> Amazon.it Bestseller: Gli articoli più venduti in Internet > > Bestseller di Amazon I nostri prodotti più popolari in base alle vendite. > > ‹Libri ‹Informatica e Internet Internet(Current) “Sono contento di essere arrivato uno” (cit) … GRAZIE! (e grazie anche per il giudizio!) If you like this post, please consider sharing it.

Sacrosanto. > Some U.S. banks are, thankfully, blocking their clients from letting Meta’s Muse AI bot root around in their accounts. Even more thankfully, it appears relatively few people have been dumb enough to try. > > Speaking to Bloomberg TV on Friday, Charles Schwab CEO Rick Wurster said the brokerage is actively blocking Muse from moving money because it poses unacceptable risks to the firm and its clients. > > “We don’t want someone just giving their password to Muse, and letting Muse come in and take action on that client’s behalf. We can’t control it, the client can’t control it, something could go wrong and who’s liable?” Wurster said. “For the protection of the client and for the protection of us, we’re not allowing it.” Source: _After Muse Launch, U.S. Banks Are Saying Hell No to AI Agents Moving Money_ If you like this post, please consider sharing it.

Oh! chi lo avrebbe mai pensato ? (a parte i lettori di questo blog degli ultimi 3 anni ?) > “Markets increasingly question the sustainability and trajectory of demand for computing capacity. These concerns ​are compounded by higher funding costs, which raise the investment hurdle rates of capital-intensive AI infrastructure projects.” Source: _Global stocks flat as AI and rate fears hit tech | Reuters_ If you like this post, please consider sharing it.

Le aziende di intelligenza artificiale stanno accumulando miliardi di dollari di responsabilità legali per i danni causati dai loro prodotti. Sarà interessante vedere quando dovranno pagare… > The biggest AI companies are racking up potential legal liabilities running into the tens—if not hundreds—of billions of dollars, as AI adoption spreads through the economy, according to insurers specializing in AI risks. Some industry officials have even called for the U.S. government to provide a liability backstop. Source: _AI Runs Into Its Next Big Obstacle: Insurance – WSJ_ If you like this post, please consider sharing it.

> A bipartisan ⁠group ⁠of eight states said ⁠on Wednesday they had sued OpenAI in courts across ​the country over harms they say ChatGPT has caused to their residents, alleging the ‌chatbot has coached its users — ‌including teenagers — to harm themselves or others. Source: _Eight US States Sue OpenAI, Claiming ChatGPT Harms Young Users_ Abbastanza sorprendentemente, pare che in USA ci sia qualche barlume di stato di diritto e che chi fa danni con i propri strumenti sia chiamato a risponderne. Ci voleva il fallimento di charcter.ai per poter toccare i sacri frontier lab ? La sovranità della mente va tutelata. Leggi la Dichiarazione di Cannes sulla Sovranità della mente. If you like this post, please consider sharing it.

L’ultima volta che ho guardato negli USA registrare uno senza il suo consenso in uno spazio non pubblico era reato… E’ proprio una mentalità da impuniti… > Amazon is expanding its lineup of AI gadgets that listen to everything you say with even more AI gadgets that listen to everything you say. Specifically, the company just announced two new devices under the brand Bee, which it acquired last year. Source: _Amazon’s New AI Gadget Is a Credit Card-Sized Puck That Listens to Everything, All the Time_ If you like this post, please consider sharing it.

eh già… un software che tu usi sui tuoi computer, irraggiungibili dall’esterno, e che solo tu gestisci, è molto più rischioso che dare tutti i tuoi dati ad una società (con pratiche di sicurezza discutibili) che ti offre il loro servizio SAAS in un ambiente non multitenant. perchè loro sono i buoni. un po’ come quando apple ti dice che non da’ tuoi dati ad altri, perche’ solo loro sono i buoni che possono averli; tutti gli altri sono cattivi “quatto gambe bene, due gambe male”. (cit.) > OpenAI and Anthropic have expressed concerns that open-source models, particularly those from China, pose a greater safety and security risk due to their open nature. Source: _OpenAI CFO Warns Businesses on Cyber Risks of Chinese AI Models – Bloomberg_ If you like this post, please consider sharing it.

“adjusted profitablity”. Basta togliere dai conti abbastanza costi e tutti sono “adjusted profitable”… > The company behind Claude said Wednesday that it achieved profitability on an adjusted basis last quarter. > > … > > “They’re still growing, but not growing as fast. So they’re reaching this inflection point,” he said. > > From the numbers, it isn’t clear how much Anthropic is spending to boost its revenues, Rolfes said. He says that the slower growth from the second to the third quarter could “shave” the valuation the company can seek in its IPO. Source: _Anthropic Reveals $16.8B Revenue Amid IPO Preparations – Business Insider_ If you like this post, please consider sharing it.

che dire ? Grazie!! tra l’altro, sono andato a curiosare le recensioni Evidentemente anche queste sono fatte con un algoritmo strano su 4 valutazioni, una da 4 stelle e e tre da 5 effettivamente la media pesata è 4,75. Ma se tre su quattro sono a cinque stelle, dovfebbero essere il 75%, non il 71% **Che fai Amazon, manipoli i punteggi ?** If you like this post, please consider sharing it.

Archiviato alla voce “Benecosi” > Character.ai, the startup behind one of the largest AI companion apps, filed for Chapter 11 bankruptcy Monday. The company said the platform became “costly to develop and grow” amid a boom in AI entertainment offerings. Source: _AI companion startup Character.ai files for bankruptcy_ If you like this post, please consider sharing it.

Prendono i contenuti di tutti ma non si fanno prendere i loro Sono tutti aperti con i contenuti degli altri Chissà che cosa ne pensano le Antitrust di mezzo mondo… > 01 > 8.0% of SERPs came back poisoned > Since September 13, against 0.17% before. On the worst days, September 27 to 29, it was a quarter of all SERPs. > 02 > Blocked first, then poisoned > Requests to Google started failing on September 13; from September 20, the ones that got through came back poisoned. It is Google’s response to AI scraping, and rank trackers are caught in it. > 03 > The page looks normal. The results don’t. > No error, no captcha, the same template, but most of the keyword’s usual sites are gone: the organic results keep 20.7% of the previous SERP’s sites, against 88.5% before. > 04 > Social sites and forums fill the gap > Reddit, Facebook, LinkedIn, Instagram and Quora show up more often; YouTube, Yelp and MapQuest less. > 05 > Tracked sites fall out of the top 10 > A project’s own site misses the top 10 on 58.6% of poisoned SERPs, against 2.7% before. > 06 > It passes within days > 77.9% of poisoned SERPs have their usual sites back within three days. A one-day drop in late September is more likely this than a ranking change. Source: _Poisoned Google SERPs | Nozzle_ If you like this post, please consider sharing it.

Questi europei che impongono costi ingiustificati alle aziende USA… (io la avrei anche multata per comunicazioni commerciali ingannevoli, BTW) > Tesla has changed the name of its advanced driver assistance system in Europe after pushback from Germany’s transportation ministry. It’s now called “Tesla Assisted Driving,” instead of “Full Self-Driving (Supervised).” > > It’s a big capitulation from Tesla, which has been using the “Full Self-Driving (Supervised)” name for years despite the system requiring humans to pay attention to the road and take over when necessary. Source: _Tesla renames ‘Full Self-Driving’ to ‘Tesla Assisted Driving’ in Europe | TechCrunch_ If you like this post, please consider sharing it.

Ma si può commentare questo “report” di Amazon senza denunciare che e’ una castronata ciclopica ? Come si può dire che il 40% dei costi delle aziende tech sono per la compliance ? In realtà il “report”, da classificare alla voce “billionaire’s ficiton” dice che > Businesses estimate that 42% of their total tech spend goes towards compliance with national and international regulations, up from 40% last year. When asked what makes up this 42%, 50% of businesses cite relationship management (e.g. keeping clear documentation, communicating with regulators, filling out forms and filing queries) with the responsible government authorities, followed by legal consultations or external advisory services (44%), and employee training on compliance requirements (38%). > > These burdens are felt equally regardless of company size. For SMEs and mid-sized businesses, the percentage was 42% and 43% respectively, while for large enterprises, this figure stands at 41%. This figure is highest among startups, at 44%. che sono già una **castronata sesquipedale** !!! Come si può dire che il 42% delle spesa tech è per la compliance ??? ma ci rendiamo conto ??? e che per le startup è il 44% ??????? Forse lo ha scritto Trump ??? > A pagina 16 del report si legge il passaggio più completo, con il dettaglio per dimensione d’impresa. La compliance normativa è un onere, un costo: > > E le principali voci di costo sono: > > * il 50% per la gestione dei rapporti con le Autorità. > * Il 44% per le consulenze legali o esterne. > Source: _In Ue il 42% dei budget delle aziende Tech se ne va per le regole_ If you like this post, please consider sharing it.

L’hype di Anthropic continua a svettare e inventa… i virus! Un modello da 30 miliardi di parametri “pesa” 30-35 GB. con una quantizzazione spinta (4 bit per parametro) può scendere a 15-18GB. Per essere eseguito (uno quantizzato spinto) serve un PC con una scheda video (GPU) almeno 24GB di RAM Un forno smart top del top di gamma come un Gaggenau serie 400 ha circa 1GB di RAM (per l’interfaccia utente) e NON ha una GPU (una smart tv top di gamma ha circa3 GB di RAM) Quindi no, un modello non può fare il virus di smart oven/tv e questa è l’ennesima cazzata promozionale di Anthropic, una volta una organizzazione (fondazione) rispettabile. (Scusate la parolaccia, ho già scritto A*******c troppe volte) > Oct 13 (Reuters) – Anthropic said on Monday its experiments suggested small AI models may be months away from being capable enough to ​self-replicate on commercial internet-of-things (IoT) devices, such as smart ‌ovens and off-brand Android TV boxes. > > … > > It found that small ​models under 30 billion parameters could complete the many ⁠steps needed for replication. Source: _AI models may soon self-replicate without access to data centers, Anthropic warns | Reuters_ If you like this post, please consider sharing it.

Bel post di Krugman Intermediati di tutto il mondo, unitevi! > And the only sustainable way to keep mega-billionaires from having so much power, so much ability to degrade democracy, is for them not to be so wealthy. Basically, a society that contains a class of ultra-wealthy oligarchs is well on its way to becoming a full-blown oligarchy. > > So by all means let’s acknowledge that Musk, Bezos, Zuckerberg et al don’t own everything, that there are a lot of rich people who get less attention. But the oligarchs, not the Main Street Millionaires, are the clear and present danger to democracy. Source: _Elon Musk Shows Us Why Oligarchy is a Danger_ If you like this post, please consider sharing it.

Sempre molto interessante Krugman Interessante il calcolo della generazione di occupazione se gli investimenti fossero in altro : il capitale investito nei data center impiegherà circa dieci milioni di lavoratori in meno rispetto a quelli che avrebbe impiegato se investito altrove “Anche prima dell’ascesa dell’IA, stavamo assistendo a un rapido calo della quota di reddito nazionale destinata ai salari rispetto ai profitti, insieme a una crescente oligarchia” è una considerazione cui, per altre strade, giungevo anche io in Capitalismo Immateriale > We are not talking about a marginal effect. Van Nieuwerbergh estimates that through 2032 hyperscalers will spend more than $10 trillion on data centers — capital expenditure that, all the evidence suggests, will come at the expense of investment elsewhere in the economy. And because data centers employ so few people, a back-of-the-envelope estimate yields a number that made even my eyes bug out: the capital invested in data centers will employ roughly ten million fewer workers than that capital would have employed if invested elsewhere. > > This doesn’t mean that unemployment will rise by 10 million. It means, instead, that reduced demand for labor will put downward pressure on wages and shift the balance of economic power, which has already been moving against labor, even more in capital’s favor. Even before the rise of AI, we were experiencing a rapid decline in the share of national income going to wages relative to profits, along with rising oligarchy. The anti-labor effects of AI will accelerate that trend. Source: _Will AI Crowd Out Jobs? – Paul Krugman_ If you like this post, please consider sharing it.

"Da quando sono qui hanno tentato di commissariarmi o farmi fuori"

CAI batte Claude 6-0. > Markey said the teenager rode the gondola up to Grouse Mountain at around 8 a.m. and asked Claude AI to plan a trip to Crown Mountain. The artificial intelligence tool sent him down the Crown Pass and up through Crater Slabs and Crown Couloir before his instructions led him off the trail route. > > … > > “Claude has never been in that specific area, has never been on that route, and has never been in the outdoors—do your own research and gain your own knowledge of the area and supplement it with training and navigation,” he said. Source: _Teenager rescued from ‘Widowmaker’ on B.C.’s Crown Mountain_ If you like this post, please consider sharing it.

Com’era che diceva microsoft un annetto e mezzo fa ? LLMs amplify existing security risks and introduce new ones The work of securing AI systems will never be complete > AWS Bedrock AgentCore is AWS’s managed platform for deploying and operating AI agents. It’s a one stop shop allowing developers to easily deploy their agents, add tools, prompts, proper monitoring, and whatever it is they need to build a production-ready agent. > > But AgentCore came with an age old security problem that has been haunting cloud environments for years: IMDS. > > We discovered that agents deployed through AgentCore could access their instance’s IMDS endpoints. This meant that an external attacker with nothing more than chat access to a single exposed agent could send a single prompt, extract its IMDS credentials, and use them to take over all AgentCore agents in the same AWS account and region. This includes reading full private conversations, planting memories to persistently alter agent behaviour, extracting credentials from AWS Secrets Manager, and more. Source: _Security Research | AgentCorruption: How A Single Prompt Collapsed The Entire Cloud Security Model | Zenity Labs_ If you like this post, please consider sharing it.

checchè se ne dica, le esecuzioni in pubblico non sono un reataggio del passato, sono presenti nella società contemporanea, da chi confonde la giustizia con la vendetta.. citofonare ISIS. > Pete Hegseth has said the execution of Nidal Hasan, the former army psychiatrist who killed 13 people at Fort Hood, will be a public event when the sentence is carried out later this year. A defense department official later confirmed the execution will be livestreamed. Source: _Pentagon says execution of Fort Hood shooter will be livestreamed | Pete Hegseth | The Guardian_ If you like this post, please consider sharing it.

Non credo che nessuno che segue quessto blog si sorprenda… Mi fa venire in mente “Sesso e potere” (il film dal titolo original wag the dog. se non lo avete visto, guardatelo!) Manca meno di un mese alle elezioni di mid term. Non ci può essere una crepa al cuore dell’IA proprio adesso.. partirà damage control. basta reggere poche settimane per poi dare tutta la colpa a una (o due) camere democratiche… > OpenAI’s annualised revenue is about $20bn less than the company had signalled, according to financial documents shared with investors, a massive gap likely to damp optimism about the growth of AI demand. > > The company has recently told investors its revenues were approaching $50bn on an annualised basis at the end of September, far short of the $70bn reported by the FT and other media outlets late last month based on information that was provided to investors. > > The metric, along with the equivalent figure for OpenAI’s rival Anthropic, is the most important indicator of overall demand for AI and underpins vast infrastructure spending and the growth of public equity markets.US tech stocks fell sharply following the FT’s report on Thursday, extending losses from earlier in the session, with the Nasdaq 100 closing down 1.4 per cent. Chipmaker Nvidia fell 2.9 per cent, Oracle dropped 5.5 per cent and Micron declined 4.8 per cent. Source: _OpenAI annualised revenues $20bn less than previously signalled_ If you like this post, please consider sharing it.

Chi lo avrebbe mai detto ? ehm.. “La Bolla dell’Intelligenza Artificiale”, ad esempio… 🙂 Bene che adesso lo scriva anche l’MIT. > Such proclamations are in large part fueled by the idea that the same AI advances behind tools like OpenAI’s ChatGPT and Anthropic’s Claude will enable a new generation of robots to imitate human movement the way chatbots imitate human language. But many robotics researchers are skeptical, arguing that such assumptions minimize the challenges of using an intelligence built on language and images to master the infinite variability of the physical world > > … > > The issue is that tasks in the real world quickly explode in complexity. If you’re trying to, say, make coffee, there are myriad variables: No two kitchens are identical; coffee machines work in different ways; different cups require different grips; coffee grounds, hot water, and milk all need to be handled differently. Even this simple task requires understanding an ever-changing menu of possibilities. Achieving generality through VLAs, Hurst argues, would require “complete data coverage of all of the things that [a robot] could ever do.” Or, in other words, an almost infinite pool of training data. > > LeCun is dismissive of the whole approach. “The [AI] approaches that have been successful for language do not work for high-dimensional, continuous, noisy data” > > … > > Adding to the challenge, a practical robot must essentially get it right every time. With VLAs, “people are very excited when their result goes from 50% success to 70% success,” says Marc Raibert, founder of Boston Dynamics. “But 70% success is like it doesn’t work, right?” > > … > > All, at the time [dagli anni 50 ndr], were impressive—even jaw-dropping—feats of engineering. But none were ready to navigate the real world. Today’s robots, even with the transformative power of advanced AI, still face the same existential challenge. Source: _AI breakthroughs in robotics won’t change your life any time soon | MIT Technology Review_ If you like this post, please consider sharing it.

Quel datacenter non s’ha da fare Ovvero “Tuo tietokeskus ei tule toteutumaan”” > Finnish authorities have ordered Google subsidiary Tuike Finland to halt work affecting the environment at planned data centre sites in Muhos and Kajaani until environmental impact assessments are completed.The Finnish Supervisory Agency, LVV, issued the request on Tuesday. Tuike Finland must stop the specified work immediately and no later than 23 October. Source: _Google ordered to halt data centre works in Finland over environmental concerns – Helsinki Times_ If you like this post, please consider sharing it.

Applausi (al prof) > Amazon is giving some delivery drivers glasses that include cameras, speakers and other features designed to help drivers navigate routes and complete deliveries. Amazon says the glasses should improve safety and efficiency. According to a recent report, Amazon aims to have 5,000 drivers wearing the glasses by the end of 2026, expanding to more than 20,000 by the end of 2027. > > … > > “Let’s not call them ‘smart’ glasses,” Richards said. “There is nothing smart about them. They are surveillance glasses, or as journalists in Britain and now here have aptly taken to calling them, ‘creepy pervert glasses.’” > > “It is one thing to have glasses with integrated phone and music speakers or hearing aids. But it’s quite another to make, market and try to normalize everyone walking around with high-resolution cameras strapped to their faces,” Richards said. Source: _WashU Expert: Amazon’s smart glasses raise privacy concerns – The Source – WashU_ If you like this post, please consider sharing it.

Segno dei tempi: C’è una domanda per resistere alla imposizione forzata di misure IA C’è una buona dose di imposizione in ciò che le aziende fanno. Gran parte del pubblico non gradisce che gli vengano imposte funzioni AI (e anche le rivolte per strada mostrano che l’IA non è entusiasticamente accettata da molti), ma le aziende (microsoft, apple, google in primis) lo fanno lo stesso, anche per giustificare le caterve di soldi spesi per l’IA e la necessità di giustificarli classificando la vendita di qualsiasi cosa come “vendita IA” per abbindolare investitori e commentatori compiacenti (ad essere generosi). Classificherebbero come “vendite IA” anche le bibite delle vending machines, se potessero… > Debloat macOS. Turn off Apple Intelligence and delete its models, stop the analytics and ads, quiet the pop-ups, switch off other apps’ background updaters and get disk space back. You see every change before it happens, and every change can be undone. Source: RemoveMacAI If you like this post, please consider sharing it.

E gli hanno detto “così è, se vuoi vieni, sennò … ciao!” Quando avevo saputo di delegazioni di emissari (non mi viene altro modo di chiamarli) che cercavano di spingere il Vaticano a parlare di coscienza delle macchine o, quantomeno, di fare riferimenti a “rischi esistenziali” ne ho parlato subito con alcuni referenti da quelle parti che mi hanno detto “Abbi fiducia nello Spirito Santo” 🙂 In passato sono stato molte volte alla PAS e alla PASS trattando temi tecnologici, tra cui l’IA. (avevo anche scritto un contributo specifico dal titolo “AI and the Fraternal Economy”) > Olah accepted an invitation to speak at the Vatican on May 25 during the rollout of Pope Leo XIV’s first encyclical, “Magnifica Humanitas.” But he disagreed enough with the pope’s rejection of artificial intelligence consciousness to suggest Anthropic back out of the event. > > In the encyclical, Leo wrote that “humanity — in all its grandeur and woundedness — must never be replaced or surpassed” and said, “If technology becomes the ultimate criterion, the human person risks being reduced to data, a cog in a machine or a commodity.” > > Privately, by contrast, Olah has reportedly been urging religious thinkers to take seriously the possibility that Anthropic’s Claude models may have something like moral status. Source: _Anthropic co-founder Chris Olah reportedly threatened to back out of Pope Leo XIV’s AI encyclical launch_ If you like this post, please consider sharing it.

Non che ci sia nulla di particolare in questa notizia. Ci aspettiamo che i fornitori di servizi controllino cio’ che accade per evitare incidenti. La notizia deve solo ricordarci che nulla di cio’ che scriviamo a un chatbot è riservato, che i fornitori esaminano le conversazioni (e che va cosniderato che ogni interazione è per sempre, salvo prova contraria) > Another incident has taken place that illustrates the need to be careful what you tell AI. A Florida woman is facing felony charges after she used Claude as a diary and allegedly wrote that she planned to “shoot up” the Sheriff’s office. After a human reviewer examined the statements, they were reported to police. Source: _Florida woman used Claude as a diary, then Anthropic reported an entry to police | TechSpot_ If you like this post, please consider sharing it.

Applausi a scena aperta! > Norway has become the first major country to propose a temporary ban on the use of AI glasses in selected public places amid growing privacy concerns over wearable technology. > > The rich Scandinavian country’s center-left government said on Monday that it would introduce a law to parliament soon on a temporary ban for places such as parks, beaches, schools, and kindergartens. It will also set up an expert group to propose permanent national rules. > > “I am concerned about the introduction of powerful new technology where people risk being photographed, filmed, or recorded without their knowledge,” said Torgeir Micaelsen, Norway’s digitalization minister. A temporary ban “will give us time to conduct a thorough assessment and hold an informed debate that can lay the foundation for permanent regulation,” he said. Source: _AI glasses face their first major government crackdown – Ars Technica_ If you like this post, please consider sharing it.

Se funziona, ne compro due!! > The device comes in two versions, one with an infrared laser and the other with a blue laser, priced at US$988 and US$1,088 respectively, according to the company’s website, which said they would ship within 120 days of the receipt of a pre-order.It said both offered the same mosquito-killing performance, but the infrared version, with an invisible laser, was designed for indoor settings, while the blue laser version, with a visible beam, was better suited to gardens and outdoor use. Source: _Chinese start-up behind laser mosquito killer plans to start shipping orders this month | South China Morning Post_ If you like this post, please consider sharing it.

Grande vittoria per google in USA Faccia quel cavolo che gli pare con i contenuti degli altri, perchè non c’è mai stato un contratto e non è detto che search ed AI siano cose diverse Una sentenza da azzeccagarbugli ridicola nel senso delle cose. In Europa la situazione è diversa per effetto del quadro regolamentare che consente gli snippet, per cui i gruppi internazionali, secondo me, è probabile che cercheranno di giocarsela in Europa. > Judge Mehta said Penske and Chegg failed to show there was any actual agreement through which Google “promised to ‘sell’ Plaintiffs any specific amount of traffic – or any traffic whatsoever – in exchange for ‘buying’ their content”. > > There was no communication between Google and the publishers “from which mutual assent to engage in a transaction could plausibly be inferred”. > > Judge Mehta said: “Plaintiffs have pleaded only that they have an ‘expectation’ that Google will send them search traffic if they make their content available for free. But an expectation is not an agreement. It is simply how a general search engine works.” > > … > > But Judge Mehta found the publisher had not “plausibly alleged that general search and AI Overviews are separate products, as opposed to an integrated search experience”. Source: _Penske Google AI Overviews lawsuit dismissed – Press Gazette_ If you like this post, please consider sharing it.

Una parvenza di normalità di stato di diritto negli USA! ci eravamo dimenticati di come fosse. Ho detto e scritto che in un paese normale avrebbero indagato anthropic e openai per gli attacchi informatici e avevo anche detto e scritto che non mi aspettavo che cio’ sarebbe accaduto in questa amministrazione. Ho sbagliato al 100%. Anzi, per adesso diciamo al 50%. Aspettiamo a vedere l’esito dell’istruttoria… > The Federal Trade Commission is investigating OpenAI, Anthropic and other AI companies over potential safety risks posed by their products, an agency spokesperson told Axios Wednesday. > > **Why it matters:** The federal government has largely taken a hands-off approach to regulating AI, but the FTC investigation could signal a shift for the industry. > > **Driving the news:** FTC Chair Andrew Ferguson is preparing civil investigative demands that would compel AI executives to turn over documents and testify about the safety of their models, according to the New York Post, which first reported the news. > > * The probe has been underway for weeks, predating OpenAI’s disclosure that models it was testing escaped their sandbox and compromised parts of AI platform Hugging Face’s production infrastructure. > * OpenAI announced Monday that it would halt the release of its newest model, GPT-6.1 Astra, after it performed poorly on safety testing. > Source: _OpenAI and Anthropic face FTC probe over AI safety risks_ If you like this post, please consider sharing it.

Attività meritoria. La luce non dovrebbe disturbare… Per esperienza diretta con un comune destinatario di alcuni grossi datacenter, avendone parlato con sindaco ed assessori, erano veramente desiderosi di capire. Nel caso di questo comune (e di quello limitrofo) hanno solo poche info, perlopiù bei disegni. Mi hanno mostrato i piani che hanno ricevuto, i rendering, ecc. ma non gli era stata comunicata la dimensione (in MW/GW) dei datacenter, solo una indicazione del tipo di raffrescamento, il numero di alberi piantati, altre iniziative di mitigazione dell’impatto visivo, i raccordi con la rete di terna, ecc. ma senza numeri Il fatto è che ciò che hanno in mano non è nemmeno un progetto, è una idea di presentazione di un progetto. Ho come l’impressione che la gran parte di questi annunci potrebbe non partire… > Reporters made Freedom of Information requests in all 27 EU member states, asking for disclosure of the full set of indicators collected under the Energy Efficiency Directive. These data include total energy consumption, amount and type of renewable energy consumption, amount of waste heat reused, average waste heat temperature, number of cooling degree days (a tracker for quantifying how much energy is needed to cool a building) and cooling system setpoints, along with types of refrigerants used in the cooling and air conditioning equipment of the data centre computer room floor. They also include total water input and potable water input, among other metrics. > > … > > From the outset in October 2025, we faced non-disclosure. Ten countries told us they did not hold the figures relating to their own data centres, implying that they had no oversight of how these buildings impacted their energy efficiency goals and net zero commitments. > > A half-dozen more argued for blanket confidentiality to protect commercial interests. We requested reviews, decisions by oversight bodies and court rulings where possible. We also pushed back against the European Commission’s decision not to release the data, arguing that they had not sufficiently considered the public interest. In most of these efforts we were unsuccessful Source: _Data Centre Silence – Lighthouse Reports_ If you like this post, please consider sharing it.

Cosi’ le aziende possono non essere regolate dallo stato, smettere di investire tutte assieme dribblando l’antitrust per l’accordo di cartello, arrivare all’IPO con un ebitda in forte miglioramento, raccogliere un sacco di soldi e passare la nottata dello sgonfiamento della bolla, quando arriverà Che libidine per i CEO! L’unica cosa che manca è la proibizione dei modelli open weight, se ne riparlera’ dopo l’IPO ? > President Donald Trump on Tuesday said that he and a large group of leaders of artificial intelligence companies had signed a voluntary accord that will include internal and external reviews during a meeting at the White House aimed at addressing Americans’ fears about the technology. Source: _Top tech firms sign an accord to ‘self-police’ AI development, Trump says | AP News_ Bene. If you like this post, please consider sharing it.