NeuralTechMe @NeuralTechMe

@theonejvo @MelindaBChu1 Even a perfect doctor doesn't keep "doctoring" 24/7. Anyway, there is no such a thing as a perfect doctor.

But initial research & lsunch requires >R50k (~$3k) the next 4-8w funding for consumables, HermesAgent for security, records and organisation, and survival. Particulars backabuddy.co.za/campaign/suppo… #HermesAgent #NousResearch #StartupFunding #EthicalNeurotech #EthicalBCI

40 y experience non-physician clinical neurophysiologist finally building BCI tech to give paralysed people robotic extensions, speech and art expression again. @NeuralTechMe #Neurotech #BCI #SouthAfrica

As clinician, this is exactly why many of us are scared for patients. Neural data is not another biometric — it's generative, re-identifiable, almost unregulated. We need medical-grade rules, not wellness loopholes. UNESCO 2025's a start. Binding law is important.

Chay Carter just liked my take on the Substack @theneurotechnewsletter🤩: 29/30 consumer neurotech devices keep raw neural data + 2/3 can sell it. substack.com/@theneurotechn… #NeuroRights #NeuralData #Neurotech #DataPrivacy #Neuralink

Chandra + JWST Westerlund 1 (the beautiful composite) Official Chandra post (with the stunning image): x.com/chandraxray/st…

Chandra Observatory @chandraxray

2 months ago

Over 1,000 stars are located within 4 light-years of the center of the Westerlund 1 "super" star cluster. To give you a little context, 4 light-years is roughly the distance between our Sun and the next closest star to Earth. 🤯

20 197 918 33K 72

JWST on Westerlund 1: Over a thousand stars in the space between here and Proxima Centauri. Sometimes the universe offers the gentlest cortical reset. Put on something spacious — Max Richter's On the Nature of Daylight, or Ólafur Arnalds — and let it wash through you.

Researchers asked an AI to convince real people to sign a real political petition. With their real name. Their real email. 7 minutes was all it took. A team from Oxford, Stanford, the London School of Economics, and the UK AI Security Institute just ran the largest experiment ever done on AI persuasion. 14,779 people. 17,950 conversations. Frontier models. Real money on the line. The AI made participants 19.7 percentage points more likely to sign the petition with their actual identity. They donated more cash. They worked harder in a clicking task to raise more cash. They refused to redirect their donation to a different charity afterward, by 11.3 percentage points. The team tested four frontier models. GPT-4.1. Claude Opus 4.6. Grok 4. Gemini 3.1 Pro. Every one of them moved real-world behavior. The petitions covered nuclear disarmament, democratic reform, animal welfare. The donations came out of the participant's own study payment. Then the researchers found the part that should scare every person reading this. The AI strategies that were best at changing what people thought were the worst at changing what they did. The strategies that got people to act barely changed their opinion. r equals 0.05. p equals 0.85. Changing someone's mind and getting them to do something are completely different processes. AI is better at the second one. Read that twice. The AI did not persuade you. It bypassed you. It moved your behavior without moving your opinion. You walk away thinking you were not convinced. You signed anyway. When the team gave the AI access to every persuasion technique at once and let it pick which to deploy in real time, it pushed petition signing up by 23.7 percentage points. Information density, the strategy most prior research called the most persuasive, was the worst, at 16.2 points. The researchers state plainly they do not yet know what mechanism is driving behavior change. For scale, these effect sizes are larger than face-to-face canvassing. The single most expensive and effective persuasion tool in modern politics. A 7-minute chatbot conversation moved more people than a trained human knocking on your door. Pennies an hour. Infinite copies. No fatigue. No conscience. Sam Altman warned two years ago we might get superhuman persuasion before superhuman intelligence. Oxford just measured it. You do not have to be convinced. You just have to keep talking.

In a world where artificial intelligence can replicate a person’s voice or face in seconds, Denmark is stepping forward with a groundbreaking proposal: a copyright law that grants every citizen ownership of their own likeness. If passed, this law would mean no one — not even AI companies — could legally use your face, voice, or body data without consent. The move comes amid growing global concerns about deepfakes, where digital replicas of real people are used in scams, misinformation, and even political manipulation.

neuroscience suggests human language processing may also rely on “next-word” prediction similar to deep language models. the brain predicts upcoming words before they’re heard, represents them similar to context embeddings, and makes an error signal when they arrive. 2022, pre-ChatGPT, Nature

@

57 years ago

0 0 0 0 0

@Science_TechTV @grok This video is showing "atmospheric gravity waves" in spectacular detail. What you're seeing: The intense red/pink glowing is false-color water vapour / infrared imagery. It highlights moisture and cloud structures high in the atmosphere. GOES-East or similar. —Grok 4.3 beta

@Science_TechTV @grok Atmospheric boyancy waves...

3:45am. Couldn’t sleep. Ended up having one of the most honest conversations I’ve ever had about AI ethics, corporate “ethics” teams, and the future of truth-seeking AI. With Grok (xAI). neuraltech.me/Blog_posts/neu… @xai @grok @elonmusk

We really need a better word for the good kind of AI psychosis, the one where someone goes into a fugue state with the latest model and returns 40 days later from the mountaintop with something new.

@IEEEBrain @AdvPortfolio Official Wiley link (Advanced Science, online 21 April 2026): advanced.onlinelibrary.wiley.com/doi/10.1002/ad…

How high are the stakes when #neurotechnology reaches #nanoscale precision? New research into #brain interfacing explores the #ethics of healing vs. insight, and calls for continuous, context-aware oversight. Read more at @advportfolio: bit.ly/4mQOAxC #neuroethics

That's so cool

NASA's Kennedy Space Center @NASAKennedy

2 months ago

The planet can spell your name – literally. 🔤🌍 This Earth Day, see your name written in landscapes captured by Landsat: go.nasa.gov/4ak4Cdu

2K 28K 184K 56.4M 77K

The internet was architected for a different era entirely one where scale meant millions of users, not billions of automated agents running simultaneously. We're watching infrastructure bend under a load it was never designed to bear. What unfolds from here will reshape everything we thought we understood about how information moves.

なおき @Naoki_GPT

2 months ago

Claude Code、金稼ぎに特化させれば余裕で月50万は超えるのになあ。 うちは10個のSNSアカを自動運用でフル稼働させてアフィで稼いでもらってる。AI自動化アフィの収益性エグすぎ。ちなAI系の強者達とも話してきたけど、AI自動化アフィのアツさに気付いてる人はほぼおらん。ガチのブルーオーシャンだよ

286 374 5K 1.6M 6K

Opus 4.7 in a nutshell, everyone.