惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

大猫的无限游戏
大猫的无限游戏
H
Hackread – Cybersecurity News, Data Breaches, AI and More
博客园_首页
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
D
Docker
酷 壳 – CoolShell
酷 壳 – CoolShell
宝玉的分享
宝玉的分享
Martin Fowler
Martin Fowler
美团技术团队
量子位
M
MIT News - Artificial intelligence
Apple Machine Learning Research
Apple Machine Learning Research
阮一峰的网络日志
阮一峰的网络日志
博客园 - 叶小钗
博客园 - 三生石上(FineUI控件)
腾讯CDC
Hugging Face - Blog
Hugging Face - Blog
博客园 - 【当耐特】
小众软件
小众软件
博客园 - 司徒正美
罗磊的独立博客
云风的 BLOG
云风的 BLOG
B
Blog RSS Feed
博客园 - 聂微东

Futurism

Man Creates Tiny Submarine for His Parakeet to Experience Life Underwater The Effects of AI-Generated Code Tearing Through Corporations Is Actually Kind of Funny Trump Hires Orbital Towing Company to Build Space Interceptors Psychologists Found Something Horrible About the Kind of Men Seeking Trad Wives To Get Swole, Teens Are Pumping Themselves Full of Drugs Meant for Fattening Cows for the Slaughterhouse Foolish Pollsters Are Now Just Asking AI What Voters Would Say in Response to Questions and Publishing It at Face Value OpenAI Says It’s Already Made $100 Million by Stuffing ChatGPT With Ads Man Punished for Breaking Into Moo Deng’s Zoo Enclosure AI Is Causing Healthcare Costs to Surge There’s a Mass Rebellion Against AI in the Workplace People Who Lose Their Job to AI Are in for a World of Pain, Goldman Sachs Report Finds OpenAI Says Not to Worry About UBI, Because It Has Another Idea Police Officer Helplessly Waves Arms at Waymo That Careened Wrong Way Through Whataburger Drive-Thru Someone Just Threw a Molotov Cocktail At Sam Altman’s House New York Times Makes Substantial Changes to Article That Glazed a Sleazy AI Startup: “Our Piece Should Have Included That Information” Space Scientists Wince as Astronauts’ Lives Depend on Artemis 2’s Controversial Heat Shield During Plunge Back to Earth The Moon Astronauts Have Been Working Out With a NASA Rowing Machine in Space First AI Model From Zuckerberg’s Wildly Expensive Superintelligence Lab Flops Compared to Virtually All Rivals Economists Starting to Admit They May Have Been Wrong About AI Never Replacing Human Jobs AI-Powered Drug Marketer Medvi Responds After Allegations About Fake Doctors and Patients As Astronauts Visit the Moon, NASA Insider Says Agency Is in Shambles Behind the Scenes Man Lights 1.2 Million Square Foot Warehouse on Fire for Not Paying Him Enough NASA Scientists Screamed With Delight When They Saw Something Smashing Into the Moon Google Says Showing Polymarket Bets on Google News Was a Mistake Las Vegas Sphere Turns Into Huge Moon to Celebrate NASA Mission The New York Times Says It’s Identified the Creator of Bitcoin We Talked to a Writer Accused of Publishing An AI-Generated Essay in The New York Times Naked Man Bursts Into Tesla Service Center With a Shotgun Student Dies When Hospital Has No ICU Doctors, Calls One on Videochat Who Pronounces Him Dead Remotely, Lawsuit Claims Analysis Finds That Google’s AI Overviews Are Providing Misinformation at a Scale Possibly Unprecedented in the History of Human Civilization
Startup Says It’s Invented a Beanie That Reads Your Mind
Joe Wilkins · 2026-05-05 · via Futurism

A person wearing a beige knit beanie and a patterned sweater is sitting at a wooden desk, turning the pages of a yellow book. The background features a dark abstract wall art and a black desk lamp with a curved gold base. There is also a vase with dried flowers on the desk. The lighting is warm and casts a golden glow on the scene.

Sabi

Sign up to see the future, today

Can’t-miss innovations from the bleeding edge of science and tech

If you thought AI-integrated smart glasses were bad, wait until you get a load of Sabi, a Palo Alto-based startup working on a beanie it says will probe your actual brain signals.

That’s not hyperbole. The company’s eponymous Sabi Cap, per New Atlas, comes lined with 100,000 electroencephalography (EEG) sensors, which will translate electrical signals from your brain into usable data for Sabi’s “Brain Foundation” AI model — all meant to transcribe your thoughts into digital text at what the company says will be a rate of 30 words per minute.

The AI model powering it is said to be trained on 100,000 hours of data from some 100 volunteers, Wired previously reported. But given that thought and speech patterns vary wildly between person to person, the challenge of building a universally workable EEG-to-speech device is enormous, and the company has yet to share any evidence that its product performs as advertised.

“These devices are going to have to be ready to go out of the box,” third-party neurotech consultant JoJo Platt told Wired. “They’re going to have to conform to me rather than me conforming to it.”

The commercial appeal is clear. It’s hard to imagine a surgically implanted brain chip like Neuralink ever gaining genuine mass traction, making a lightweight alternative compelling. And certain evidence does suggest that you could get usable data from outside the skull; as one non-peer-reviewed paper found a few years ago, AI models fine-tuned with EEG data represent a “significant advancement towards portable, low-cost ‘thoughts-to-text’ technology with potential applications in both neuroscience and natural language processing.”

Yet as a peer-reviewed paper published in Scientific Reports last year found, the efficacy of EEG-to-text models remains “unclear due to limitations in evaluation methodologies.” The early promise of EEG, the later study argues, is likely the result of flashy pattern memorization rather than a novel tech that can decode human brain waves.

In other words, it’s possible Sabi’s founders have massively underestimated how advanced their brain foundation model really is. Until we get a glimpse at the product, set to release later in 2026, it’s anyone’s guess whether this is a true mind-reading device, or just a very expensive hat.

More on tech companies: There’s Something Bizarre About the Offices of AI Startups