Agregátor RSS

Filmy o věznicích, vězních a jejich útěcích se natáčejí už desítky let. Vybrali jsme ty nejlepší – od vážných dramat, která poukazují na problémy systému, až po akční, dobrodružné i komediální příběhy. Zkrátka filmy, kde hraje vězení klíčovou roli.

European Union member states and the European Parliament agreed early Thursday to push back the toughest deadlines under the bloc’s AI Act, giving enterprises more time to prepare for high-risk compliance.

Under the provisional deal between negotiators for the European Parliament and European Council, high-risk AI systems will face new deadlines of Dec. 2, 2027 for stand-alone systems and Aug. 2, 2028 for AI used in products covered by EU sectoral safety rules, a European Parliament statement said. The original deadline was Aug. 2, 2026.

The deal still needs formal adoption by both Parliament and Council before it can enter into law. The co-legislators intend to complete that step before Aug. 2. Until they do, the original deadline applies as drafted.

“Today’s agreement on the AI Act significantly supports our companies by reducing recurring administrative costs,” Marilena Raouna, Cyprus’s deputy minister for European affairs, said in a statement from the Council, which is composed of representatives of each of the EU’s 27 member states. Cyprus holds the rotating presidency of the Council, which negotiates on behalf of member states.

The breakthrough comes nine days after previous discussions collapsed without agreement.

Fewer restrictions, more time to implement

The provisional agreement removes overlapping rules for AI in machinery products, Parliament said. These will now follow only sectoral safety rules, with safeguards meant to ensure equivalent health and safety protection.

It also narrows what counts as a “safety component” under the AI Act. AI features that only assist users or improve performance will not automatically be treated as high-risk, the Parliament said, as long as a failure does not create health or safety risks.

For wider sectors such as medical devices, toys, lifts, machinery and watercraft, the co-legislators agreed on a mechanism to resolve overlaps between the AI Act and existing sectoral laws, the Council said in its statement.

The deadline for member states to set up AI regulatory sandboxes has been pushed back by a year to Aug. 2, 2027, the Council said. Watermarking obligations on AI-generated content, on the other hand, will apply earlier than the Commission proposed, from Dec. 2, 2026 instead of Feb. 2, 2027, the Parliament said.

Mid-size firms get more breathing room. Exemptions previously available only to small and medium-sized enterprises now extend to small mid-cap companies, the Council said. The deal also clarifies that the EU’s AI Office will supervise general-purpose AI systems centrally, with national authorities keeping responsibility in areas including law enforcement, border management, judicial authorities and financial institutions.

“With this agreement, we show that politics can move just as quickly as technology,” said Arba Kokalari, the Parliament’s co-rapporteur for the Internal Market and Consumer Protection committee. “We now make the AI rules more workable in practice, remove overlaps and pause the high-risk requirements.”

Parliament and Council also agreed to ban AI systems that create child sexual abuse material or that depict identifiable people in sexually explicit content without consent, the Parliament said. The ban covers placing such systems on the EU market, doing so without safety measures to prevent misuse, and using them to generate the content. Companies have until Dec. 2, 2026 to comply.

“Alongside simplification measures, we are banning nudification apps, a key part of the Parliament’s mandate, and, of course, the creation of child sexual abuse material using AI systems,” said Michael McNamara, the Parliament’s co-rapporteur for the Civil Liberties, Justice and Home Affairs committee.

What still applies

Several parts of the AI Act keep moving on their original schedule. Bans on unacceptable-risk AI have applied since February 2025, according to the European Commission. The general-purpose AI rules came into force in August 2025. The transparency obligations under Article 50, including disclosure for chatbot interactions, are set to apply from Aug. 2, 2026.

The provisional agreement is part of the seventh omnibus package on simplification, proposed by the Commission on Nov. 19 last year in response to the Draghi report on EU competitiveness.

Ivanti warned customers today to patch a high-severity remote code execution vulnerability in Endpoint Manager Mobile (EPMM) exploited in zero-day attacks. [...]

Chrome začal potají stahovat lokální AI model Gemini Nano. • Pomáhá s úpravou textů, odhalováním podvodů nebo souhrny. • Soubor má 4 GB a po odstranění se znovu stáhne.

Linux administrators rely on AppArmor to contain compromised applications. If a browser, container, or Snap package is exploited, the profile is supposed to limit what that process can touch on the host. This mechanism is the backbone of Linux container isolation.

Your security controls aren't failing, they're missing where most of today's work actually happens. Keep Aware shows how browser activity like copy/paste and AI prompts bypass traditional protections. [...]

The heart’s constant motion makes it largely immune to cancer. The discovery could help protect other organs.

The heart is a biological wonder. It beats roughly 2.5 billion times in an average lifetime. Unlike skin cells, which regularly die off and regrow, a healthy adult heart hardly regenerates at all—even through all the wear and tear.

The heart has another superpower: Resistance to tumors. Nearly every tissue in the body turns cancerous, but the heart almost never does. Cancers in heart tissue show up in less than 0.3 percent of autopsies, or about 1.5 cases per million people each year.

How the heart keeps cancer at bay has baffled researchers. Pinning down its hidden defenses could inspire treatments for more vulnerable tissues, including top killers such as breast, lung, or colorectal.

Persistent mechanical strain may be the key. A new study from the University of Trieste suggests that with every beat, the heart pushing against pressure dampens gene activity tied to tumor growth. In a rather Frankenstein experiment, researchers transplanted living hearts into the necks of mice, where they survived but didn’t experience mechanical stress.

When the team injected cancer cells, the mice’s own beating hearts slowed the invasion, while the transplanted hearts were nearly overtaken within weeks. Beating heart tissue grown in the lab also fought off tumors compared to tissue that didn’t beat.

Heart cells don’t uniquely feel stress. Lung, skin, and muscle cells do too, just in different, often less rhythmic ways. It’s possible that recreating heartbeat-like forces—potentially through wearable gadgets—could extend this type of natural protection to more common cancers.

Growing Pains

Cell growth is a double-edged sword. On the one hand, it’s essential for healing and regenerating the body. The skin is constantly blasted with radiation and toxins. It suffers cuts and bruises. To repair damage, skin cells turn over every 40 to 56 days. Bombarded with chemicals from food, medications, or alcohol, the liver’s cells regenerate to keep it in working order even after substantial injury.

But cancer is the price we pay for growth. Tumors arise as cell division damages DNA. Over time, cancers grow and spread. This is why we don’t get cancer in our teeth, nails, or hair—the cells making them up are dead. Cells that rarely divide also largely escape cancer. Mature neurons barely renew and seldom form cancers. Red blood cells, which lack a nucleus and DNA, can’t become cancerous at all. Heart muscle cells are similar. Despite nonstop contraction and damage, only about one percent or fewer renew themselves each year.

This partially explains why primary heart cancers or so rare. But the organ also wards off invading secondary cancers metastasized from other tissues, which are usually far more deadly.

“Even cardiac metastases are frequently clinically silent [no detectable symptoms], with many cases identified only incidentally or at autopsy,” wrote Wyatt Paltzer and James Martin at the Baylor College of Medicine, who were not involved in the study.

It’s a paradox. The heart is flooded with oxygen and nutrients, an ideal environment for wandering cancer cells to settle and thrive. Yet they don’t. One reason may be the heart’s inability to regenerate. Previous studies have suggested that the mechanical forces of heartbeats limit cell division. The team wondered if the same forces also shield the heart from cancer.

Under Pressure

To test their idea, the researchers had to make a living heart with no beat.

“That was the most tricky part, because keeping the heart still is very difficult,” study author Giulio Ciucci told Science.

They adapted a technique used in end-stage heart failure patients to remove mechanical strain. In people, an implanted device takes over the pumping of blood. Here, the team transplanted a donor heart into a mouse’s neck and hooked it up to blood vessels. The animal’s own heart kept circulation going as usual. The transplanted heart stayed alive but didn’t do any work.

They then injected lung cancer cells, which often spread to the heart, into both organs. Within two weeks, nearly all healthy cells in the transplanted heart had been overtaken. In the beating heart, tumors rarely filled over 20 percent of a single chamber. Under constant pressure, the cancer cells struggled to divide.

One mouse with two hearts is hardly conventional. And transplantation risks immune attack and infection that could influence how cancers develop. “You have a lot of confounding factors,” said Ciucci.

So, the team moved to an “artificial heart” seeded with cancer cells, where mechanical forces could be dialed up or down in isolation. Like in the heart transplant results, the cancer spread throughout the tissue after removing strain. But it was mostly confined to the surface of beating tissues and in smaller amounts.

Looking for a reason, they compared gene activity in patient tissues with cancers that had spread to the heart, liver, and lungs and found a unique gene expression signature in the heart. In engineered tissues, mechanical stress changed how DNA was packaged, limiting access to genes related to growth and cancer. A protein on the surface of the nucleus, the cell’s DNA hub, translates physical forces from outside the cell into which genes are turned on or off. Knock this protein out, and invading cancer cells became “blind” to the heartbeat and grow freely.

Scientists have long known mechanical stress shapes cancer. As cancers grow, the cells stiffen surrounding tissue, which boosts survival, growth, immune evasion, and drug resistance. The new findings suggest that the movements of their host tissues also play a role, and the newly pinpointed protein could be a drug target.

The team is now exploring if mimicking heartbeat-like forces in other organs could prevent cancer growth. Lung, skin, and other tissues already stretch and relax, but remain susceptible.

“We really think that the key here is the continuous compression that you have in the heart,” said Ciucci. Working with engineers, they’re developing a wearable for melanoma—a type of skin cancer—that compresses the cells similar to a heartbeat. Early results look promising.

The post The Heart Rarely Gets Cancer. Scientists Think They Know Why. appeared first on SingularityHub.

The hardest part of cybersecurity isn't the technology, it’s the people. Every major breach you’ve read about lately usually starts the same way: one employee, one clever email, and one "Patient Zero" infection. In 2026, hackers are using AI to make these "first clicks" nearly impossible to spot. If a single laptop gets compromised on your watch, do you have a plan to stop it from taking down

The hardest part of cybersecurity isn't the technology, it’s the people. Every major breach you’ve read about lately usually starts the same way: one employee, one clever email, and one "Patient Zero" infection. In 2026, hackers are using AI to make these "first clicks" nearly impossible to spot. If a single laptop gets compromised on your watch, do you have a plan to stop it from taking down [email protected]

Flexibilní systém HyTEM nahrazuje na experimentálním letounu pohyblivé klapky • Umělá inteligence přizpůsobuje tvar křídla aktuálním povětrnostním podmínkám • Inovativní přístup zvyšuje celkovou účinnost a snižuje spotřebu paliva

Two U.S. nationals were sentenced to 18 months in prison each for operating so-called laptop farms that helped North Korean IT workers fraudulently obtain remote employment at nearly 70 American companies. [...]

Palo Alto Networks has disclosed that threat actors may have attempted to unsuccessfully exploit a recently disclosed critical security flaw as early as April 9, 2026. The vulnerability in question is CVE-2026-0300 (CVSS score: 9.3/8.7), a buffer overflow vulnerability in the User-ID Authentication Portal service of Palo Alto Networks PAN-OS software that could allow an unauthenticated attacker

Palo Alto Networks has disclosed that threat actors may have attempted to unsuccessfully exploit a recently disclosed critical security flaw as early as April 9, 2026. The vulnerability in question is CVE-2026-0300 (CVSS score: 9.3/8.7), a buffer overflow vulnerability in the User-ID Authentication Portal service of Palo Alto Networks PAN-OS software that could allow an unauthenticated attacker Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

Nejlepší LLM už je dostupný pro všechny uživatele ChatuGPT. • GPT-5.5 Instant slibuje chytřejší, přesnější a stručnější odpovědi. • ChatGPT zároveň zlepšuje paměť a řekne, kdy z ní čerpal.

A 20-year-old California man was sentenced to 78 months in prison for serving as a home invader and money launderer in a criminal ring that stole over $250 million in cryptocurrency. [...]

Ačkoli je papež Lev XIV. hlavou katolické církve a stojí v čele více než miliardy věřících po celém světě, také on někdy řeší všední potíže. A kdo v životě neměl problémy se zákaznickou linkou? Krátce poté, co nastoupil do úřadu, musel papež se svou bankou řešit změnu údajů. Operátorka ale nechtěla uvěřit, s kým mluví, a Svatému otci zavěsila.

Modern attacks don't stop at initial compromise. This webinar explores why security and recovery must work together to reduce downtime and improve resilience. [...]

Apple’s Worldwide Developer Conference (WWDC) takes place in just a few weeks. Everyone expects the company to explain its approach to AI deployment on its platforms. With that in mind, here’s what several months of speculation suggest Apple will announce, though the details remain to be disclosed.

Apple is investing billions of dollars in these plans; R&D spending reached 10.3% of revenue in the second quarter, up from 7.6% in Q1. Given Apple’s accelerating revenue, on a dollar basis this means the company’s R&D spend is up 34% from a year ago.

“We believe AI is a really important investment area for Apple, and we’re going to be doing that incrementally on top of what we normally invest in our product roadmap,” said Apple CFO Kevan Parekh during Apple’s latest fiscal call. (AI isn’t Apple’s only spending target, either.)

While the billions Apple is investing are dwarfed by the huge infrastructure investments made by pure AI players, Apple’s infrastructure already exists — in the form of 2.5 billion actively used devices, the vast majority of which can already run some AI models natively on device. So, how is Apple exploiting this deployment advantage?

BYO-AI

Only this week, Bloomberg once again confirmed Apple intends to permit its customers to select their choice of AI service on their device. You’ll be able to pick Gemini, ChatGPT, or Claude as your default supporting AI service. This means that while you might continue to use Apple Intelligence for most queries, you’ll also be able to use one of those server-based AI choices for more complex tasks. Y

ou will also be able to opt to use those services to provide all your AI needs, but not until iOS 27, which will be unveiled at WWDC. These services will be provided within a new Extension system, which I suspect will work with the apps those AI developers are already building for iPhones. That could lead to a new ‘App Store for AI’ approach, which the company may be able to monetize. So, you’ll be able to select between AI services for tasks such as text generation and editing and powering Siri.

Work with Gemini

While allowing users AI choice, Apple is also building out Apple Intelligence to be a competitive option. The company’s engineers have been working with Google Gemini to build their own Foundation Models for common tasks.

That work includes use of a customized version of Gemini to improve Siri’s conversational abilities and understanding of natural language as well as a degree of contextual intelligence. This should enable Siri to perform complex tasks across more than one app. “Apple Intelligence will continue to run on Apple devices and Private Cloud Compute, while maintaining Apple’s industry-leading privacy standards,” Apple and Google confirmed earlier this year.

Private by design

Apple is building privacy into these tools, meaning those Foundation Models and new Siri features will maintain user privacy by relying on Private Cloud Compute, Apple’s uniquely private infrastructure that lets you use powerful server-based AI to resolve complex tasks without eroding privacy. This is good for everyone, but particularly promising to enterprises seeking AI efficiency along with guaranteed data security. That provision of privacy should let Apple offer its services to businesses that might otherwise opt for sovereign or on-premises AI data services. A legal team might feel more confident to use AI to summarize confidential contracts on-device, for example, while a medical practitioner might be able to use AI to securely analyze patient notes. In many cases, IT will almost certainly gravitate towards Apple’s privacy posture.

The hardware thing

Apple Silicon, Unified Memory, the Neural Engine and so many additional hardware and software achievements available in Apple’s products mean the entire ecosystem is already resolutely AI ready. 

In an important statement, CEO Tim Cook recently explained how this all works together, and the extent to which Apple’s hardware advantage already translates into an experience advantage for AI users and developers:

“What truly sets Apple apart is how Apple Intelligence is woven into the core of our platforms, powered by Apple Silicon, and designed from the ground up to deliver intelligence that is fast, personal, and private. This is not AI as a standalone feature, but AI as an essential intuitive part of the experience across our devices. It builds on years of innovation, from the neural engine, to advanced on-device processing, enabling capabilities that are not only incredibly powerful but also respectful of user privacy.

“Increasingly, that same foundation is drawing developers and researchers to our products as powerful platforms for building and running agentic AI, thanks to the unique combination of performance, efficiency, and on-device capabilities. When you combine this level of integration with our relentless focus on the customer experience, it becomes clear why Apple platforms are the best place to experience AI.”

You don’t need an LLM to see which way the wind blows

In short, Apple’s platforms are already the best place to build AI and the best place to use AI. Apple has decided to lean into that advantage while adopting a multi-state approach to AI services; it recognizes the inevitable commoditization of those services, while also continuing to develop its own to compete with third parties. With Apple’s own ecosystem as the battleground, that approach means Apple wins, even if its AI services lose.

“With Apple Silicon and its powerful unified memory architecture, leading AI developers like Perplexity are choosing Mac as their preferred platform to build enterprise-grade AI assistants that power autonomous agents and boost workplace productivity,” Apple said.

Asymco’s Horace Dediu notes the significance of such a shift: “If foundation models are heading toward commodity status, then the strategic value shifts to whoever controls the integration layer and the user relationship,” he wrote. 

People, get ready

Expect to learn much more concerning Apple’s AI plans at WWDC in a few weeks. While Apple is doubtless still burning at the challenges it has encountered so far, like any marathon runner the company continues to race through those short-term pains. “We can’t wait to share what we’ve been working on, from AI advancements to exciting new software and developer tools. It’s going to be an incredible week,” Cook promised during Apple’s fiscal call.

We’ll all be watching.

You can follow me on social media! Join me on BlueSky,  LinkedIn, and Mastodon.

Garmin si většina lidí spojí hlavně s hodinkami, navigacemi a různou sportovní elektronikou. Jenže firma už v roce 2023 koupila americkou značku JL Audio, která patří mezi známá jména ve světě subwooferů, systémů pro auta a profesionálního ozvučení. A právě pod touto značkou nyní Garmin představil ...

Bad week. Turns out the easiest way to get hacked in 2026 is still the same old garbage: shady packages, fake apps, forgotten DNS junk, scam ads, and stolen logins getting dumped into Discord channels like it’s normal. Some of these attack chains don’t even feel sophisticated anymore. More like some tired guy with a Telegram account and too much free time. The worst part is how often this stuff