Agregátor RSS

Na všech herních platformách je každou chvíli nějaká slevová akce. Každý týden proto vybíráme ty nejatraktivnější, které by vám neměly uniknout. Pokud chcete získat hry zdarma nebo s výhodnou slevou, podívejte se na aktuální přehled akcí!

Austrian privacy non-profit None of Your Business (noyb) has filed complaints accusing companies like TikTok, AliExpress, SHEIN, Temu, WeChat, and Xiaomi of violating data protection regulations in the European Union by unlawfully transferring users' data to China. The advocacy group is seeking an immediate suspension of such transfers, stating the companies in question cannot shield user data Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

Some of you have apparently already botched chatbots or allowed ‘shadow AI’ to creep in

Cisco and Nvidia have both recognized that as useful as today's AI may be, the technology can be equally unsafe and/or unreliable – and have delivered tools in an attempt to help address those weaknesses.…

We'll defo ask for permission next time, automaker tells FTC

General Motors on Thursday said that it has reached a settlement with the FTC "to address privacy concerns about our now-discontinued Smart Driver program."…

Od začátku roku už přijímají eDoklady i banky a pošty. Ačkoli systém běží dobře a chyby jsou spíš výjimečné, existují i situace, kdy je nezbytné nebo aspoň rychlejší ověřit se přes klasickou plastovou občanku.

Krom očekávaných novinek jako Ryzen 9 9950X3D, Ryzen 9 9900X3D a Ryzen 5 9600 došlo i na jednu nečekanou záležitost. AMD bez tiskových zpráv vydává nejpomalejší šestijádrový Zen 4 bez grafiky…

…aneb Číselné ohlédnutí za uplynulým rokem v dinosauří paleontologii

Dnešní klasické qubity mají dva stavy spinu a jsou kvůli tomu velmi zranitelné. Antimonová kočka z USNW Sydney má osm spinových stavů, které jí dávají sedm životů. Krásně popularizovaný výzkum moderní Schrödingerovy kočky otevírá dveře dalšímu vývoji kvantových počítačů a slibuje velezajímavé aplikace.

Fedora hodlá cílit oficiální variantu s desktopem KDE Plasma jako systém pro multimédia a digitální obsah. Získají ve Fedoře současní tvůrci systém, který jim bude optimálně vyhovovat?

Non-profit privacy advocacy group "None of Your Business" (noyb) has filed six complaints against TikTok, AliExpress, SHEIN, Temu, WeChat, and Xiaomi, for unlawfully transferring European user's data to China and infringing European Union's general data protection regulation (GDPR). [...]

It’s accurate and nearly as fast as expert human interpreters.

The dream of a universal AI interpreter just got a bit closer. This week, tech giant Meta released a new AI that can almost instantaneously translate speech in 101 languages as soon as the words tumble out of your mouth.

AI translators are nothing new. But they generally work best with text and struggle to transform spoken words from one language to another. The process is usually multistep. The AI first turns speech into text, translates the text, and then converts it back to speech. Though already useful in everyday life, these systems are inefficient and laggy. Errors can also sneak in at each step.

Meta’s new AI, dubbed SEAMLESSM4T, can directly convert speech into speech. Using a voice synthesizer, the system translates words spoken in 101 languages into 36 others—not just into English, which tends to dominate current AI interpreters. In a head-to-head evaluation, the algorithm is 23 percent more accurate than today’s top models—and nearly as fast as expert human interpreters. It can also translate text into text, text into speech, and vice versa.

Meta is releasing all the data and code used to develop the AI to the public for non-commercial use, so others can optimize and build on it. In a sense, the algorithm is “foundational,” in that “it can be fine-tuned on carefully curated datasets for specific purposes—such as improving translation quality for certain language pairs or for technical jargon,” wrote Tanel Alumäe at Tallinn University of Technology, who was not involved in the project. “This level of openness is a huge advantage for researchers who lack the massive computational resources needed to build these models from scratch.”

It’s “a hugely interesting and important effort,” Sabine Braun at the University of Surrey, who was also not part of the study, told Nature.

Self-Learning AI

Machine translation has made strides in the past few years thanks to large language models. These models, which power popular chatbots like ChatGPT and Claude, learn language by training on massive datasets scraped from the internet—blogs, forum comments, Wikipedia.

In translation, humans carefully vet and label these datasets, or “corpuses,” to ensure accuracy. Labels or categories provide a sort of “ground truth” as the AI learns and makes predictions.

But not all languages are equally represented. Training corpuses are easy to come by for high-resource languages, such as English and French. Meanwhile, low-resource languages, largely used in mid- or low-income countries, are harder to find—making it difficult to train a data-hungry AI translator with trusted datasets.

“Some human-labeled resources for translation are freely available, but often limited to a small set of languages or in very specific domains,” wrote the authors.

To get around the problem, the team used a technique called parallel data mining, which crawls the internet and other resources for audio snippets in one language with matching subtitles in another. These pairs, which match in meaning, add a wealth of training data in multiple languages—no human annotation needed. Overall, the team collected roughly 443,000 hours of audio with matching text, resulting in about 30,000 aligned speech-text pairs.

SEAMLESSM4T consists of three different blocks, some handling text and speech input and others output. The translation part of the AI was pre-trained on a massive dataset containing 4.5 million hours of spoken audio in multiple languages. This initial step helped the AI “learn patterns in the data, making it easier to fine-tune the model for specific tasks” later on, wrote Alumäe. In other words, the AI learned to recognize general structures in speech regardless of language, establishing a baseline that made it easier to translate low-resource languages later.

The AI was then trained on the speech pairs and evaluated against other translation models.

Spoken Word

A key advantage of the AI is its ability to directly translate speech, without having to convert it into text first. To test this ability, the team hooked up an audio synthesizer to the AI to broadcast its output. Starting with any of the 101 languages it knew, the AI translated speech into 36 different tongues—including low-resource languages—with only a few seconds of delay.

The algorithm outperformed existing state-of-the-art systems, achieving 23 percent greater accuracy using a standardized test. It also better handled background noise and voices from different speakers, although—like humans—it struggled with heavily accented speech.

Lost in Translation

Language isn’t just words strung into sentences. It reflects cultural contexts and nuances. For example, translating a gender-neutral language into a gendered one could introduce biases. Does “I am a teacher” in English translate to the masculine “Soy profesor” or to the feminine “Soy profesora” in Spanish? What about translations for doctor, scientist, nanny, or president?

Mistranslations may also add “toxicity,” when the AI spews out offensive or harmful language that doesn’t reflect the original meaning—especially for words that don’t have a direct counterpart in the other language. While easy to laugh off as a comedy of errors in some cases, these mistakes are deadly serious when it comes to medical, immigration, or legal scenarios.

“These sorts of machine-induced error could potentially induce real harm, such as erroneously prescribing a drug, or accusing the wrong person in a trial,” wrote Allison Koenecke at Cornell University, who wasn’t involved in the study. The problem is likely to disproportionally affect people speaking low-resource languages or unusual dialects, due to a relative lack of training data.

To their credit, the Meta team analyzed their model for toxicity and fine-tuned it during multiple stages to lower the chances of gender bias and harmful language.

“This is a step in the right direction, and offers a baseline against which future models can be tested,” wrote Koenecke.

Meta is increasingly supporting open-source technology. Previously, the tech giant released PyTorch, a software library for AI training, which was used by companies, including OpenAI and Tesla, and researchers around the globe. SEAMLESSM4T will also be made public for others to build on its abilities.

The AI is just the latest machine translator that can handle speech-to-speech translation. Previously, Google showcased AudioPaLM, an algorithm that can turn 113 languages into English—but only English. SEAMLESSM4T broadens the scope. Although it only scratches the surface of the roughly 7,000 languages spoken, the AI inches closer to a universal translator—like the Babel fish in The Hitchhiker’s Guide to the Galaxy, which translates languages from species across the universe when popped into the ear.

“The authors’ methods for harnessing real-world data will forge a promising path towards speech technology that rivals the stuff of science fiction,” wrote Alumäe.

The post Meta’s New AI Translates Speech in Real Time Across More Than 100 Languages appeared first on SingularityHub.

Certifikační autorita Let's Encrypt oznámila, že bude volitelně nabízet krátkodobé certifikáty s šestidenní platností a navíc s možností vystavit je na IP adresu. Zvolit typ certifikátu bude možné v certifikačním profilu ACME.

A severe flaw in the W3 Total Cache plugin installed on more than one million WordPress sites could give attackers access to various information, including metadata on cloud-based apps. [...]

Herní konzole Nintendo Switch 2 byla oficiálně potvrzena. Vyjde letos. Trailer na YouTube. Více ve středu 2. dubna na Nintendo Direct.

Microsoft has expanded its Windows 11 administrator protection tests, allowing Insiders to enable the security feature from the Windows Security settings. [...]

OpenAI has started rolling out a number of new features to ChatGPT, according to Techcrunch.

The big news is support for reminders; the feature allows users, for example, to ask ChatGPT to remind them when it’s time to renew an ID or passport. Users can also ask ChatGPT to deliver a news summary or weather forecast at the same time every day — something OpenAI called Tasks.

Initially, the new abilities will only be available to paying customers of Chat GPT Plus, Team, and Pro.

FSB cyberspies venture into a new app for espionage, Microsoft says

updated  Star Blizzard, a prolific phishing crew backed by the Russian Federal Security Service (FSB), conducted a new campaign aiming to compromise WhatsApp accounts and gain access to their messages and data, according to Microsoft.…

Posted by Erik Varga, Vulnerability Management, and Rex Pan, Open Source Security Team

In December 2022, we announced OSV-Scanner, a tool to enable developers to easily scan for vulnerabilities in their open source dependencies. Together with the open source community, we’ve continued to build this tool, adding remediation features, as well as expanding ecosystem support to 11 programming languages and 20 package manager formats. 

Today, we’re excited to release OSV-SCALIBR (Software Composition Analysis LIBRary), an extensible library for SCA and file system scanning. OSV-SCALIBR combines Google’s internal vulnerability management expertise into one scanning library with significant new capabilities such as:

• SCA for installed packages, standalone binaries, as well as source code

• OSes package scanning on Linux (COS, Debian, Ubuntu, RHEL, and much more), Windows, and Mac

• Artifact and lockfile scanning in major language ecosystems (Go, Java, Javascript, Python, Ruby, and much more)

• Vulnerability scanning tools such as weak credential detectors for Linux, Windows, and Mac

• SBOM generation in SPDX and CycloneDX, the two most popular document formats

• Optimization for on-host scanning of resource constrained environments where performance and low resource consumption is critical

OSV-SCALIBR is now the primary SCA engine used within Google for live hosts, code repos, and containers. It’s been used and tested extensively across many different products and internal tools to help generate SBOMs, find vulnerabilities, and help protect our users’ data at Google scale.

We offer OSV-SCALIBR primarily as an open source Go library today, and we're working on adding its new capabilities into OSV-Scanner as the primary CLI interface.

Using OSV-SCALIBR as a library

All of OSV-SCALIBR's capabilities are modularized into plugins for software extraction and vulnerability detection which are very simple to expand.You can use OSV-SCALIBR as a library to:

1.Generate SBOMs from the build artifacts and code repos on your live host:

import (

 "context"

 "github.com/google/osv-scalibr"

 "github.com/google/osv-scalibr/converter"

 "github.com/google/osv-scalibr/extractor/filesystem/list"

 "github.com/google/osv-scalibr/fs"

 "github.com/google/osv-scalibr/plugin"

 spdx "github.com/spdx/tools-golang/spdx/v2/v2_3"

)

func GenSBOM(ctx context.Context) *spdx.Document {

 capab := &plugin.Capabilities{OS: plugin.OSLinux}

 cfg := &scalibr.ScanConfig{

   ScanRoots: fs.RealFSScanRoots("/"),

   FilesystemExtractors: list.FromCapabilities(capab),

   Capabilities: capab,

 }

 result := scalibr.New().Scan(ctx, cfg)

 return converter.ToSPDX23(result, converter.SPDXConfig{})

}

2. Scan a git repo for SBOMs:

Simply replace "/" with the path to your git repo. Also take a look at the various language extractors to enable for code scanning.

3. Scan a remote container for SBOMs:

Replace the scan config from the above code snippet with

import (

 ...

 "github.com/google/go-containerregistry/pkg/authn"

 "github.com/google/go-containerregistry/pkg/v1/remote"

 "github.com/google/osv-scalibr/artifact/image"

 ...

)

...

filesys, _ := image.NewFromRemoteName(

 "alpine:latest",

 remote.WithAuthFromKeychain(authn.DefaultKeychain),

)

cfg := &scalibr.ScanConfig{

 ScanRoots: []*fs.ScanRoot{{FS: filesys}},

 ...

}

4. Find vulnerabilities on your filesystem or a remote container:

Extract the PURLs from the SCALIBR inventory results from the previous steps:

import (

 ...

 "github.com/google/osv-scalibr/converter"

 ...

)

...

result := scalibr.New().Scan(ctx, cfg)

for _, i := range result.Inventories {

 fmt.Println(converter.ToPURL(i))

}

And send them to osv.dev, e.g.

$ curl -d '{"package": {"purl": "pkg:npm/[email protected]"}}' "https://api.osv.dev/v1/query"

See the usage docs for more details.

OSV-Scanner + OSV-SCALIBR

Users looking for an out-of-the-box vulnerability scanning CLI tool should check out OSV-Scanner, which already provides comprehensive language package scanning capabilities using much of the same extraction as OSV-SCALIBR. 

Some of OSV-SCALIBR’s capabilities are not yet available in OSV-Scanner, but we’re currently working on integrating OSV-SCALIBR more deeply into OSV-Scanner. This will make more and more of OSV-SCALIBR’s capabilities available in OSV-Scanner in the next few months, including installed package extraction, weak credentials scanning, SBOM generation, and more.

Look out soon for an announcement of OSV-Scanner V2 with many of these new features available. OSV-Scanner will become the primary frontend to the OSV-SCALIBR library for users who require a CLI interface. Existing users of OSV-Scanner can continue to use the tool the same way, with backwards compatibility maintained for all existing use cases. 

For installation and usage instructions, have a look at OSV-Scanner’s documentation here.

What’s next

In addition to making all of OSV-SCALIBR’s features available in OSV-Scanner, we're also working on additional new capabilities. Here's some of the things you can expect:

• Support for more OS and language ecosystems, both for regular extraction and for Guided Remediation

• Layer attribution and base image identification for container scanning

• Reachability analysis to reduce false positive vulnerability matches

• More vulnerability and misconfiguration detectors for Windows

• More weak credentials detectors

We hope that this library helps developers and organizations to secure their software and encourages the open source community to contribute back by sharing new plugins on top of OSV-SCALIBR.

If you have any questions or if you would like to contribute, don't hesitate to reach out to us at [email protected] or by posting an issue in our issue tracker.

When considering new hires, 80% of corporate executives will prioritize skills over degrees, with half planning to increase freelance hiring this year to fill in for a gap in AI and other skills, according to a new study from freelancing platform Upwork.

The study, released this week, showed “unprecedented growth” in specialized AI skills, which have surged 220% year-over-year.

At the same time, degrees continue to lose relevance when it comes to hiring freelancers, with 74% of execs focused more on proven expertise. Moreover, 78% of CEOs say top freelancers deliver more value than degree-holding employees, emphasizing skills over credentials to stay competitive. And 29% of C-suite executives consider freelancers essential to their operations, with 51% saying their business would be difficult to run without freelancer support.

Skills-based hiring has been on the rise for several years, as organizations seek to fill specific tech needs such as big data analytics, programing (such as Rust) and AI prompt engineering. In fact, demand for genAI courses is surging, passing all other tech skills courses and spanning fields from data science to cybersecurity, project management, and marketing.

The top 10 highest paid skills in tech can help workers earn up to 47% more — and the top skill among them is generative artificial intelligence (genAI), according to employment website Indeed and other sources.

Skills such as genAI modeling now earn freelancers up to 22% higher hourly rates than traditional AI and machine learning roles, according to Upwork.

Even as freelancers are reshaping workforce strategies, their rise doesn’t necessarily threaten full-time roles. “It complements them,” said Kelly Monahan, managing director of the Upwork Research Institute.

In a study released in October, Upwork found that 85% of top-performing companies — which it labels “work innovators” — view freelancers as vital, with 91% planning to expand their use over the next year. Only 71% of non top-performing companies see freelancers as critical to success, Monahan said.

While cost savings, such as not paying benefits, could sometimes be a factor in hiring freelancers, it is not the primary driver of freelance hiring, according to Monahan.“Businesses prioritize freelancers for their agility and specialized expertise, which enable them to scale resources up or down as needed and address skill gaps effectively,” she said.

According to Upwork, other reasons for the increase in freelance hiring include:

• 94% of top-performing companies say hiring freelancers gives them access to specialized skills
• 89% say freelancers make their business more innovative
• 84% say hiring a freelancer is faster than a hiring full-time employee

In addition to hard skills, soft, human-centric roles such as personal coaching have emerged among the fastest-growing skills on Upwork’s platform, with demand increasing by 74% year-over-year. “This underscores the growing importance of guidance and adaptability as businesses invest in reskilling their workforces to navigate technological change,” Monahan said. “Freelancers are enabling companies to innovate rapidly and adapt to changing market demands.”

Upwork is not alone in its findings. According to research firm Gartner, organizations are struggling to find skilled talent, and universities — once vital for workforce preparation — are lagging in updating curricula to match modern demands. As technology and work methods advance, graduates are left with outdated skills, making specific competencies more important than degrees in proving a candidate’s value.

According to Gartner, 74% of HR leaders believe organizations are shifting to skills-based talent management, but only 41% have implemented it, while 50% are still considering it.

“Approximately half of HR leaders say that a skills-based approach to talent management has the potential to solve many of the challenges their organizations face, though only one-third are actually investing in a skills-based approach to talent management, Gartner said in its report.

HR leaders, Gartner said, should prepare for a skills-focused future by:

• Assessing: Review role requirements to reduce or remove degree mandates.
• Fortifying: Ready the organization to onboard and support non-degreed talent.
• Attracting: Target skilled non-degreed talent and adjust EVP messaging to appeal to them.
• Evolving: Plan for talent management changes to adopt a skills-based approach.

New methods of assessing skills

Companies are adopting more advanced approaches to assessing potential and current employee skills, blending AI tools with hands-on evaluations, according to Monahan. 

AI-powered platforms are being used to match candidates with roles based on their skills, certifications, and experience. “Our platform has done this for years, and our new UMA (Upwork’s Mindful AI) enhances this process,” she said.

Gartner, however, warned that “rapid skills evolutions can threaten quality of hire, as recruiters struggle to ensure their assessment processes are keeping pace with changing skills. Meanwhile, skills shortages place more weight on new hires being the right hires, as finding replacement talent becomes increasingly challenging. Robust appraisal of candidate skills is therefore imperative, but too many assessments can lead to candidate fatigue.”

In Upwork’s In-Demand Skills 2025 report, the skills that are growing in importance include:

• AI Development: GenAI modeling and AI data annotation are among the fastest-growing skills, reflecting the need for technical expertise in building and managing AI solutions.
• >Data Science & Analytics: >Skills such as data visualization and data extraction remain essential for making sense of complex information.
• >Project Management: >Both in supply chain logistics and business operations, project managers are critical for keeping teams aligned and projects on track.
• >Professional Development: >Skills such as personal coaching and training and development are increasingly sought as companies prioritize workforce reskilling.

The shift toward skills-based hiring is further driven by a readiness gap in today’s workforce. Upwork’s research found that only 25% of employees feel prepared to work effectively alongside AI, and even fewer (19%) can proactively leverage AI to solve problems.

“As companies navigate these challenges, they’re focusing on hiring based on practical, demonstrated capabilities, ensuring their workforce is agile and equipped to meet the demands of a rapidly evolving business landscape,” Monahan said.

According to Upwork, 47% of Gen Z professionals already engage in freelance or portfolio work, reflecting their preference for autonomy and skills diversification over traditional career paths.

“This aligns with modern businesses’ needs for agile talent who can deliver measurable results, driving the shift toward skills-based hiring,” Monahan said. “If you are just looking to fill job roles, you will miss out on the rising portfolio career talent.”

The U.S. Treasury Department has sanctioned a network of individuals and front companies linked to North Korea's Ministry of National Defense that have generated revenue via illegal remote IT work schemes. [...]