Science Corporation implants PRIMA in its first Stargardt patient

Share
Science Corporation implants PRIMA in its first Stargardt patient

An implant company just showed the far edge of what neural engineering can do today, and OpenAI went shopping for the firm that hacks open-source software for a living.


Science Corporation has put its PRIMA vision chip into the first patient with Stargardt disease, CEO Max Hodak said on stage at TechCrunch Disrupt on Monday. The implant went in during May under a trial at Sydney Eye Hospital in Australia, and during the first rehabilitation session the patient recognized a letter; in the second he read a word. The device is a two-millimetre square chip placed under the back of the eye where photoreceptors have died, paired with camera-equipped glasses that transmit a view of the world to it. Hodak is blunt about what that view is worth right now: "It's like looking through a straw… it's black and white, but it is sharp," good for a couple of letters at a time. Stargardt is an inherited disease that destroys central vision and has no treatment; roughly 150,000 people worldwide have lost sight to it.

Here's why this is more than a milestone press release. PRIMA already holds a European CE mark for geographic atrophy, an advanced form of age-related macular degeneration, where the company's trial showed 84% of 38 patients reading letters or words with a mean gain of 25.5 letters at 12 months. In other words, there is a commercial product, a revenue line, and a second disease now being attacked with the same hardware. The honest caveat: the Stargardt result is a single company-reported patient, not peer-reviewed data — a first-in-disease proof of concept, which is exactly what Hodak calls it. He says the next generation is a year from human testing and near-native acuity is a five-year target. If a retina can be driven by correct electrical signal, the same logic is the company's whole roadmap to the wider brain interface it was founded for.


OpenAI is in early talks to acquire Trail of Bits, the security firm that finds and fixes vulnerabilities in open-source software, according to The Information. Neither company has confirmed anything and no price has surfaced; the report describes talks, not a deal. Trail of Bits, founded in 2012, is not a stranger to OpenAI — it runs the company's Patch the Planet program, which pays to fix bugs in the open-source dependencies that everything else is built on, and it has already stress-tested OpenAI's own defences.

Read against OpenAI's year, the logic is straightforward. The company has been building a cybersecurity business in public, most visibly by committing $1 billion in subsidized cyber model access to underfunded defenders — we covered that push in September in OpenAI puts $1 billion behind the defenders nobody funds. Owning the firm that audits open-source code and the labs' own sandboxes would fold the supply-chain layer inside the company rather than renting it. But this stays a rumor from unnamed sources until someone signs: treat it as evidence about OpenAI's intent, not as a transaction.

What to watch: whether OpenAI or Trail of Bits confirms the talks, and whether PRIMA's Stargardt program expands beyond a single Australian trial site.

If an implant can return a patch of central vision, is the bionic eye the AI product of the decade — or still a science project with a market of thousands? Tell us in the comments.

Read more

Physical Intelligence launches a hosted API for its robot models

Physical Intelligence launches a hosted API for its robot models

A robot-model lab opens its fine-tuning stack to outsiders, and a new paper says one policy can out-assemble twenty specialists — in simulation, for now. Physical Intelligence opened a beta of its Pi API, letting outside developers upload demonstration data, fine-tune the lab's robot foundation models on Pi's own compute, and run the result through a hosted inference gateway — no ML pipeline of their own. The company, co-founded by Sergey Levine, Chelsea Finn and Karol Hausman and valued at $5

DOJ names the OpenAI engineer accused of bugging a conference room

DOJ names the OpenAI engineer accused of bugging a conference room

The company's account of a hidden recorder in its offices was the first half of the story. The Justice Department's complaint, unsealed today, is the second half — and it is stranger than the first. Federal prosecutors charged Abdulrahman Alfozan, 35, a Saudi national who had been an OpenAI engineer since July 2025, with using a device to intercept oral communications — and the complaint describes both the recorder and the chatbot prompts allegedly used to plan it. According to the filing, Ope

Character.AI's bankruptcy filing admits the economics never worked

Character.AI's bankruptcy filing admits the economics never worked

The filing that landed Monday said the company was out of money. The affidavit behind it, reported today, says it never had a path to the money in the first place — and puts numbers on a claim the whole AI industry has been arguing about for three years. Character.AI's own chief executive told a bankruptcy court that compute cost the company roughly five times what its subscriptions brought in — and that "at all times, Character has struggled with profitability." According to the first-day aff

Anthropic migrated off git after merges hit 30,000 a day

Anthropic migrated off git after merges hit 30,000 a day

The story under the story today is the infrastructure: what happens when agents write code faster than the tools built for humans, and what a lab does when the threat turns out to be inside the building. Anthropic says its merge rate grew more than tenfold this year, so it replaced git with a version control system it wrote itself — built and fully migrated by seven engineers in five months. In May 2026, the company averaged around 2,500 code changes merged per day across its monorepo; by Sept