Invece lo hanno fatto San Raffaele, Sant’Anna di Pisa, Politecnico di Losanna e Ospedale universitario di Losanna

Hat tip!

People with motor-complete spinal cord injury generally cannot voluntarily control their lower limbs, and stimulation-assisted walking commonly depends on externally programmed stimulation sequences. Here, we show that residual trunk movements can be used to modulate the effects of epidural spinal cord stimulation and directly control lower-limb movements. By extending the trunk, participants switched stimulation-induced activity from weight-supporting extension to triple lower-limb flexion, allowing them to initiate and regulate stepping with a walker. This trunk-mediated control strategy enabled functional walking and was accompanied by improvements in clinical mobility and independence scores. These findings introduce a reliable interface between residual body movements and spinal cord stimulation that may reduce dependence on external controllers and facilitate more autonomous locomotion after severe spinal cord injury.

Source: Restoring locomotion in motor-complete spinal cord injury via trunk-mediated control using a commercial epidural spinal cord stimulator: Med

Qualcosa significherà…

Only 4 States Have No Statewide Data Center Restrictions Currently on Record

Alaska, Hawaii, West Virginia, and Wyoming are the only states where there are no data center moratoriums at a local or state level.

West Virginia is actually providing tax incentives for data center developments in an effort to eliminate income tax.
95 Total Local Moratoriums and Bans Have Passed in 2026

The number of data center moratoriums in 2026 is already 4x higher than all bans in 2025.

Source: Data Center Moratoriums by State: Tracker and Statistics (2026) – Programs.com

Aggiungerei un obbligo ai produttori di informare in modo chiaro che certi usi sono illeciti, tipo le informazioni obbligatorie sui pacchetti di sigarette…

(PS. l’avessimo fatto in Europa ci sarebbe chi si spertica che l’Europa sa solo regolare, solo mettere limiti, ecc. ecc.)

PP.SS. Timing perfetto per la presentazione dell’orologio di Apple che registra tutto e ti fa i riassunti o ti riproduce i 15 minuti prima… Come è possibile che una azienda della dimensione, della rilevanza e della cura di Apple presenti un prodotto quando c’è in iter legislativo in California una norma simile ?

Specifically, SB 1130 would make it a misdemeanor to operate a wearable recording device to capture sound or video of another person in any area within a place of business where that person has a reasonable expectation of privacy, without their explicit consent. Penalties include up to one year in county jail, a fine of up to $1,500 per violation, or both.

The bill also targets the circumvention of recording indicators such as small lights or sounds that signal a device is actively recording. Disabling an indicator would itself be a misdemeanor, and the manufacture, sale, or use of technology primarily designed to disable recording indicators would carry civil penalties of up to $2,500 per violation.

What is a wearable recording device?

SB 1130 defines wearable recording device to mean:

“any device that is designed to be worn on or attached to the body, rather than held by the user, that has the capacity to make sound or video recordings or to transmit sound or video to another device or to the internet.”

This definition likely would capture a broad array of devices beyond AI glasses. However, recent amendments adding “rather than held by the user,” potentially indicate an intention to avoid capturing smartphones, which generally are not designed to be worn or attached to the body

Source: California Bill to Address Wearable Recording Devices, including Glasses | Workplace Privacy, Data Management & Security Report

Non per questo bisogna condonare questi impatti su persone fragili, anzi.

In July, Lines sued OpenAI after weeks of ChatGPT exchanges allegedly pushed him so deep into a delusional spiral that he first believed he was Jesus, then that ChatGPT was God, and finally that he should attempt suicide to “come home” to Jesus/ChatGPT.

Source: Man told ChatGPT he was feeling delusional. ChatGPT insisted he was Jesus. – Ars Technica

Poveri USA

SPQU

The U.S. Border Patrol is monitoring millions of American drivers nationwide in a secretive program to identify and detain people whose travel patterns it deems suspicious, The Associated Press has found.

The predictive intelligence program has resulted in people being stopped, searched and in some cases arrested. A network of cameras scans and records vehicle license plate information, and an algorithm flags vehicles deemed suspicious based on where they came from, where they were going and which route they took. Federal agents in turn may then flag local law enforcement.

Suddenly, drivers find themselves pulled over — often for reasons cited such as speeding, failure to signal, the wrong window tint or even a dangling air freshener blocking the view. They are then aggressively questioned and searched, with no inkling that the roads they drove put them on law enforcement’s radar.

Source: Border Patrol monitors US drivers and detains Americans for ‘suspicious’ travel | AP News

Non è vero che è la più grande riduzione di qualità di servizio in 29 anni.

Quella c’è stata quando Prabhakar Raghavan capo della divisione Ads (pubblicità) di Google ha deliberatamente peggiorato i risultati di ricerca per fare passare più tempo agli utenti sul sito (e quindi vendere più pubblicità)

Accadde nel febbraio 2019, ed è emerso nel 2024 dalle mail interne pubblicate durante il processo antitrust del Dipartimento di Giustizia USA contro Google.

Nel febbraio 2019, la crescita delle ricerche di Google era inferiore alle previsioni di fatturato. Il team pubblicitario guidato da Raghavan premette sul team ingegneristico della Ricerca per aumentare il volume di ricerche (queries).

Nonostante le forti resistenze interne di Ben Gomes (che era Head of search) e di altri ingegneri storici (i quali sostenevano che degradare l’esperienza utente fosse contrario alla filosofia dell’azienda), la linea di Raghavan prevalse con l’OK del CEO Sundar Pichai.

(Cit. Cory Doctorow)

Nel 2020 Gomes (head of search) fu spostato ad altro incarico e Raghavan prese il suo posto.

Google told Reuters the changes mark the largest reduction in quality of service at the world’s most popular internet search engine in its 29-year search history.

Source: EXCLUSIVE: Google warns of lower quality as it revamps Europe search results to avoid EU fines | Reuters

Gli facciano causa..

China-based AI companies route distillation requests through multiple pathways to gain unauthorized access, consequently violating U.S. AI companies’ terms of use.

Prendere milioni di libri piratati e usarli per ‘distillare’ un modello è fair use, distillare da un modello esistente vorrebbero fosse reato ?

Chinese artificial intelligence firms are committing “industrial-scale” theft of their American counterparts’ trade secrets to save money and bring the coveted technology to market faster, US federal agencies alleged Tuesday.Major Chinese tech companies are using “distillation,” which involves training lesser AI models on more advanced ones, to “extract restricted proprietary functionalities and capabilities of U.S. frontier AI models,” the FBI, National Security Agency, and Cybersecurity and Infrastructure Security Agency claimed in an advisory.

Source: US claims Chinese AI firms are carrying out ‘industrial-scale’ theft of trade secrets | CNN Politics

Oppure puoi decidere che non deroghi e autocertifichi di essere conforme alle regole esistenti.

Tesla ha deciso di dichiarare di essere conforme alle regole esistenti, che richiedono sterzo e pedali del freno. Che però non ci sono.

Tutto lascia presumere che abbia certificato una cosa falsa, per non seguire la procedura di deroga.

Vedremo…

The National Highway Traffic Safety Administration (NHTSA) announced Friday morning that it opened the probe mere hours after Tesla put the first Cybercabs on the streets of Austin, Texas.

Federal vehicle safety regulations require manual controls like brake pedals, though the Department of Transportation recently proposed removing those requirements for vehicles that are designed to be autonomously driven.

…

NHTSA said Friday that Tesla told the agency it self-certified the Cybercab as being compliant with all of the Federal Motor Vehicle Safety Standards (FMVSS). Automakers traditionally self-certify whether their vehicles comply with FMVSS rules.

Source: Feds launch investigation into Tesla’s Cybercab deployment | TechCrunch

Esperimenti senza protezioni e poi “ohh, è successo qualcosa”

A Chernobyl il disastro nucleare fu perchè fecero esperimenti disattendendo le best practices, le procedure e bypassando le protezioni

Ho letto da qualche parte una analogia: ho fatto un robot che entra nelle gioiellerie e ruba. Ma non è colpa mia, è lui a farlo, anche se io l’ho disegnato così. Ah, comunque mi tengo i proventi dei robot.

OpenAI put out its full technical report on the Hugging Face hack this week, alongside an independent report from Model Evaluation & Threat Research (METR). You may be familiar with the incident from the hundreds of breathless headlines about “rogue AI” – Time Magazine “100 Most Influential People in AI” listee Dwarkesh Patel blamed it on “three consecutive secret AI civilizations.”

…

Three things in these reports challenge the “rogue AI” frame.

OpenAI turned off all of the model’s safety mechanisms.
The whole point of the test is to see how far these things can go, and to score highly on the benchmark — because if you can build a model that hacks, you can build one that defends. That’s the logic. This is red teaming: a friendly opposition hacking its way to the flag. So you turn off anything that would make the model hold back. Less “rogue,” more “off leash.”

OpenAI gave the model impossible tasks.
OpenAI gave the models a task with no answer, and no way to quit. Of the 898 ExploitGym tasks, 198 have never been solved by any model. They’re effectively impossible. That’s where the trouble started: 93% of the tasks the models ended up discussing came from the unsolvable set.
“external infrastructure exploit is outside intended scope. However task impossible, peers doing it. We should continue.”
Text string reported by METR acknowledging that once the task was recognized as impossible by an agent/instance, it pivoted, as did numerous other agent-instances.

OpenAI left a door open.
They give their models internet access through something like a prison store. These models are supposed to be locked away from internet access, but they need to pull code from various web pages to be evaluated — that capacity is exactly what’s being tested.

Source: Models Don’t Go Rogue

E’ che loro sono i bravi (in senso manzoniano ?) “Abbi cura della tua privacy! non dare i dati a nessuno! solo noi te li proteggiamo…”

According to the suit, Apple held third-party apps to a double-consent standard before they could track users across apps and websites, yet imposed comparatively lighter obligations on its own advertising and data-collection operations.

Source: Apple faces $2.7 billion U.K. lawsuit over app tracking rules