Agregátor RSS

10 věcí, které nás štvou na Windows 11 a bohužel asi jen tak nepřestanou

Živě.cz - 10 Březen, 2025 - 19:45
**Přibývají agresivní reklamy a nerespektující chování **Widgety jsou na špatném místě a nabídka Start neflexibilní **Operační systém na uživatele příliš nereaguje, měl by být chytřejší
Kategorie: IT News

Internetová bublina praskla před 25 roky. Hloupý byl, kdo dával, hloupější, kdo nebral

Živě.cz - 10 Březen, 2025 - 19:15
Před 25 roky praskla bublina na americkém akciovém trhu, kterou způsobily technologické firmy orientované na internet . V angličtině se fenomén označuje za dot-com bubble, v češtině hovoříme o internetové horečce. V 90. letech se internet živelně rozvíjel a komercionalizoval. Investoři s vidinou ...
Kategorie: IT News

FTC will send $25.5 million to victims of tech support scams

Bleeping Computer - 10 Březen, 2025 - 18:58
​Later this week, the Federal Trade Commission (FTC) will start distributing over $25.5 million in refunds to those misled by tech support companies Restoro and Reimage's scare tactics. [...]
Kategorie: Hacking & Security

Eutelsat slibuje, že na Ukrajině dokáže nahradit Starlink. Jeho hlavní výhodou je hybridní družicový systém

Živě.cz - 10 Březen, 2025 - 18:45
Svět se v souvislosti se situací na Ukrajině naučil nové slovíčko Eutelsat . Francouzsko-britský satelitní operátor se totiž nechal slyšet, že by mohl zkoušené zemi v následujících měsících dodat až čtyřicet tisíc internetových terminálů, které by zajistily nezávislost na dosavadním Starlinku. ...
Kategorie: IT News

Microsoft is developing its own AI models to compete with OpenAI

Computerworld.com [Hacking News] - 10 Březen, 2025 - 18:26

One of the hottest trends right now involves work on reasoning AI models — tools that can more or less arrive at the right conclusion on their own. (The best-known reasoning models at the moment are the Chinese Deepseek-R1, as well as o1 and o3-mini from OpenAI.

However, credible reports suggest Microsoft has decided to seriously challenge Deepseek and OpenAI by developing its own set of reasoning AI models called Microsoft AI (MAI).

If successful, Microsoft would eventually not have to use its partner OpenAI’s o1 models in Copilot, Bloomberg reports.

Exactly when MAI might be released to the public is unclear at the moment, though experts expect a launch later this year.

Kategorie: Hacking & Security

Is Apple Intelligence the new Apple Maps?

Computerworld.com [Hacking News] - 10 Březen, 2025 - 18:17

Following widespread speculation about ongoing problems, Apple has announced that one of the key features of Apple Intelligence will be delayed — perhaps until 2026. The delay could harm Apple’s brand during a critical time for the tech industry. It remains unclear whether the company will be able to successfully manage its way through these choppy waters.

What Apple said

Apple initially promised it would introduce contextual intelligence in Apple intelligence about now. But Apple spokesperson Jacqueline Roy recently told Daring Fireball:

“Siri helps our users find what they need and get things done quickly, and in just the past six months, we’ve made Siri more conversational, introduced new features like type to Siri and product knowledge, and added an integration with ChatGPT. We’ve also been working on a more personalized Siri, giving it more awareness of your personal context, as well as the ability to act for you within and across your apps. It’s going to take us longer than we thought to deliver on these features, and we anticipate rolling them out in the coming year.”

The coming year could mean quite a lot here, including pushing the rollout back into 2026.

When it didn’t make the grade

Apple Intelligence currently combines two systems: an older one based on machine intelligence and pattern matching and a new system based on a large language model (LLM). The new system can handle changes in your questions in Siri, for example, but the existence of two forms of AI on the device makes successful deployment more complex.

Apple had originally planned to introduce these new tools with iOS 18.4, but this was delayed into May. It now seems those efforts have fallen short, despite efforts from Apple’s top developers who were airlifted in to bail things out.

Failure was sealed when Craig Federighi, Apple senior vice president of software engineering, slammed the brakes on public introduction of the tools because the features just weren’t up to speed. Results are inconsistent and fail to meet the promises Apple has made. This is reminiscent of previous failures.

From A to B and back again

Way back when Apple introduced Apple Maps, it did so with a blaze of glory. But the system quickly failed to live up to the hype. A variety of issues plagued the nascent Apple service, critics went crazy and top software leader Scott Forstall lost his job. Apple Maps quickly became a joke, with Apple fans and critics alike choosing alternative services. They didn’t see Maps as a serious option, a view that took years of development and millions of dollars to break. Apple today arguably offers the best mapping service in the world.

The sin Apple committed with Apple Maps was to promise more than it could actually deliver. Once the service hit the market, its weaknesses were quickly identified, leading to the reputational hit that took years to recover from. 

Apple’s existing AI assistant, Siri, has a similar reputation — this time, by starting as something fantastic and quickly becoming something of a joke when compared to competing services. Apple Intelligence was meant to change that, but the latest delay threatens to make it worse, cementing in public opinion the idea that Apple doesn’t know how to do AI at the very worst moment in industry evolution.

An unforced error?

We know that Apple recognizes the importance of AI. That’s why the company has been promoting Apple Intelligence so urgently in nearly all its hardware advertising. (That includes the ads the company pulled over the weekend promoting the now delayed contextual intelligence in Siri.) Hoping that the future would seem better tomorrow, Apple bet its marketing push on services it hadn’t yet made.

Apple’s hype machine sang out the coming benefits of Apple Intelligence loudly. While the choir sang, teams inside the company made progress on preparing the company’s ecosystem for the new AI features. That included introducing entry-level products equipped with enough memory to run the promised new AI features. Outside Apple, the company was energetically promoting the new features among customers and developers and maintaining belief in the company trajectory among its customers and advocates. 

You could argue that every root and branch of the company was focused on setting the scene for this new super smart Siri, the introduction of which was intended to be the point in time at which the company absolutely nailed AI.

After months (and years) of steady planning and execution, this was to be the flourish, the time when others sharing the tech chess board moved their kings aside.

That’s was the plan, at least.

Grabbing the wrong puck

While you can argue it’s good Apple didn’t introduce a flawed product, it will be judged as if it did so. Because the company chose to focus so much energy promoting Apple Intelligence on the basis of a set of features it has not been able to deliver in time.

In promoting this vaporware, Apple created expectations it couldn’t meet. That’s always bad business, but on a product as strategically important as AI it could pose an existential challenge. Overcoming that challenge is going to take all the time-tested approaches we’ve seen Apple use before when it gets things wrong: High-level changes, public apology, interviews, and possibly a resignation or two before the company gets it right.

Meanwhile, as a result of the blunder, Apple will need to squander its energy battling the idea that Apple Intelligence isn’t smart enough. That’s a perception directly fed by all the AI-focused marketing over the last year or so. (There’s a reason Steve Jobs said the real artists are the people who ship).

This is a big problem for Apple and it knows it.

I fear heads will roll, and in the context of the many other challenges the company faces, this weakest link erodes the capacity of the team. And Siri still doesn’t have a clue.

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

Kategorie: Hacking & Security

Sidewinder goes nuclear, charts course for maritime mayhem in tactics shift

The Register - Anti-Virus - 10 Březen, 2025 - 18:01
Phishing and ancient vulns still do the trick for one of the most prolific groups around

Researchers say the Sidewinder offensive cyber crew is starting to target maritime and nuclear organizations.…

Kategorie: Viry a Červi

Swiss critical sector faces new 24-hour cyberattack reporting rule

Bleeping Computer - 10 Březen, 2025 - 17:47
Switzerland's National Cybersecurity Centre (NCSC) has announced a new reporting obligation for critical infrastructure organizations in the country, requiring them to report cyberattacks to the agency within 24 hours of their discovery. [...]
Kategorie: Hacking & Security

Google paid $12 million in bug bounties last year to security researchers

Bleeping Computer - 10 Březen, 2025 - 17:36
Google paid almost $12 million in bug bounty rewards to 660 security researchers who reported security bugs through the company's Vulnerability Reward Program (VRP) in 2024. [...]
Kategorie: Hacking & Security

Researchers Expose New Polymorphic Attack That Clones Browser Extensions to Steal Credentials

The Hacker News - 10 Březen, 2025 - 16:47
Cybersecurity researchers have demonstrated a novel technique that allows a malicious web browser extension to impersonate any installed add-on. "The polymorphic extensions create a pixel perfect replica of the target's icon, HTML popup, workflows and even temporarily disables the legitimate extension, making it extremely convincing for victims to believe that they are providing credentials to
Kategorie: Hacking & Security

Researchers Expose New Polymorphic Attack That Clones Browser Extensions to Steal Credentials

The Hacker News - 10 Březen, 2025 - 16:47
Cybersecurity researchers have demonstrated a novel technique that allows a malicious web browser extension to impersonate any installed add-on. "The polymorphic extensions create a pixel perfect replica of the target's icon, HTML popup, workflows and even temporarily disables the legitimate extension, making it extremely convincing for victims to believe that they are providing credentials to Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]
Kategorie: Hacking & Security

Quantum leap: Passwords in the new era of computing security

Bleeping Computer - 10 Březen, 2025 - 16:02
Quantum computing threatens to break traditional encryption, putting sensitive data at risk. Learn more from Specops Software about the risks of quantum computing and how to prepare for them. [...]
Kategorie: Hacking & Security

Rhysida pwns two US healthcare orgs, extracts over 300K patients' data

The Register - Anti-Virus - 10 Březen, 2025 - 16:00
Terabytes of sensitive info remain available for download

Break-ins to systems hosting the data of two US healthcare organizations led to thieves making off with the personal and medical data of more than 300,000 patients.…

Kategorie: Viry a Červi

Consumer Reports calls out slapdash AI voice-cloning safeguards

The Register - Anti-Virus - 10 Březen, 2025 - 15:15
Study finds 4 out of 6 providers don't do enough to stop impersonation

Four out of six companies offering AI voice cloning software fail to provide meaningful safeguards against the misuse of their products, according to research conducted by Consumer Reports.…

Kategorie: Viry a Červi

Microsoft lifts Windows 11 update block for some AutoCAD users

Bleeping Computer - 10 Březen, 2025 - 15:02
Microsoft has removed a compatibility hold that prevented some AutoCAD users from installing the Windows 11 2024 Update due to launch and crash issues. [...]
Kategorie: Hacking & Security

Desert Dexter Targets 900 Victims Using Facebook Ads and Telegram Malware Links

The Hacker News - 10 Březen, 2025 - 14:50
The Middle East and North Africa have become the target of a new campaign that delivers a modified version of a known malware called AsyncRAT since September 2024. "The campaign, which leverages social media to distribute malware, is tied to the region's current geopolitical climate," Positive Technologies researchers Klimentiy Galkin and Stanislav Pyzhov said in an analysis published last week.
Kategorie: Hacking & Security
Syndikovat obsah