Agregátor RSS

Hackers are exploiting a maximum-severity vulnerability, tracked as CVE-2025-59528, in the open-source platform Flowise for building custom LLM apps and agentic systems to execute arbitrary code. [...]

200 orgs and 5,000 devices compromised so far in Vlad's latest intelligence grab, Microsoft reckons

The UK's National Cyber Security Centre (NCSC) has issued a fresh warning about Russia's ongoing targeting of routers to steal passwords and other secrets.…

The Russia-linked threat actor known as APT28 (aka Forest Blizzard) has been linked to a new campaign that has compromised insecure MikroTik and TP-Link routers and modified their settings to turn them into malicious infrastructure under their control as part of a cyber espionage campaign since at least May 2025. The large-scale exploitation campaign has been codenamed Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

Televizory zlevňují a velké úhlopříčky už dávno nejsou luxus. Přesto má smysl připlatit si za vyšší střední a prémiovou třídu. Vybrali jsme 13 modelů, které dnes dávají smysl – od absolutní špičky až po nejlepší poměr cena/výkon.

In the rapid evolution of the 2026 threat landscape, a frustrating paradox has emerged for CISOs and security leaders: Identity programs are maturing, yet the risk is actually increasing. According to new research from the Ponemon Institute, hundreds of applications within the typical enterprise remain disconnected from centralized identity systems. These "[email protected]

The Common Unix Printing System (CUPS) still sits on millions of Linux systems, usually in the background, rarely monitored, and often trusted more than it should be. We saw a wake-up call in late 2024 when a series of vulnerabilities revealed how printer auto-discovery could be abused to enable remote code execution.

OpenAI has released a new policy paper outlining several proposals to address the economic consequences of rapid AI development. The document comes amid growing concerns that AI could quickly take over job roles and fundamentally transform entire industries.

Among the proposals is a public wealth fund, in which the government and AI companies would invest in assets linked to the AI boom, according to Business Insider. The returns would then be distributed directly to citizens.

OpenAI also called for modernizing the tax system, with a greater focus on corporate profits and capital rather than earned income. In the same vein, ideas about special taxes on automated work — sometimes called a “robot tax” — were also raised.

The company also wants governments to encourage companies to test a four-day workweek without pay cuts, where productivity gains from AI are used to the benefit of employees.

An international operation from law enforcement authorities in partnership with private companies has disrupted FrostArmada, an APT28 campaign hijacking local traffic from MikroTik and TP-Link routers to steal Microsoft account credentials. [...]

Když se NASA v úterý odpoledne pochlubila prvními snímky obletu Měsíce z palubní zrcadlovky, sociální sítě se zaplnily komentáři typu: „A fakt to není ilustrace z AI?“ Některé záběry totiž vypadají opravdu až surrealisticky a Měsíc místy spíše jako CGI. NASA je ale federální agentura, která musí ...

A high-severity security vulnerability has been disclosed in Docker Engine that could permit an attacker to bypass authorization plugins (AuthZ) under specific circumstances. The vulnerability, tracked as CVE-2026-34040 (CVSS score: 8.8), stems from an incomplete fix for CVE-2024-41110, a maximum-severity vulnerability in the same component that came to light in July 2024. "Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

Google na webu vids.new zpřístupnil svůj generátor videí i bezplatným účtům. Kdokoliv si může každý měsíc vytvořit až deset osmisekundových klipů v rozlišení 720p a se stereo zvukem. Novinka funguje na nejpokročilejším modelu Veo 3.1 a video vytvoří čistě podle textového zadání, případně může ...

Automated pentesting tools deliver strong early results, then quickly plateau. Picus Security explains how the "PoC cliff" leaves major attack surfaces untested and creates a dangerous validation gap. [...]

An AI system unearthed a trove of CRISPR-like proteins in minutes instead of weeks or months.

CRISPR is a breakthrough technology with humble origins. Scientists first discovered the powerful gene editor in bacteria that were using it as a weapon against invading viruses called phages. Phages can wipe out up to a quarter of a bacterial population in a day. Under assault, bacteria have evolved a hefty arsenal of defenses in a relentless arms race.

These bacterial immune systems often chop up the DNA or RNA of invading viruses and are relatively easy to manufacture, making them alluring targets for scientists developing genetic engineering tools. CRISPR is just one example. There are many more. But traditional methods of searching for them are slow and labor-intensive, leaving most CRISPR-like proteins unexplored.

Now, MIT scientists have released an AI called DefensePredictor that can root out new bacterial defense systems in five minutes, instead of weeks or months. As proof of concept, DefensePredictor churned through hundreds of thousands of proteins in multiple strains of Escherichia coli (E. coli). Over 600 proteins not previously linked to immune defense popped up. Added to a vulnerable strain of bacteria, a subset of these protected them against attack.

“E. coli harbors a much broader landscape of antiphage defense than previously realized, expanding the likely number of systems by multiple orders of magnitude,” wrote the team.

These systems might hold secrets about how immunity evolved. And because the proteins may work in different ways, they could be a goldmine for next-generation precision molecular tools.

Unrivaled Success

Around three decades ago, Japanese scientists discovered a curious, repetitive DNA sequence in E. coli. Other researchers soon realized it was widespread across bacterial species and matched viral DNA sequences—suggesting it could be part of the bacteria’s immunity against phages.

The system now known as CRISPR stores snippets of DNA from past infections and uses protein “scissors” to cut apart matching viral DNA during reinfection. Intrigued by its precision, scientists repurposed CRISPR into a variety of gene editing tools and launched a gene therapy revolution.

CRISPR is the most famous, but a range of bacterial defense systems have transformed genetic engineering. One, containing an enzyme that cuts specific sequences of foreign DNA, is widely used to add genetic material into cells. Another encodes a balance of toxins and antitoxins that can trigger bacterial death after phage infection. This one has been adapted into a kill switch to prevent engineered microbes or genetically modified crops from spreading uncontrollably.

Researchers are also exploring the use of newly discovered systems—with video game-like names like Zorya and Thoeris—as molecular sensors and programmable signaling in synthetic biology.

There are likely more undiscovered tools in the universe of bacterial defense, and scientists have ways of hunting them down. Some defense genes are grouped close to one another, so a known gene could guide the discovery of others. Researchers have also found genes by screening libraries of free-floating circular genome fragments across bacterial populations.

Over 250 systems have been painstakingly validated. But plenty more could escape current detection methods if, for example, their components are spread across the genome.

“The full repertoire of antiphage defense systems in bacteria remains unknown,” wrote the team. “We currently lack the tools to systematically identify systems with high speed, sensitivity, and specificity.”

AI Discoverer

The new DefensePredictor algorithm bridges that gap.

At its core is a protein language model called ESM-2. Proteins are made of 20 molecular “letters” that combine into strings and fold into complex 3D shapes. Similar to large language models, algorithms like ESM-2 learn the language of proteins and can predict their structure and purpose based on sequence alone.

ESM-2 and other similar algorithms have already helped scientists decipher mysterious proteins in bacteria, viruses, and other microorganisms previously unknown to science. Researchers hope their unique shapes could inspire antibiotics, biofuels, or even be used to build synthetic organisms.

To build their AI, the team first established a training ground. With a previous model, DefenseFinder, they screened roughly 17,000 microbial genomes for genes related—and unrelated—to defense systems. They translated these genes into corresponding proteins and built up a database with some 15,000 antiphage proteins and 186,000 proteins unrelated to defense.

These numbers are far too staggering for a human to tackle, but the AI took the work in stride. Alongside ESM-2, the model used several algorithms to distinguish between defense and non-defense proteins. Eventually DefensePredictor learned some general characteristics that make a protein more likely to be part of the immune system. (Like other language models, it’s hard to fully understand the system’s reasoning, which the team is still trying to unpack.)

When tested on 69 strains of E. coli, DefensePredictor surfaced a treasure trove of over 600 new defense-related proteins, including more than 100 that were different than any yet discovered. Although some were encoded near one another or in circular DNA—like previous findings—nearly half weren’t. They were instead littered across the genome yet may still work together.

To test the results, the team engineered a highly vulnerable E. coli strain to express candidate defense proteins—predicted to work either alone or as part of a system—and exposed them to two dozen aggressive phages. Nearly 45 percent of the proteins offered protection against at least one phage.

Beyond E. coli, the scientists expanded their search to 1,000 more microorganisms and found thousands of potential defense proteins unlike anything seen before. “New immune mechanisms remain to be found,” wrote the team.

The race is on. Also published this week, a Pasteur Institute team combined multiple AI models to look for antiphage systems in protein sequences. Across over 32,000 bacterial genomes, the model predicted nearly 2.4 million antiphage proteins—most previously unknown. They released an atlas of AI-predicted bacterial immunity proteins for others to explore.

“The diversity of antiphage defense systems is vast and largely untapped,” they wrote.

Microorganisms harbor a colossal repertoire of biological tools we’re only just beginning to uncover at scale. More species are constantly found thriving in diverse environments, from pond scum to boiling sulfuric springs to the crushing pressure of the Mariana Trench. Every new genome scientists discover and pick apart, now with AI’s help, could be hiding the next CRISPR.    

The post MIT Mined Bacteria for the Next CRISPR—and Found Hundreds of Potential New Tools appeared first on SingularityHub.

Odborníci z AŽD otestují na Jičínsku upravenou bateriovou vlakotramvaj • Vyřazená brněnská šalina proto projde rozsáhlou přestavbou pro železnici • Tento inovativní projekt nabídne levnější alternativu ke klasickým vlakům

An active campaign has been observed targeting internet-exposed instances running ComfyUI, a popular stable diffusion platform, to enlist them into a cryptocurrency mining and proxy botnet. "A purpose-built Python scanner continuously sweeps major cloud IP ranges for vulnerable targets, automatically installing malicious nodes via ComfyUI-Manager if no exploitable node is already Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

Robotické vysavače dnes působí skoro chytře. Umějí si zmapovat byt, vyhnout se překážkám a vrátit se přesně do stanice. Přesto je dokážou rozhodit věci, které člověku připadají úplně banální, jako je tmavý koberec nebo velké zrcadlo. Není to tím, že by byly hloupé. Jen svět kolem sebe nevnímají ...

Nvidia’s recent acquisition of SchedMD, the company behind the Slurm workload manager, is raising concerns among AI industry executives and supercomputing specialists who fear the chip giant could use its new position to favour its own hardware over competing chips, whether through code prioritization or roadmap decisions.

The concern, as industry sources frame it, is straightforward: Nvidia now controls scheduling software that also runs on hardware from its rivals, including AMD and Intel. A vendor that controls workload scheduling software has significant leverage over how efficiently competing hardware performs within shared computing environments — whether it exercises that leverage or not, Reuters reported, citing five anonymous sources, three of whom work in the AI industry and two with knowledge of supercomputer operations.

Analysts who spoke to InfoWorld said Nvidia’s open-source commitment — the company said during the acquisition announcement that it would “continue to develop and distribute Slurm as open-source, vendor-neutral software” — may not be sufficient protection.

“Slurm’s open-source foundation offers safeguards such as transparent code, forking ability, and community governance, but SchedMD’s control gives Nvidia soft power rather than hard lock-in,” said Manish Rawat, semiconductor analyst at TechInsights. Rawat said Nvidia could subtly shape the roadmap, prioritising GPU-aware scheduling and topology optimisations that favour its own hardware, and that integration timelines already showed faster support for the CUDA ecosystem compared to alternatives such as AMD’s ROCm or Intel’s oneAPI – creating what he described as a “best-supported path effect.”

What is Slurm, and why does it matter

Slurm, originally developed at Lawrence Livermore National Laboratory, runs on roughly 60% of the world’s supercomputers. The software is in active use at major AI companies, including Meta Platforms, French AI startup Mistral, and Anthropic for elements of AI model training, Reuters reported.

Government supercomputers used for weather forecasting and national security research also depend on it. Nvidia acquired Slurm developer SchedMD in December 2025 and described the deal as a push to strengthen its open-source ecosystem and help users adopt newer AI techniques alongside traditional supercomputing work.

Is the concern valid?

Dr. Danish Faruqui, CEO of Fab Economics, a US-based AI hardware and datacenter advisory, said the risk was real.

“The skepticism that Nvidia may prioritize its own hardware in future software updates, potentially delaying or under-optimizing support for rivals, is a feasible outcome,” he said. As the primary developer, Nvidia now controls Slurm’s official development roadmap and code review process, Faruqui said, “which could influence how quickly competing chips are integrated on new development or continuous improvement elements.”

Owning the control plane alongside GPUs and networking infrastructure such as InfiniBand, he added, allows Nvidia to create a tightly vertically integrated stack that can lead to what he described as “shallow moats, where advanced features are only available or performant on Nvidia hardware.”

One concrete test of that, industry observers say, will be how quickly Nvidia integrates support for AMD’s next-generation chips into Slurm’s codebase compared with how quickly it integrates its own forthcoming hardware and networking technologies, such as InfiniBand.

Does the Bright Computing precedent hold?

Analysts point to Nvidia’s 2022 acquisition of Bright Computing as a reference point, saying the software became optimized for Nvidia chips in ways that disadvantaged users of competing hardware. Nvidia disputed that characterization, saying Bright Computing supports “nearly any CPU or GPU-accelerated cluster.”

Rawat said the comparison was instructive but imperfect. “Nvidia’s acquisition of Bright Computing highlights its preference for vertical integration, embedding Bright tightly into DGX and AI Factory stacks rather than maintaining a neutral, multi-vendor orchestration role,” he said. “This reflects a broader strategic pattern — Nvidia seeks to control the full-stack AI infrastructure experience.”

However, he said Slurm presented a fundamentally different challenge. “Deeply entrenched in supercomputing centers and academia, and effectively community-governed, Slurm carries high switching costs,” Rawat said. “Nvidia may influence but is unlikely to replicate the same tightly integrated control in markets dominated by established, neutral, and community-driven platforms.”

The open-source safety valve and its limits

Faruqui acknowledged that Slurm’s open-source licensing under a GNU GPL v2.0 licence offers some protection, including the community’s right to fork the project if Nvidia’s stewardship is seen as biased. But he cautioned that the option carried its own risks. “Slurm’s open-source status provides a safety valve with its limitations, but it is not a complete shield against vendor-neutrality,” he said.

The acquisition brought many of the world’s leading Slurm developers inside Nvidia, he noted, meaning a community-led fork would struggle to sustain the same pace of development.

Rawat described the situation as “a strategic dependency risk, not a crisis,” and said organisations should diversify GPU procurement, benchmark workloads across multiple vendor ecosystems, and develop internal expertise to modify or switch orchestration tools if needed.

Faruqui recommended that enterprise buyers negotiating Slurm support agreements seek service-level guarantees that apply equally to non-Nvidia hardware, covering response times, bug fixes, and feature parity across heterogeneous clusters. On architecture, he said organisations should consider containerising AI workloads to isolate applications from the underlying scheduler, making migration to alternative schedulers such as Flux or Kubernetes more feasible if required.

The article originally appeared in InfoWorld.

It’s not just your imagination; you are seeing more Macs being used in business environments these days — and that trend is expected to continue.

The latest Omdia/Informa US PC market data found that Apple took an 11% share of the US enterprise market last year. “For full-year 2025…, the biggest story at the vendor level was Apple, which has been making market share gains in US businesses, reaching an 11% share in full year 2025: up 2.4 percentage points from 2024,” said Kieren Jessop, research manager at Omdia:

Selected data points include :

• Across all segments (education, consumer, business) Macs grabbed 15.7% share in the last quarter of 2025.
• The Mac achieved an overall 16% market share across the year.
• Mac growth hit 11.2% in 2025 compared to industry average growth of 3.3%.

The data shouldn’t mask that enterprises are still dominated by Windows devices, but does give us a fairly useful temperature showing where the industry heat is right now.

What drove growth? 

The MacBook Air maintained its place on the throne as “Most Popular Notebook,” Jessop said. Apple, now in its 50th year,  also boosted memory to 16GB while pruning $100 off the cost of the ‘Air during the year. Those moves helped keep sales healthy. 

There may be more to come, Jessop suggested, particularly as the MacBook Neo enters public consciousness: “The $599 Neo extends that value trajectory and is expected to significantly disrupt the entry-level segment,” he said.

The Neo (reviewed here) is most certainly an inflection point for intense competition, the analyst noted. Available at just $499 in the weak education market, and $599 everywhere else, the new Mac aims squarely at entry-level users. The thing is, it breaks into this part of the market at the same time as widely-reported component cost increases kick in.

“Looking ahead, the outlook for 2026 is significantly more cautious,” said Jessop, predicting huge price increases in RAM and storage components. “Memory and storage costs have risen 40%–70% since the start of 2025,” he said. 

PC sales will feel the geopolitical heat

Climbing component prices are unlikely to change trajectory anytime soon, with the problem made worse by the growing conflagration in the Middle East. Oil is used in everything, from ferrying finished products around to creating the casing around cabling and manufacturing equipment of all kinds. Shortage in this one raw material will inevitably pour problems across the industry.

“Omdia expects at least a further 60% increase in mainstream PC memory and storage costs in Q1 2026,” the analyst said, predicting the greatest impacts on the sub-$500 segment, which includes most education and entry-level consumer devices. To put that into context, Omdia is currently forecasting a $90 to $165 increase in PC build costs  due to component shortages. These steep increases are expected to affect everyone, and while some of the larger manufacturers might be able to swallow a hit against margins, others will be unable to do the same. 

Apple is knocking on the door

We learned last week that Apple is moving quite aggressively, allegedly purchasing memory at top dollar prices and choosing to handle the pain. This secures its own supply, of course, but also makes it harder for others to buy the memory they need at a price they can afford.

The competitive threat it is putting in place is quite real. “As thinner margins and lower allocation priority constrain the low-end market, smaller vendors are especially at risk of being squeezed out of the market,” Jessop said.

Looking ahead, what seems most likely is that Windows systems comparable to the MacBook Air will begin to see increase prices, a move that will make the lower-cost Mac even more competitive. That’s particularly true across enterprises that need to deploy new kit, but face their own existential cost and supply chain-related challenges.

We have to wait and see how these forces play out, but it seems plausible to think Apple is nowhere near hitting the ceiling of its enterprise market share gains. The combination of its own strategies (from its various platforms, OSes, and Apple Silicon) and market reality seems to be forming a structural advantage the company should be able to exploit for years. 

“Apple’s vertical integration (own silicon, own OS) gives it more levers than competitors reliant on third-party chips and Microsoft licensing,” Hexnode CEO Apu Pavithran told me recently.

It’s almost as if years of carving out its own independent place means Apple now has in place strengths its competitors do not possess.

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

Tyto filmy a seriály jsou teď na českém Netflixu nejoblíbenější. Nerozlišujeme žánr, stáří ani hodnocení na filmových webech. Jde o souhrnnou oblíbenost za poslední týdny, kterou zjišťuje web FlixPatrol.

When talking about credential security, the focus usually lands on breach prevention. This makes sense when IBM’s 2025 Cost of a Data Breach Report puts the average cost of a breach at $4.4 million. Avoiding even one major incident is enough to justify most security investments, but that headline figure obscures the more persistent problems caused by recurring credential [email protected]