, okay, maybe two years from now this is real. She starts actually — I mean, she's probably Googling ClinicalTrials.gov right now. And there's nothing there for her to enroll in.
Ryan Castillo: And here's my surprise on this — the monkey clock. At Y Combinator in June 2025, Musk tied the six-to-twelve month window to a specific condition: a monkey had to have carried the implant for three years first. Fine. Except nobody has published when that three-year clock started. It's a trigger buried inside a contingency that's unverifiable from the outside.
Jordan Hale: Wait — no one knows when the clock started? That's not vague, that's... that's constructed vagueness.
Ryan Castillo: No published start date, no published end date. And the three resets themselves — what that pattern actually signals for credibility, and then there's the superhuman vision framing layered on top of the near-term medical case — that's the part that makes this messier than it looks.
Jordan Hale: And that superhuman vision framing — you know, that's the part that makes my head tilt, because Musk has talked about infrared, ultraviolet, radar-like perception as part of the Blindsight roadmap. Like those are in the same breath as 'six to twelve months.'
Ryan Castillo: That's where the credibility math breaks. Those capabilities — infrared, ultraviolet, radar — there's no technical timeline for them, no regulatory pathway anybody's published. And when you put them next to the near-term medical device claim, the whole frame sounds further along than it is.
Jordan Hale: It's doing work it shouldn't be doing.
Ryan Castillo: Exactly — and here's my actual surprise. PRIME, the motor-function BCI, started authorized human trials in 2024. Still ongoing. That's Neuralink's own pace on a device that's already through the regulatory gate. Blindsight is targeting a different brain region, harder problem, no ClinicalTrials.gov entry — and somehow the timeline is supposed to be faster?
Jordan Hale: Wait — PRIME is still in early human trials after two years and Blindsight is supposed to leap ahead of it?
Ryan Castillo: That's the number that matters. PRIME is the baseline. And it doesn't support 'months away' for a device in a completely different cortical region with, as far as we can tell publicly, no filed IND.
Jordan Hale: Picture her. She's in her sixties, lost her vision at eighteen from retinal disease, her visual cortex is intact, she is literally the patient Blindsight is built for. She heard the April 2025 announcement. She heard June 2025. She's hearing August 2026. That's three times she's been told 'soon.' What does 'soon' mean to her now?
Ryan Castillo: And the engineering might genuinely be advancing — we can't see inside the lab. But the public record, the ClinicalTrials.gov absence, the PRIME pace — none of it confirms the window. The three resets aren't evidence it's failing. They're evidence we can't tell either way.
Jordan Hale: Which is maybe the most uncomfortable place to sit — not cynicism, just... genuine uncertainty, with someone's hope on the other end of it.
Ryan Castillo: The science underneath it is real, though. That part I want to be clear about. Visual cortex stimulation — phosphenes, bypassing the optic pathway entirely — that's documented. Neuralink didn't invent that concept out of nothing.
Jordan Hale: No, and I think that's what makes the timeline resets genuinely frustrating rather than just, you know, cynical PR noise. Like the work is real. Blindsight is real neuroscience. It's just that three resets in sixteen months — April 2025, June 2025, August 2026 — that pattern is now its own variable. Separate from whether the device actually functions. If it launches and restores even partial vision, that's medical history. But the credibility debt is sitting there waiting to be paid back.
Ryan Castillo: And the trust math is the thing nobody can calculate right now. Because the ClinicalTrials.gov absence, the PRIME pace — none of it tells you the device fails. It just tells you the gap between Musk's announcement cadence and the regulatory record is real, and it compounds every time.
Jordan Hale: That's where I land, I think. Genuine uncertainty — not verdict.
Ryan Castillo: Same. Appreciate you sitting in the uncomfortable part with me.
