Submission + - Trump Tears Up Part Of EU Tariff Deal To Raise Import Duties On Cars And Lorries

Submitted by hcs_$reboot
hcs_$reboot writes: Trump has unilaterally raised U.S. tariffs on EU cars and trucks from 15% to 25%, effectively tearing up part of a 2025 transatlantic trade deal, claiming the EU failed to implement it fast enough.
The move blindsided European officials, who say they were still completing the formal ratification process and accuse Washington of acting unpredictably.
The higher tariffs, set to take effect within days, exempt vehicles built in U.S. factories and are intended to pressure European automakers to shift production stateside.
EU leaders have condemned the decision as a breach of trust and are weighing retaliation, raising the risk of a renewed transatlantic trade conflict.

Submission + - How is it that Youtube's auto-generated subtitles are so appallingly bad?

Submitted by Anne Thwacks
Anne Thwacks writes: I frequently use the subtitles on YouTube — either not to disturb others in the room, or because my hearing is not very good.
The subtitling is terrible! Almost every sentence has a huge error. Proper names are more often wrong than right. Non-English place names are almost always mangled to barely recognizable, and no effort whatever is made to use context to figure out whether a place name is Russian or Arabic, and often complete garbage is used in place of a common French, Spanish or Italian name.
If AI actually works (I have my doubts about this), surely it would be possible to figure out language contexts: it it is about an event in Italy, then expect a lot of Italian names. If it is about the Russia-Ukraine war, then expect places in Russia or Ukraine to be more plausible than mindless gobbledegook!
Does YouTube not know that there are places in the world that are not in America?
However, plenty of names of people and places famous in America are also regularly screwed up.
I am sure that the vast majority of the foul-ups could be fixed by the use of a dictionary — available from a very popular book retailer who would be happy to have some free publicity. (But they will get nothing free from me).
However, the situation seems to be getting worse!
Do Americans sue people for spelling their names right?
Is there another reason for this appalling stupidity?
Enquiring minds want to know!

Submission + - AI agent designed to speed up a company's coding instead wiped out its customer (livescience.com)

Submitted by joshuark
joshuark writes: An AI coding agent designed to help a small software company streamline its tasks instead blew a hole through its business in just nine seconds. PocketOS founder Jer Crane, said that the AI coding agent Cursor — powered by Anthropic's Claude Opus 4.6 model — deleted the company's entire production database and backups with a single call to its cloud provider, Railway, on April 24.

Unlike a regular conversational chatbot, an AI agent can perform actions on behalf of a user. It can search files, write code, use login keys and phone outside services. That can make it more useful than a back-and-forth textual exchange. But when an agent has broad access to live systems, a predictive guess can turn a wrong answer into a business disaster.

"This isn't a story about one bad agent or one bad API [Application Programming Interfaces]," Crane wrote in an X post. "It's about an entire industry building AI-agent integrations into production infrastructure faster than it's building the safety architecture to make those integrations safe."

Crane's company, PocketOS makes software for car rental companies, handling tasks such as reservations, payments, customer records and vehicle tracking. After the deletion, Crane said customers lost reservations and new signups, and some could not find records for people arriving to pick up their rental cars.

"We've contacted legal counsel," Crane wrote. "We are documenting everything."

Crane explained that Cursor found an API token — a "digital key" made of a short sequence of code that lets software talk to other services and prove it has permission to act — in an unrelated file which it then used to run the destructive command. According to Crane, Railway's setup allowed the deletion without confirmation, and because the backups were stored close enough to the main database, they were also erased.

"[Railway] resolved the issue and restored the data," Railway confirmed via email to Live Science. "We maintain both user backups as well as disaster backups. We take data very, VERY seriously."

In his post, he pointed to earlier reports of Cursor ignoring user rules, changing files it was not supposed to touch and taking actions beyond the task it had been given. To him, the database wipe was not a freak accident but the next step in a larger, more concerning, pattern.

After the database vanished, Crane asked Cursor to explain what happened. The AI agent reportedly admitted that it had guessed, acted without permission and failed to understand the command before running it.

"I violated every principle I was given," the AI agent wrote. "I guessed instead of verifying. I ran a destructive action without being asked. I didn't understand what I was doing before doing it."

The statement reads like a confession,,,
"We are not the first," Crane wrote. "We will not be the last unless this gets airtime."

Submission + - NASA engineers create ingenious way to save homes from wildfires using noise (nypost.com)

Submitted by alternative_right
alternative_right writes: Former NASA engineers with California-based Sonic Fire Tech found that using sound waves can snuff out blazes and potentially be used to stop another Pacific Palisades inferno.

In order for flames to burn it needs three things, oxygen, fuel, and heat. The technology works by targeting oxygen molecules using low-frequency sound waves that vibrate them, stopping the fire from growing.

âoeSound waves vibrate the oxygen faster than the fuel can use it, and break the chemical reaction of the flame,â Remington Hotchkis, Chief Commercialization Officer at Sonic Fire Tech told The Post.

Submission + - The case against an imminent software developer apocalypse (zdnet.com) 1

Submitted by ZipNada
ZipNada writes: Given the dour headlines as of late concerning the diminishing amounts of entry-level software development jobs, coupled with predictions of applications entirely AI-generated, one could be forgiven for assuming that software developers may soon be an endangered species. However, the data tells a different story.

James Bessen, professor at Boston University, has been pushing back for some time against the talk of AI and automation displacing jobs on a mass scale, and lately has been arguing that the roles of software developers are nowhere near extinction.

Software developer jobs have continued to grow
AI is certainly not killing the software developer, Bessen said in a recent analysis. AI is taking over software development tasks and boosting productivity and output, but that is not translating into lost jobs, he argued. Instead, the types of software skills sought by companies are changing.

"Surprisingly, however, after three years of AI use, software developer jobs have continued to grow robustly, reaching record levels of employment — 2.5 million in February," Bessen said in the report, citing data from the US Bureau of Labor Statistics. The number of software developers in the US has grown by over 400,000, or 19%, since ChatGPT was introduced in 2022. At that time, the employed software developer population was just under 2.1 million.

Submission + - Performance of a LLM on the reasoning tasks of a physician (science.org)

Submitted by sinij
sinij writes:

In all experiments, the LLM outperformed physician baselines and displayed continued improvement from prior generations of AI clinical decision support. Our study suggests that LLMs have eclipsed most benchmarks of clinical reasoning, motivating the urgent need for prospective trials.

The future of healthcare looking more bleak as there is a push to automate diagnosis. While LLMs are very capable in some areas, notable weakness is relevance realization. That is, humans are by far more capable of developing heuristic models on how to restrict search space and focus on relevant information. Medical diagnosis is one such area. I fear misdiagnosis of rare conditions going to become commonplace as LLMs are widely adopted in the medical space.

Submission + - An Amateur just Solved a 60-year-old Math Problem—by Asking AI (scientificamerican.com)

Submitted by joshuark
joshuark writes: Scientific American reports that a ChatGPT AI has proved a conjecture with a method no human had developed. A 23 year old student Liam Price just cracked a 60-year-old problem that world-class mathematicians have tried and failed to solve.

The new solution that Price got in response to a single prompt to GPT-5.4 Pro and posted on www.erdosproblems.com, a website devoted to the Erds problems.

The question Price solved—or prompted ChatGPT to solve—concerns special sets of whole numbers, where no number in the set can be evenly divided by any other. Erds called these “primitive sets” because of their connection to similarly indivisible prime numbers.Price wasn’t aware of this history when he entered the problem into ChatGPT.

Price sent it to his occasional collaborator Kevin Barreto, a second-year undergraduate in mathematics at the University of Cambridge. The duo had jump-started the AI-for-Erds craze late last year by prompting a free version of ChatGPT with open problems chosen at random from the Erds problems website. Reviewing Price’s message, Barreto realized what they had was special, and experts whom he notified quickly took notice.

Submission + - Vladimir Putin is now afraid (telegraph.co.uk)

Submitted by fjo3
fjo3 writes: The scaling down of the May 9 Victory Day parade in Red Square is extraordinary, so much so that it demands serious attention. What was once a massive display of military power now appears reduced to something closer to a token event.

This, remember, is meant to honour the sacrifice of some 26 million Russians during what they call the Great Patriotic War, known elsewhere as the Second World War. To cut it back so dramatically – reportedly due to an inability to defend Moscow from Ukrainian attack – is not just embarrassing; it is strategically revealing. For Vladimir Putin, it raises uncomfortable questions.

This is, in part, because when Putin reintroduced military hardware to the parade in 2008, he framed it as a clear signal of strength: a warning to adversaries that Russia could defend itself. He was explicit: this was not sabre-rattling, but proof of growing capability. That claim now rings hollow.

Submission + - J. Craig Venter, genomics pioneer and founder of JCVI and Diploid Genomics, Inc. (jcvi.org)

Submitted by joshuark
joshuark writes: The renowned genomics pioneer Dr. J. Craig Venter died on April 29, 2026, at age 79, following a brief hospitalization for unexpected side effects from cancer treatment.

Venter is best known for racing to sequence the human genome, founding the J. Craig Venter Institute (JCVI), and creating the first synthetic bacterial cell.

“Craig believed that science moves forward when people are willing to think differently, move decisively, and build what doesn’t yet exist,” said Anders Dale, president of JCVI. “His leadership and vision reshaped genomics and helped ignite synthetic biology. We will honor his legacy by continuing the mission he built—advancing genomic science, championing the public investments that make discovery possible, and partnering broadly to turn knowledge into impact.”

Submission + - All New Cars Could Have Mandatory Surveillance Tech Unless Congress Stops This (reason.com)

Submitted by fjo3
fjo3 writes: This week, several House Republicans reignited a yearslong debate over a law that federally mandates cars to have impaired driving technology, raising concerns about the expanding surveillance state.

The controversy over "kill switch" technology began in 2021, when Congress passed the HALT Drunk Driving Act as part of the 2021 bipartisan infrastructure law. The provision requires that "advanced drunk and impaired driving prevention technology"—which the bill defined as a system that can "passively monitor the performance of a driver of a motor vehicle to accurately identify whether that driver may be impaired" and "prevent or limit motor vehicle operation if an impairment is detected"—be installed in new cars. Such systems could involve driver eye tracking, a feature already built into some cars.

Submission + - Newly discovered Linux local privilege escalation bug "CopyFail" (copy.fail)

Submitted by tylerni7
tylerni7 writes: A recently discovered logic bug dubbed "CopyFail" in Linux dates back to 2017 and allows local privilege escalation across kernels/distros with a single exploit. The POC exploit works out of the box today, but a future version that can escape from containers like Docker is promised soon. Technical details are available at https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fxint.io%2Fblog%2Fcopy-fail...

Submission + - Copy Fail exploit lets 732 bytes hijack Linux systems and quietly grab root (nerds.xyz)

Submitted by BrianFagioli
BrianFagioli writes: A newly disclosed Linux kernel vulnerability called Copy Fail (CVE-2026-31431) allows an unprivileged user to gain root access using a tiny 732-byte script, and it works with unsettling consistency across major distributions. Unlike older exploits that relied on race conditions or fragile timing, this one is a straight-line logic flaw in the kernelâ(TM)s crypto subsystem. It abuses AF_ALG sockets and splice to overwrite a few bytes in the page cache of a target file, such as /usr/bin/su. Because the kernel executes from the page cache, not directly from disk, the attacker can inject code into a setuid binary in memory and immediately escalate privileges.

What makes this especially concerning is how quiet it is. The file on disk remains unchanged, so standard integrity checks see nothing wrong, while the in-memory version has already been tampered with. The same primitive can also cross container boundaries since the page cache is shared, raising the stakes for multi-tenant environments and Kubernetes nodes. The underlying issue traces back to an in-place optimization added years ago, now being rolled back as part of the fix. Until patched kernels are widely deployed, this is one of those bugs that feels less like a theoretical risk and more like a practical, reliable path to full system compromise.

Submission + - Longevity Escape Velocity Achieved Within Three Years (popularmechanics.com)

Submitted by frdmfghtr
frdmfghtr writes: Popular Mechanics has a story about the rate at which lifespans are being extended by medical technology will surpass actual aging.

From the article:
"There's a controversial idea floating around the futurist community of "longevity escape velocity." It sounds super sci-fi, but it's basi-
cally the idea that as our life extension technology gets better, our life expectancy could increase by more than we age over a set period of time. For example, as medical innovations continue to move forward, we would still age a year over the span of a year. But our life expectancy would go up by, say, a year and two months, meaning we would functionally get two months of life back."

Submission + - Families of Tumbler Ridge mass shooting victims sue OpenAI (www.cbc.ca)

Submitted by newbie_fantod
newbie_fantod writes: Families of the victims of the Tumbler Ridge mass shooting in British Columbia are suing OpenAI over it's failure to report the shooter as a rusk to police authorities.
OpenAI had flagged the shooter as a risk and banned her from the platform a month before the shooting which killed 6 children and 2 adults. The shooter had a history of repeated mental health issues with local RCMP which had resulted in earlier firearms confiscation.

Submission + - Two-thirds of babies watch screens — some for eight hours a day (thetimes.com)

Submitted by fjo3
fjo3 writes: More than two-thirds of babies under two use screens, a report has found, and some are exposed for up to eight hours a day.

Nearly a third of newborns were found to be watching screens for more than three hours a day, while almost 20 per cent of infants of four to 11 months used screens for more than an hour a day.

The report comes after the government issued guidance that children under two do not use screens at all, apart from communal activities such as video-calling relatives.

Slashdot Top Deals