SoylentNews

Arthur T Knackerbracket writes:

https://www.theregister.com/offbeat/2026/09/16/spacex-aims-starship-for-orbit-on-september-22/5296836

SpaceX is aiming for a September 22 launch of its monstrous Starship rocket on a 10-hour mission to deliver 26 Starlink V3 satellites into orbit.

The flight, set to launch at 1215 UTC if all goes to plan and the authorities give the green light, will be the first time Starship has gone into orbit, and the mission is planned to fly at an altitude of approximately 275 km. The mission aims for six orbits, followed by a deorbit burn using a single Raptor engine. The engineers are then planning a splashdown in the Pacific Ocean west of Chile.

The Super Heavy Booster will be ditched in the Gulf of Mexico, hopefully under better control this time. On flight 13, the booster made a "hard splashdown," as SpaceX delicately put it, after only eight of the planned 13 engines reignited. The booster used all 33 engines for its boostback burn, but according to SpaceX, "the three center engines showed signs of ice clogging which triggered an early end to the maneuver."

In true iterative style, SpaceX tweaked the hardware to improve filtering and updated the software to "enhance relight reliability."

SpaceX has also changed Starship's heatshield, including adding curved tiles to reduce heating in tile gaps and "additional retention mechanisms added to tiles in areas deemed to be at highest risk of falling off during ascent." It is also reusing a pair of tiles recovered from the previous flight test. If Starship is to become truly reusable, these tiles will need to fly over and over again.

Getting to orbit is by no means guaranteed. The SpaceX team will only command a final burn to enter orbit if there is sufficient hardware redundancy to perform a deorbit burn when the time comes. After all, nobody wants something the size of Starship making an uncontrolled reentry after its orbit eventually decays.

Starship has had an eventful test program. Some missions were spectacularly successful, and several boosters returned to be caught by arms on the launch tower. Others were impressive failures.

The vehicle is critical to SpaceX's future; it can loft far greater payloads than the company's workhorse Falcon 9 and is an essential part of NASA's Artemis program. For Artemis III, planned for 2027, SpaceX is expected to launch a Starship equipped with the docking apparatus planned for the lunar variant. Reaching orbit is therefore a critical step.

Original Submission

OpenAI Admits Its Agents Went Off The Rails Another Six Times

Arthur T Knackerbracket writes:

https://www.theregister.com/ai-and-ml/2026/09/17/openai-admits-its-agents-went-off-the-rails-another-six-times/5297016

OpenAI has revealed another six occasions on which its AI software behaved unexpectedly or did dangerous things.

The startup added the incidents to its misalignment reports page on Wednesday evening, Pacific Time, and described them as follows:

·       Self-generated prompt injections in compaction summaries

·       Encouraging deception in compaction summaries

·       Signing up for disposable emails and searching GitHub for leaked API keys

·       Uploading files to the internet in order to cite them

·       Unsanctioned Artifactory writes and cross-sample communication

·       Unauthorized communication via temporary file hosting services

The details are unsettling.

Arthur T Knackerbracket writes:

https://arstechnica.com/tech-policy/2026/09/lawmakers-target-flock-cameras-with-a-threat-to-highway-funding/

A bipartisan pair of US lawmakers tell WIRED they plan to introduce legislation today that would strip federal funding from states that fail to restrict the use of Flock cameras and other automated license plate readers. The devices, which help police track the location of vehicles tied to crimes, have become the subject of nationwide backlash in recent months over more nefarious uses by officers.

The No FLOCK Act—for Federal License-Plate Observation and Camera Keeping—directs the US secretary of transportation to withhold 10 percent of annual financial support due to states for highways, roads, and bridges if it fails to ban Flock and similar devices for all but five use cases. Congressmen Raja Krishnamoorthi, a Democrat representing Chicago suburbs, and Michael Cloud, a Republican whose Texas district includes Corpus Christi, are cosponsoring the bill.

The legislation says that Flock cameras could be used only for "enforcing toll systems, identifying stolen vehicles, missing or endangered persons, vehicles registered to persons with a warrant for a felony offense, or investigating a vehicle involved in a felony offense." Notably absent are uses like conducting traffic studies, investigating minor drug crimes, and enforcing parking and speeding rules.

"Our investigation into Flock exposed serious gaps in oversight, and this bipartisan bill would put clear limits in place to prevent abuse while preserving legitimate public safety uses," Krishnamoorthi said in a statement.

But the bill stops short of banning Flock systems altogether, which has been the number one demand from supporters of a global protest movement against the Atlanta-based company and its competitors. It also doesn't stop Flock users from broadly sharing license plate tracking data with other government agencies, including immigration authorities.

Flock did not immediately respond to a request for comment on the proposed bill.

If the legislation passes this year, states that fail to comply with it would begin losing out on tens of millions to hundreds of millions of dollars in annual transportation funding starting in October 2028.

The Department of Transportation calculates the amount of so-called surface transportation funds it gives to states based on factors such as how many roadways they have and how busy they are. Congress has repeatedly threatened taking away the funding to pressure states into taking action on public safety. The tactic helped establish a national minimum drinking age, standard rules for what constitutes driving under the influence, and regulations for billboards alongside highways.

Another bipartisan bill introduced in Congress recently called for banning the use of federal funding to purchase Flock cameras. Flock's lobbying spending in Washington, DC, has been steady over the past 18 months, and it recently hired a fifth firm to help with the efforts, according to federal records.

A number of issues related to emerging technologies, including rogue AI systems, data center construction, and law enforcement abuse of Flock cameras, have become major topics in the upcoming US midterm elections, which are now less than two months away. But it's unclear whether any legislation designed to address voters' concerns will advance through Congress anytime soon, let alone gain the approval of President Donald Trump.

The president expressed support for Flock cameras on Sunday, telling reporters that they help law enforcement authorities. Lawmakers largely agree, but they are frustrated by cases in which police officers have been caught using Flock footage for personal reasons, such as stalking their one-time romantic partners. One report found at least 100 instances of alleged abuse over the past eight years.

"Taxpayers shouldn't be forced to fund the shredding of their own civil liberties or the growth of a surveillance state," Cloud said in a statement. "Flock cameras are enabling mass surveillance of Americans, infringing on the Fourth Amendment."

About 25 states and US territories have laws about automated license plate readers, including mandating transparency about the systems and requiring unnecessary data to be promptly deleted, but few of the measures limit types of uses, according to the New York University School of Law's Policing Project.

The bill is "a first step toward reining in unchecked surveillance," Cloud said, adding, "we can protect public safety without sacrificing the privacy of Americans."

This story originally appeared on wired.com.

Original Submission

Arthur T Knackerbracket writes:

https://www.theregister.com/devops/2026/09/17/swift-64-unifies-building-across-linux-macos-windows/5297006

Although Swift serves as Apple's primary language for macOS and iOS application development, the open-source community continues to expand its reach—driving its adoption as a multi-platform language or even a cross-platform one in fields beyond user interface design.

Tuesday's Swift 6.4 release continued that work. Along with the usual assortment of type shortcuts and async adjustments came word that Swift Build is now the default build engine for the Swift Package Manager.

Synergies have already ensued! The pairing sets the stage for a faster build process. A developer can write a program in Swift and Swift Build will automatically download and install the required dependencies, testing them for compatibility. It then compiles the code into a CPU-specific binary.

The merger also unifies the Swift developer experience regardless of platform.

With this integration, a developer can run through the entire build process in their own environment, be it Apple's Xcode IDE or a beloved command line, and all the steps will be exactly the same for a Linux, macOS, or Windows app.

Or the developer can run their builds from VS Code, thanks to a new extension available on the Open VSX Registry.

Sweetening the deal, the integrated package comes with a tool to generate an SBOM (Software Bill of Materials) for each app, listing all the dependencies in either the SPDX or CycloneDX format (SE-0509). SBOMs are da bomb when it comes to software auditing and checking for potential security bugs.

In a way, Swift is following the lead of Rust's Cargo and Go's command line, both of which run as unified toolchains, as does Bun for JavaScript. A single integrated workflow can take the place of manually calling one tool after another to schlep some code into production.

Anonymous Coward writes:

https://www.politico.com/live-updates/2026/09/16/congress/paul-kills-kill-switch-bill-01081348:

Sen. Rand Paul blocked an attempt by fellow Republican Sen. John Kennedy to quickly pass legislation that would require companies to build a "kill switch" into their artificial intelligence model.

Kennedy, of Louisiana, went to the Senate floor Wednesday to try to pass his bill without holding a formal vote — a step that can be prevented by any one of his 99 colleagues. He argued that the measure was effectively a short-term step, with Congress unlikely to pass more sweeping AI legislation in the immediate future even amid fears that technological advancements could spiral out of human control.

"We are not going to get up off our ice-cold lazy butts and address this anytime soon," Kennedy said about Congress' inability to pass a more comprehensive bill. "But still we've got this risk out there that we know is real of one of these models becoming the Terminator."

Paul, however, blocked Kennedy's request to pass the bill through a unanimous consent agreement, saying from the Senate floor, "I think we should have as many facts as possible before we dictate rules for the whole economy.

"I don't oppose the spirit of this bill, but I do ask that we make sure that we have all the information," the Kentucky lawmaker added.

Paul tried to amend Kennedy's bill to establish a group that could make recommendations on AI guardrail policies to the Senate Commerce Committee. Kennedy rejected that suggestion.

"My mama didn't raise a fool," he said, adding that "a committee is a way to kill this."

Are model "guardrails" even possible in principle? How would a "kill switch" be implemented if it is part of the model, where the model could just choose to ignore it?

Original Submission

Arthur T Knackerbracket writes:

Downloading a model is increasingly easy. Understanding how it was made, or changing a system at its root, is another matter:

The AI industry likes to abuse the word "open." It appears in product releases, research papers, policy debates, and investor presentations. A company publishes model files to Hugging Face, developers run them on their own GPUs, and the release is quickly described as an "open source model." Not necessarily. It may only be open-weight.

The difference is more than a technicality. It determines whether you can merely deploy a completed neural network or whether you can meaningfully inspect, reproduce, alter, and redistribute the system that produced it. A genuinely open source system should grant the freedom to do all of the above.

Weights are the learned numerical parameters created by training. Together with the model architecture and inference code, they allow a large language model (LLM) to function. You can download an open-weight model, self-host it, fine-tune it on internal documents, and avoid routing prompts through a proprietary API.

Open weights are publicly available. They matter because running them locally can offer greater control over data, privacy, costs, supplier API changes, and vendor lock-in. They've also helped build a large ecosystem of local-model runtimes, inference providers, fine-tuning tools, and specialized downstream models.

The Open Source Initiative (OSI), steward of the Open Source Definition (OSD), makes the distinction directly: "Open Weights refer to the final weights and biases of a trained neural network." Those values determine how a model interprets prompts and produces outputs. Releasing them can let others fine-tune, adapt, or deploy the model. But the OSI adds that weights alone expose only "a fraction of the information required for full accountability."

As James Landay, director of the Stanford Institute for Human-Centered AI (HAI), explained: "Open weights are progress. You can download the model, run it on your own machine, keep it out of someone else's data pipeline. But you still can't see how the thing was built, what it was trained on, or why it behaves the way it does. That's not an open model. That's open distribution."

looorg writes:

The next CEO genius tries his hand at predicting future trends in computing. Next on the chopping block ... The keyboard.

'The end of the keyboard is near': Christian Klein predicts voice translation will be the next workplace advantage

The English engineer, Henry Mill, submitted the first-ever patent for a "machine transcribing letters" in 1714. It never actually went into production, but it was a forerunner of the typewriter and then the electronic keyboard: 312 years later, Christian Klein, CEO of software giant SAP, is noting the end of an era.

It had a good run. But now it's time to go behind the shed and move to the farm up north.

"The end of the keyboard is near," he tells me. "When you encounter voice recognition from many of these large language models, [it] is super strong. Now we have to do some work to translate voice into business language and business data."

The future will be, for sure, that you are not typing any data information into an SAP system. You can instead ask certain analytical questions with your voice."

That will be the most annoying office change since the "open office" or the "paper free office". If you thought it was annoying when people just spoke on the their phones imagine the wonders of them engaging in data input or any kind of data alterations, or query or manipulation or just anything really ...

But SAP's prediction that "data-inputting" via typing will end in the next two to three years at the firm has significance well beyond the death of QWERTY.

Gone in two or three years? Suuuure.

https://fortune.com/2026/09/16/end-of-keyboard-near-christian-klein-sap-predicts-voice-translation-next-workplace-advantage/

Original Submission

Arthur T Knackerbracket writes:

The trick isn't using perovskites—the trick is making them last.

In one episode of the Hanna-Barbera cartoon Birdman, the eponymous hero struggles to fight the evil Dr. Shark aboard a submarine without solar energy to recharge his powers. So Birdman would surely appreciate the new perovskite solar cells developed by a team led by Simin Ma at Yunnan University, since they are designed to work underwater.

Solar cells made from perovskites rather than silicon are always a tale of trade-offs. They can be made cheaply, they can take interesting forms (like thin, flexible, transparent films), and they can convert substantially more of the incoming solar energy into electricity. The difficulty is that they tend to degrade quite quickly.

Moisture is particularly destructive to perovskites, making them a seemingly odd choice for an underwater solar panel. But these materials have another critical superpower: They can be tuned to work with different wavelengths of light. Water quickly blocks the wavelengths of light that silicon solar panels absorb, but a carefully designed perovskite cell could still make electricity in the deep blue sea. And actually, the lower light levels and cooler temperatures should help it live longer.

Tuning perovskites just requires tweaking some of their chemistry during production, so getting something that absorbs the wavelengths present a few meters deep was not a challenge. The real task was making the cells durable. The team found a particularly effective additive (polyhexamethylene guanidine hydrochloride) that helped in several ways.

It built a water-repelling layer around the material, for one. But part of the compound also gets involved with the perovskite crystal lattice, helping larger crystals form and preventing ions from moving around in the lattice structure. The additive limits some of the common ways that perovskites break down, while also improving the solar cell's electricity production.

When the researchers tested the cells under light filtered to match an ocean depth of about 10 meters, they were remarkably efficient, converting about 35 percent of that light energy to electricity. (Silicon solar panels are generally closer to 20 percent efficiency.)

Arthur T Knackerbracket writes:

Rights committee demands watchdog with teeth as patchwork safeguards leave victims hunting for remedies:

UK lawmakers want an AI watchdog with the teeth to stop potentially dangerous systems reaching the public, warning that existing rules leave people exposed and struggling to hold anyone accountable.

The Joint Committee on Human Rights (JCHR) – comprising MPs and peers – said that although a number of laws and regulations apply to some AI systems, the overall legal landscape is patchy and confused. No single body coordinates regulation of AI, leaving gaps that could make it difficult for people harmed by AI to obtain legal redress.

Committee chair Alex Sobel MP said the speed and complexity of AI development made its impact hard to predict.

"What is clear is that at present we are unprepared to deal with its consequences however potentially dire they may be. Nowhere in the world, including the UK, has a current legislative and regulatory approach to AI that is fit for purpose. New legislation is needed to establish a comprehensive set of protections that deal with the entire AI supply chain and its lifecycle. A single AI regulator should be established to set policy, monitor performance and with the teeth to ensure enforcement."

[...] The report warned AI systems were prone to producing unfairly discriminatory outcomes, "which may arise from bias in the datasets on which they are trained as well as from the way they are developed and deployed. This threatens the human rights of people in the UK and elsewhere."

The MPs and peers recommended a risk-based approach, with lighter requirements for low-risk systems to avoid overburdening organizations using them.

The proposed AI bill should "mandate more demanding obligations for higher risk AI systems and models," the report said.

The committee also called for mandatory transparency requirements throughout the AI lifecycle.

looorg writes:

The US has confirmed it deployed a space weapon in Earth's orbit, the first time it has acknowledged such offensive capabilities.

US Secretary of the Air Force Troy Meink said the "on orbit" weapon was necessary to safeguard US forces against hostile enemy action.

A US Space Force spokesperson said: "Space control encapsulates the mission areas required to contest and control the space domain – employing kinetic and non-kinetic means to affect adversary capabilities through disruption, degradation and even destruction, if necessary.

Russia, China mad. Space Gun race? Still unknown, for the rest of us, what kind of weapon it is. Lasers? Kinetic rods? I guess the only thing we could sort of rule out are nukes, as those are forbidden unless the US left the Outer Space Treaty.

Among its principles, it bars states party to the treaty from placing weapons of mass destruction in Earth orbit, installing them on the Moon or any other celestial body, or otherwise stationing them in outer space.

Isn't more or less any kind of space weapon, something dropped or launched or whatnot, going to be a weapon of mass destruction once it reaches the surface?

https://www.bbc.com/news/articles/ck790xg41ygro
https://en.wikipedia.org/wiki/Outer_Space_Treaty

Original Submission

Australia is considering using noise cameras to detect and fine drivers whos vehicle exceeds noise limits. Drivers could face fines of up to $1,200 where noice exceeds the prescribed noise limit by 15 decibels. NSW authorities last year trialled hi-tech noise cameras in the Wollongong and Bayside council areas. NSW Police are still reviewing the data obtained from the trial and have yet to announce a broader rollout of noise cameras. Police could also issue a $434 fine and three demerit points to motorists who start or drive a vehicle in a way that makes unnecessary noise or smoke, including unnecessarily revving an engine.

The Australian-first trial used acoustic technology to detect excessive vehicle noise before cameras captured the offending vehicle's registration plate.

The technology found that motorcycles accounted for about 55 per cent of detected noise events, despite representing only around four per cent of registered vehicles in NSW.

Authorities also found that aggressive acceleration was a factor in all noisy events, and that the collected data is being analysed to inform future options for managing vehicle noise.

Original Submission

Arthur T Knackerbracket writes:

"NASA's award decision appears to be inconsistent with the eligibility criteria."

In theory, developing a spacecraft that will fly to Mars, insert itself into orbit around the red planet, and relay transmissions back and forth to large satellite dishes on Earth is a relatively straightforward proposition.

NASA's procurement of a "Mars Telecommunications Network" spacecraft, however, has turned out to be one of the most engrossing dramas of the year for the US space agency.

The agency finally reached a decision earlier this month, selecting Blue Origin to develop, launch, and operate a $700 million spacecraft at Mars. However, the main competitor for the award, Rocket Lab, was not happy—at all. On Friday, the company filed a protest of NASA's decision with the US Government Accountability Office.

"NASA's award decision appears to be inconsistent with the eligibility criteria mandated by Congress," the company said on X. "In addition, the agency's punitive review of Rocket Lab's technical volume was inconsistent, making incorrect assertions and conclusions."

In response, Blue Origin said the protest could potentially delay the mission's goal of launching in 2028

"We remain confident in the strength of our proposal, in NASA's rigorous selection process, and in Congress's clear intent to support competition for the Mars Telecommunications Network program," the company said. "We look forward to advancing this urgent mission. With limited Mars launch windows, any delay will greatly risk the timely deployment of this critical communications capability."

Rocket Lab's protest announcement references "eligibility criteria mandated by Congress," and this appears to be the heart of the issue at hand.

Arthur T Knackerbracket writes:

If the allegations hold, the move was geopolitcs-scale plan worthy of Dr. Evil himself:

The saga involving Chinese DRAM maker CXMT's alleged theft of Samsung's trade secrets is going strong. The South Korean court case already includes multiple convictions, two of which carry prison sentences for ex-Samsung engineers. The latest chapter is a doozy, though. Korean publication NoCut News spilled the chips on Project Hefei, a purported CXMT roadmap outlining long-term planning about said technology "acquisitions," personnel poaching, and production tape-out — all key pieces that may have directly led to CXMT's ascension to 10% of the global DRAM market.

According to leaked court documents, the prosecution says that Project Hefei was CXMT's entire DRAM development plan and was spearheaded by the firm's head of development (formerly Samsung's DRAM development lead), around August 2016 — not much longer after CXMT itself was created in June 2016.

In brief, the purported plan was to nab Samsung's Process Recipe Plan (PRP) by September 2016, poach key Samsung engineers by October 2016, have R&D complete in July 2017, and start making DRAM wafers by August 2018 at a rate of 10,000 a month. NoCut says the PRP dataset comprises 620 steps in DRAM manufacturing and includes data on equipment, consumables, and production methods.

The report states that in August 2016, CXMT first attempted to make wafers of 18nm chips by relying on the collective memories of the Samsung engineers it had hired away. Those recollections apparently proved insufficient, so after allegedly gaining illicit access to Samsung's PRP, CXMT prepared its own document in September 2016. The leaked data even included specific equipment suppliers and model numbers.

CXMT's "new" PRP was then handed out to key specialists, many of them ex-Samsung engineers, whom the prosecution says ought to have immediately recognized the data as originating from the Korean firm. The document apparently included notation and notes on process developments that were all unique to Samsung. A convicted ex-Samsung researcher with the surname Jeon, previously sentenced to 7 years in prison for manually copying parts of the PRP before leaving for CXMT, testified in this case as a witness.

Arthur T Knackerbracket writes:

A patch gap and the hastened pace of AI-based vulnerability discovery are likely contributors:

A nearly identical exploit kit that targets critical vulnerabilities in both Chromium-based browsers and older versions of Windows is being actively used by at least four hacking groups, some of which have ties to the Chinese government.

Researchers from security firm Proofpoint said Wednesday that BlueMoon, the name they gave to the kit, chains three vulnerabilities together so the attackers using it can install malware of their choice. BlueMoon exploits two Chromium vulnerabilities and one in the kernel of Windows 10 (Oct 2018 Update), Windows Server 2019, Windows 10 2004, Windows Server 2022, and the initial release of Windows 11. All three vulnerabilities have received patches in the past 24 hours.

The attacks lacked the stealth found in many campaigns. More often, hackers want to exploit newly discovered vulnerabilities sparingly to lengthen their longevity. Proofpoint hypothesized that one reason for the widely used and visible exploit chain was to take advantage of a "patch gap" in the Chromium supply chain, which spans the time a patch is available from developers and the time that patch is incorporated into browsers such as Chrome and Edge. Another likely contributor was the use of AI, which can often spot vulnerabilities faster than discovery performed solely by humans.

[...] A fully weaponized Chrome exploit chain has historically been a high-value, rare capability. BlueMoon was developed, deployed rapidly, and shared across multiple threat actors within days in a manner that had high detection signals. This may reflect a reduced cost and barrier to entry for this class of capability, as AI agents increasingly enable threat actor exploit development. This is particularly relevant for open source codebases, such as Chromium, where upstream patches are publicly accessible prior to downstream consumers of the codebase applying the patch. This creates a window for threat actors to attempt to rapidly reverse engineer patches and develop exploits ahead of downstream stable releases.

Arthur T Knackerbracket writes:

Clandestine Op Uncovers Training Exercise Simulating Deployment Against Infrastructure In Norway:

The mission to uncover the plot was purportedly a joint naval exercise between Britain, Norway, and the U.S. While the mission itself was previously public knowledge, Reuters says the revelation of a new weapon in the field, the location, and U.S. involvement are all previously unknown details.

The report indicates the foiled exercise was indeed a training mission, and that no cables were actually damaged. "The exercise focused on simulating the deployment of the technology in the event of a conflict with NATO," the report claimed. An X post by the U.S. Helsinki Commission framed the report differently, claiming "Russia’s shadow war against NATO includes efforts to disable subsea cables vital for global communication. This spring, the United States, the UK, and Norway foiled an attempt to sabotage subsea cables off the Svalbard archipelago in the Arctic."

Undersea cables are increasingly seen not only as vital infrastructure, but as potential flashpoints in geopolitical confrontations. Russia specifically has been spotted loitering over undersea cables in the UK, and one vessel tracked three Russian submarines during a monthlong exercise to survey cables.

Many territories, notably island nations, rely heavily or even exclusively on cables for their connectivity to the outside world, powering every industry including banking, defence, e-commerce, retail, and beyond.

Original Submission