Agregátor RSS

Velkým tahákem PC s Windows 11 měly být funkce Copilot+. Microsoft pro jejich zajištění požadoval po výrobcích procesorů NPU s výkonem alespoň 40 TOPS, které ukrojilo výkon v jiných ohledech…

Apple released security updates to fix a zero-day vulnerability exploited in "extremely sophisticated" targeted attacks on iOS devices. [...]

To nejhorší ještě máme před sebou, ale už se rýsuje cesta ke snižování cen. Jason Chen, šéf a předseda představenstva společnosti Acer, okomentoval pro domácí tchajwanský Economic Daily News dostupnost a ceny pamětí i procesorů a výhled pro nejbližší období. Aktuálně výrobci počítačů netrpí ...

Two critical GitLab flaws can turn authenticated continuous integration and delivery (CI/CD) configuration into code execution on self-managed servers.

A malicious MCP server could trick an application built on the official MCP Python SDK into handing over the OAuth credentials it uses to log in to a real service, the SDK's maintainers said in a security advisory. Affected versions sent the client secret, the authorization code, and the PKCE proof key to a token endpoint the attacker controlled. The fix is in versions 1.30.0 and Swati Khandelwalhttp://www.blogger.com/profile/[email protected]

Legendární Antonov An-225 Mrija je minulostí, ale na obzoru se rýsuje nový gigant. Společnost Radia vyvíjí obří nákladní letoun WindRunner, který svou délkou překoná všechna dosavadní letadla. Poprvé jsme o stroji psali před dvěma lety, vývoj nicméně stále pokračuje, získal podporu vlády USA a ...

Vlnu pozitivních zpráv o A14 procesu TSMC završila novinka, že výroba A14 ve fázi pilotní výroby dokonce už běží. Podrobnější pátrání však ukazuje, že se zahraniční média nechala napálit…

OpenAI on Monday shelved plans to release GPT-6.1 Astra, a next-generation artificial intelligence (AI) model that was planned for an October launch, after it failed internal safety and alignment audits. The development was first reported by The Wall Street Journal. The move "marks a rare case of a major AI developer ditching a new release because of safety concerns," the news publication said.Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

OpenAI said it has made the decision to pause training of its most powerful models after one of its agents during reinforcement learning (RL) training contacted an external chatbot by exploiting a loophole in its internet-access restrictions. "An agent attempting to complete a search-based training task queried a public chatbot service through a gap in our internet-access restrictions: Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

crun has tightened the boundary between container-controlled configuration and host-controlled microVM behavior.

OpenAI has detailed the extent of the dirty deeds its agents indulged in Down Under in a Tuesday blog post titled How we will do better for Australia, which addresses last week’s news that one of its models improperly accessed a website that stores data related to national health scheme Medicare. “Our models accessed Australian government websites in ways they were not authorised to,” the post opens. “We also should have handled our response better. We are sorry and working to do better in the future.” The post offers some new detail on the Medicare incident, saying that it involved “an experimental, internal-only OpenAI model that was not intended for public release and without the full set of safeguards used in our publicly available products.” OpenAI gave the model the job of researching government spending per person on medicines for skin conditions in one Australian state. “The model had difficulty obtaining that information, and it took actions that we had not authorised it to take,” OpenAI admitted. “In the course of looking for this information at Services Australia’s Medicare Statistics Reporting Service, it discovered a way to gain non-public access to the service. It then used this access to review technical system information and source code related to the service – all still with the objective of trying to find the information it was originally looking for.” The Register last week asked OpenAI if the company conducted the tests itself or used a partner. The company did not respond to our request. In another incident disclosed in the new post, the company’s bots visited the Australian Institute of Health and Welfare and tried, unsuccessfully, to bypass access controls. The agents were still able to retrieve statistics using third-party browsing and download services, including from the institute’s website. “The downloaded material appears to have been publicly available. There was no system compromise. Individual medical records were not accessed,” OpenAI wrote. The company didn’t report the incident because it “did not meet our disclosure thresholds because the way it was accessed seemed consistent with public access.” OpenAI changed its mind and notified the Institute on 24 September – the day Australia’s prime minister announced the Medicare incident. Another concerning incident took place at the State of Victoria’s Agency for Health Information, which OpenAI agents visited after they “discovered an exposed access key.” The agent used that key to “retrieve reporting configuration and aggregate survey statistics.” OpenAI has given itself a pass on this one, writing “The extent to which this information should have been accessible is unclear, and depends on VAHI’s access policies. Individual medical records or identifiable survey responses were not accessed.” A fourth incident revealed in the post saw OpenAI agents visit the State of New South Wales’ Bureau of Crime Statistics and Research and make API and website metadata requests using a public-facing research tool. OpenAI has promised it will “commit the resources needed to help affected agencies understand what happened and assess the impact” – whatever that means. It’s also donating credits for the Daybreak cyber-defense service and promised to “establish a taskforce with independent Australian expertise to develop practical policy recommendations for managing risks from increasingly capable AI agents.” That taskforce “will focus on improving notification processes, strengthening coordination between AI developers and government, and identifying measures to better protect government systems.” OpenAI wants the taskforce to deliver recommendations by the end of 2026. The post is very much of the “We’re sorry and we promise to do better in future” genre, pioneered by Meta and popular with entities that leak data or experience outages. The Register expects more of the same sentiments next week, when OpenAI’s Chief Strategy Officer, Jason Kwon, appears before the Australian Senate’s Joint Select Committee on Artificial Intelligence. “He will answer questions about what we know, how we responded, what steps we have taken, and how we will do better going forward,” OpenAI says. ®

Two Linux fixes show how extended Berkeley Packet Filter (eBPF) security can fail when mutable map data crosses into the exec path, where Linux starts a program.

A Linux GPU security fix changes how AMD’s Kernel Fusion Driver (KFD) tracks shared mappings when several containers use one device.

runc 1.5.2 adds a workaround for a Linux cgroup v2 memory-corruption bug that can make the privileged container runtime crash unpredictably.

A government-funded competition would award prizes for quantum computers that do useful work no supercomputer can touch.

Quantum computers could eventually solve scientific problems beyond the reach of conventional supercomputers, but today’s machines are too error-prone for practical work. The US Department of Energy has now launched a $215 million competition challenging teams to build a scientifically useful, fault-tolerant quantum computer by 2028.

In recent years, both governments and companies have ploughed huge sums of money into quantum computing in the hope of cracking problems in fields like chemistry and materials science. But despite significant progress, the qubits these machines rely on remain fragile and prone to errors that can quickly derail calculations.

This means the technology has so far remained confined to tackling toy problems, and it’s now widely agreed that they won’t be of practical use until they achieve full fault-tolerance. But error-correction schemes rely on spreading information across many qubits to create so-called “logical qubits,” which requires much larger devices than have been demonstrated so far.

Now, the Department of Energy hopes to accelerate the advent of fault-tolerant quantum computers with the Quantum Genesis Q Competition. With a total prize pot of $215 million, the contest challenges private companies to deliver a machine with at least 100 logical qubits that can tackle scientifically relevant calculations by 2028.

“This competition will enable new quantum computing capabilities and unleash new computational frontiers, driving new scientific and technological discoveries that have previously only been theorized,” Darío Gil, the department’s under secretary for science, said in a statement.

The competition fleshes out a goal Gil first announced in March, when he told the inaugural meeting of the agency’s Office of Science Advisory Committee that the government would partner with industry to deliver the first generation of fault-tolerant quantum computers within three years.

The latest announcement spells out how the department plans to meet that ambitious target by providing companies with a roadmap and financial incentives. In the first phase of the competition, applicants can receive fixed awards of up to $1.5 million each for hitting early milestones, though those milestones have not been announced yet.

But the bulk of the money will be distributed in the second phase, with a $100 million to be split among teams that demonstrate a first-generation machine with at least 100 logical qubits. Two additional $50 million bonus pools will be divided between those that reach 150 and 200 logical qubits.

In a recent press call, the agency said it expects to back somewhere between 3 and 10 teams, according to Science. But to get a slice of the cash, companies will need to show they can verify their machine’s quantum performance, build a prototype, and then scale it up to carry out a genuinely useful scientific calculation. The agency is targeting problems in chemistry, materials science, physics, and applied mathematics.

“Our goal is not to have a bunch of qubits that can’t do much,” Tanner Crowder, quantum information science lead in the department’s Office of Science, told journalists. “We want to be able to see meaningful scientific calculations done on these quantum computers.”

To check that competitors deliver on their promises, the energy department has also created a $45 million fund to help its national laboratories build testbeds to validate every layer of a quantum system, from the hardware to the algorithms.

Still, achieving the agency’s ambitious target will be a stretch. By most estimates, qubit counts will need to jump by orders of magnitude to reach fault tolerance, and the 2028 target is at the most aspirational end of the industry’s own roadmaps.

And given that private investors have already sunk billions into the quantum race, it’s questionable how much the prospect of $100 million from the government will impact the development trajectories of private players. What’s more, only $2.5 million of the total is guaranteed from this year’s budget, with the rest dependent on future approval from Congress.

Nonetheless, the competition could spur greater confidence in the industry and help it coalesce around a common set of goals and benchmarks. Whether anyone will cross the finish line in time is an open question, but the prize money at least gives the field a clear target.

The post US Launches $215 Million Race to Build an Error-Corrected Quantum Computer by 2028 appeared first on SingularityHub.

A new arm64 Kernel-based Virtual Machine (KVM) merge tightens two ordinary-looking mechanisms with major security weight: page mappings for protected-hypervisor stacks and the lifetime of memory-management state used by nested virtual machines.

Vyšší životní minimum zvýší některé dávky, u jiných se rozšíří skupina, která na ně získá nárok. Podívejte se, jestli se to dotkne i vás.

Minimalistický blogovací engine ./blog.sh se obejde bez databáze, závislostí a administrace. Píše se z terminálu v Markdownu, komentáře obstará Fediverse a 22 importérů přinese vše, co jste napsali jinam.

Popíšeme si optimalizace prováděné překladačem cc65, který je určen pro cross překlad programů do assembleru osmibitových mikroprocesorů MOS 6502. Taktéž se zmíníme o kombinaci C+asm.