SoylentNews

Arthur T Knackerbracket has processed the following story:

The human brain contains about 86 billion neurons. These cells fire electrical signals that help the brain store memories and send information and commands throughout the brain and the nervous system.

The brain also contains billions of astrocytes — star-shaped cells with many long extensions that allow them to interact with millions of neurons. Although they have long been thought to be mainly supportive cells, recent studies have suggested that astrocytes may play a role in memory storage and other cognitive functions.

MIT researchers have now put forth a new hypothesis for how astrocytes might contribute to memory storage. The architecture suggested by their model would help to explain the brain’s massive storage capacity, which is much greater than would be expected using neurons alone.

“Originally, astrocytes were believed to just clean up around neurons, but there’s no particular reason that evolution did not realize that, because each astrocyte can contact hundreds of thousands of synapses, they could also be used for computation,” says Jean-Jacques Slotine, an MIT professor of mechanical engineering and of brain and cognitive sciences, and an author of the new study.

Dmitry Krotov, a research staff member at the MIT-IBM Watson AI Lab and IBM Research, is the senior author of the open-access paper, which appeared May 23 in the Proceedings of the National Academy of Sciences. Leo Kozachkov PhD ’22 is the paper’s lead author.

Astrocytes have a variety of support functions in the brain: They clean up debris, provide nutrients to neurons, and help to ensure an adequate blood supply.

Astrocytes also send out many thin tentacles, known as processes, which can each wrap around a single synapse — the junctions where two neurons interact with each other — to create a tripartite (three-part) synapse.

Within the past couple of years, neuroscientists have shown that if the connections between astrocytes and neurons in the hippocampus are disrupted, memory storage and retrieval are impaired.

Arthur T Knackerbracket has processed the following story:

Germany’s Braunschweig Regional Court has reportedly sentenced four Volkswagen executives to jail over “Dieselgate” – the 2015 scandal in which the automaker was found to have fudged software used to test its vehicles’ pollution emissions.

The matter concerned nitrogen oxide emissions, which software installed by Volkswagen rated as low enough to comply with environmental laws. Independent tailpipe testing produced different results, leading the USA’s Environmental Protection Agency to claim the German auto giant had violated clean air laws.

Volkswagen staff admitted to a conspiracy to fudge emission results, the company pled guilty and paid substantial fines and recalled around 11 million cars.

In 2019 German authorities charged four execs of the matter, and Braunschweig Regional Court decided one of the resulting cases on Monday.

German media reports explain that four VW execs were jailed for between 15 and 54 months, with the judge finding all knowingly contributed to the scandal.

The trial ran for almost four years, saw over 150 witnesses take the stand and the court sit for over 170 days. This is not the end of the matter because cases against another 31 defendants remain on the books of German courts. Another case, against Martin Winterkorn, the former CEO of Volkswagen AG, remains stalled due to his age and ill health.

Dieselgate also claimed execs at Audi and Nissan.

In an unrelated 2022 incident, Toyota-owned truck brand Hino admitted to have reported inaccurate emissions data for 20 years.

Software continued to bedevil Volkswagen, which showed CEO Herbert Diess the door in 2022 after his efforts to turn around the company’s software unit stalled.

Original Submission

_{Processed by jelizondo}

hubie writes:

A simple text editor that dates back to Windows 1.0 is getting smartified:

Microsoft has continued to shovel AI into its built-in Windows inbox apps, and now it's rolling out a Notepad update that will use Copilot to write text for you.

The updates come in the same week that Redmond released a snappy, lightweight command line editor that is the antithesis of what the venerable Notepad has become.

Notepad's Write feature requires users to sign in with their Microsoft account, select where they want the new content to go (or make a selection for reference), and then choose Write from the Copilot menu to prompt the AI to generate text, which you can review and insert into Notepad if it fits your needs.

The output can then be kept, discarded, or refined with follow-up prompts. Copilot can be disabled in the app's settings.

It is unclear who asked for this, or why Microsoft thinks users of a once-simple text editor require this assistance. If it were necessary, then surely an app like WordPad would have been a better place (if Microsoft hadn't killed it off, presumably so it could better focus on bloating Notepad).

At least with Outlook in Microsoft 365, the idea of letting an AI write emails for you could appeal to overly busy middle managers or people who struggle to come up with written communication. But Notepad is primarily used for quick and dirty tasks like jotting down ideas or pasting fussily formatted files into plain text to make the actual text more easily portable.

Microsoft's AI ambitions for Notepad first appeared just over a year ago. In November 2024, a "Rewrite" function turned up, with options to tweak text based on the tone, format, and length requirements of a user.

Notepad dates back to Windows 1.0 and remained more or less unchanged for decades, with only an occasional fix. In recent years, however, it has undergone multiple tweaks and enhancements at the hands of Microsoft, culminating in the new generative AI features.

Arthur T Knackerbracket has processed the following story:

If you’ve been around a while you might recall that Verizon used to be utterly obnoxious when it came to absolutely everything about using your mobile phone. Once upon a time, the company banned you from even using third-party apps (including basics like GPS), forcing you to use extremely shitty Verizon apps. It also used to be absolutely horrendous when it came to unlocking phones, switching carriers, and using the device of your choice on the Verizon network.

Two things changed all that. One, back in 2008 when the company acquired spectrum that came with requirements that users be allowed to use the devices of their choice. And two, as part of merger conditions affixed to its 2021 acquisition of Tracfone. Thanks to those two events Verizon was dragged, kicking and screaming, into a new era of openness that was of huge benefit to the public.

Now, with the Trump administration openly destroying whatever’s left of U.S. federal corporate oversight and consumer protection standards, Verizon sees an opportunity. As Jon Brodkin at Ars Technica notes, Verizon’s attempting to get the Trump administration to kill all unlocking requirements, in a bid to drag everyone back to the dark ages of cellphone use.

Verizon being Verizon, they can’t help but lie about it in a petition to the Trump FCC, claiming that they simply must be allowed to unfairly lock down mobile devices, because doing anything else harms competition and helps criminals:

[...] These openness requirements are somewhat scattershot across carriers, which is why the Biden FCC had been proposing a uniform rule that would have required that all wireless providers unlock devices within 60 days of purchase.

Not only is that effort dead now thanks to Trump’s election, but Verizon’s pushing to eliminate all such requirements, driving progress violently backward. Verizon’s hoping that such rollbacks can be part of FCC boss Brendan Carr’s “Delete, Delete, Delete” deregulatory bonanza, in which he’s destroying longstanding consumer protection standards under the pretense of government efficiency.

Verizon even name drops Elon Musk’s DOGE efforts in their petition, insisting that longstanding and popular consumer protection standards on wireless devices are “the perfect example of the type of rule that the Commission should eliminate as part of the Department of Government Efficiency’s Deregulatory Initiative.”

Even if the rules aren’t destroyed by the Trump FCC, numerous recent Trump court rulings and executive orders make it all but impossible for regulators to enforce most consumer protection rules. But Verizon, ever a fan of crushing consumer protection standards and competition, wants to make doubly sure.

Original Submission

oaklandwatch writes:

Stack Overflow will test paying experts to answer questions. That's one of many radical experiments they're now trying to stave off an AI-induced death spiral.

Questions and answers to the site have plummeted more than 90% since April of 2020. So here's what Stack Overflow will try next.

• They're bringing back Chat, according to their CEO (to foster "even more connections between our community members" in "an increasingly AI-driven world").
• They're building a "new Stack Overflow" meant to feel like a personalized portal. "It might collect videos, blogs, Q&A, war stories, jokes, educational materials, jobs... and fold them together into one personalized destination."
• They're proposing areas more open to discussion, described as "more flexible Stack Exchanges... where users can explore ideas or share opinions."
• They're also licensing Stack Overflow content to AI companies for training their models.
• Again, they will test paying experts to answer questions.

Original Submission

An Anonymous Coward writes:

Over 9,000 ASUS routers are compromised by a novel botnet dubbed "AyySSHush" that was also observed targeting SOHO routers from Cisco, D-Link, and Linksys.

The campaign was discovered by GreyNoise security researchers in mid-March 2025, who reports that it carries the hallmarks of a nation-state threat actor, though no concrete attributions were made.

The threat monitoring firm reports that the attacks combine brute-forcing login credentials, bypassing authentication, and exploiting older vulnerabilities to compromise ASUS routers, including the RT-AC3100, RT-AC3200, and RT-AX55 models.

Specifically, the attackers exploit an old command injection flaw tracked as CVE-2023-39780 to add their own SSH public key and enable the SSH daemon to listen on the non-standard TCP port 53282. This modifications allow the threat actors to retain backdoor access to the device even between reboots and firmware updates.

"Because this key is added using the official ASUS features, this config change is persisted across firmware upgrades," explains another related report by GreyNoise.

"If you've been exploited previously, upgrading your firmware will NOT remove the SSH backdoor."

The attack is particularly stealthy, involving no malware, while the attackers also turn off logging and Trend Micro's AIProtection to evade detection.

Characteristically, GreyNoise reports logging just 30 malicious requests associated with this campaign over the past three months, though 9,000 ASUS routers have been infected.

Still, three of those requests were enough to trigger GreyNoise's AI-powered analysis tool that flagged them for human inspection.

The campaign likely overlaps with the activity Sekoia tracks as "Vicious Trap," disclosed last week, though the French cybersecurity firm reported that threat actors leveraged CVE-2021-32030 to breach ASUS routers.

In the campaign seen by Sekoia, the threat actors were observed targeting SOHO routers, SSL VPNs, DVRs, and BMC controllers from D-Link, Linksys, QNAP, and Araknis Networks.

The exact operational goal of AyySSHush remains unclear, as there are no signs of distributed denial of service (DDoS) or using the devices to proxy malicious traffic through the ASUS routers.

However, in the router breaches observed by Sekoia, a malicious script was downloaded and executed to redirect network traffic from the compromised system to third-party devices controlled by the attacker.

Currently, it appears the campaign quietly builds a network of backdoored routers to create the groundwork for a future botnet.

ASUS has released security updates that address CVE-2023-39780 for the impacted routers, though the exact time of availability varies per model.

Users are recommended to upgrade their firmware as soon as possible and look for suspicious files and the addition of the attacker's SSH key (IoCs here) on the 'authorized_keys' file.

Also, GreyNoise lists four IP addresses associated with this activity, which should be added to a block list.

101.99.91[.]151
101.99.94[.]173
79.141.163[.]179
111.90.146[.]237

If a compromise is suspected, a factory reset is recommended to clean the router beyond doubt and then reconfigure it from scratch using a strong password.

Links in article:

https://www.greynoise.io/blog/stealthy-backdoor-campaign-affecting-asus-routers
https://nvd.nist.gov/vuln/detail/CVE-2023-39780
http://www.labs.greynoise.io/grimoire/2025-03-28-ayysshush/
https://blog.sekoia.io/vicioustrap-infiltrate-control-lure-turning-edge-devices-into-honeypots-en-masse/
https://www.labs.greynoise.io/grimoire/2025-03-28-ayysshush/

Original Submission

taylorvich writes:

https://phys.org/news/2025-05-bed-bugs-human-pest.html

Ever since a few enterprising bed bugs hopped off a bat and attached themselves to a Neanderthal walking out of a cave 60,000 years ago, bed bugs have enjoyed a thriving relationship with their human hosts.

Not so for the unadventurous bed bugs that stayed with the bats—their populations have continued to decline since the Last Glacial Maximum, also known as the ice age, which was about 20,000 years ago.

A team led by two Virginia Tech researchers recently compared the whole genome sequence of these two genetically distinct lineages of bed bugs. Published in Biology Letters, their findings indicate the human-associated lineage followed a similar demographic pattern as humans and may well be the first true urban pest.

"We wanted to look at changes in effective population size, which is the number of breeding individuals that are contributing to the next generation, because that can tell you what's been happening in their past," said Lindsay Miles, lead author and postdoctoral fellow in the Department of Entomology.

According to the researchers, the historical and evolutionary symbiotic relationship between humans and bed bugs will inform models that predict the spread of pests and diseases under urban population expansion.

By directly tying human global expansion to the emergence and evolution of urban pests like bed bugs, researchers may identify the traits that co-evolved in both humans and pests during urban expansion.

"Initially with both populations, we saw a general decline that is consistent with the Last Glacial Maximum; the bat-associated lineage never bounced back, and it is still decreasing in size," said Miles, an affiliate with the Fralin Life Sciences Institute. "The really exciting part is that the human-associated lineage did recover and their effective population increased."

Miles points to the early establishment of large human settlements that expanded into cities such as Mesopotamia about 12,000 years ago.
"That makes sense because modern humans moved out of caves about 60,000 years ago," said Warren Booth, the Joseph R. and Mary W. Wilson Urban Entomology Associate Professor.

fliptop writes:

At 22, I left LA for a traditional Oxford education. I mocked it then—but this stodgy approach might be our best hope right now:

The level of AI cheating has reached such an extreme that many fear we've reached a point of no return.

• "Everyone is cheating their way through college," New York magazine recently announced. "ChatGPT has unraveled the entire academic project."
• The Wall Street Journal has reached a similar conclusion: "There's a good chance your kid uses AI to cheat," declared a recent article.
• "AI cheating is getting worse," The Atlantic agrees, and "colleges don't have a plan."

Even worse, cheaters are getting all the rewards. A Columbia student recently got kicked out for cheating—and he turned around and raised millions to turn his system into a startup.

[...] When I was 22 years old, I woke up one morning in a hot dingy apartment on the cusp of South Central LA—where I'd been sleeping on the floor. I grabbed two suitcases I'd packed the night before, and caught a ride to the Los Angeles International Airport.

[...] My destination was the University of Oxford—a place I'd never seen and only knew through hearsay and Hollywood movies.

[...] How would the Oxford system kill AI?

[...] If implemented today, the Oxford system would totally elminate AI cheating—in these five ways:

1. EVERYTHING WAS HANDWRITTEN—WE DIDN'T EVEN HAVE TYPEWRITERS.
2. [...] MY PROFESSORS TAUGHT ME AT TUTORIALS IN THEIR OFFICES. THEY WOULD GRILL ME VERBALLY—AND I WAS EXPECTED TO HAVE IMMEDIATE RESPONSES TO ALL THEIR QUESTIONS.
3. [...] ACADEMIC RESULTS WERE BASED ENTIRELY ON HANDWRITTEN AND ORAL EXAMS. YOU EITHER PASSED OR FAILED—AND MANY FAILED.
4. [...] THE SYSTEM WAS TOUGH AND UNFORGIVING—BUT THIS WAS INTENTIONAL. OTHERWISE THE CREDENTIAL GOT DEVALUED.
5. [...] EVEN THE INFORMAL WAYS OF BUILDING YOUR REPUTATION WERE DONE FACE-TO-FACE—WITH NO TECHNOLOGY INVOLVED.

The author goes into detail on each of the 5 points, and speculates why it wouldn't be feasible in the US.

Related:

• Parents Sue After Student Punished For Using AI In Class
• OpenAI Afraid to Release ChatGPT Detection Tool That Might Piss Off Cheaters
• Dishonor Code: What Happens When Cheating Becomes the Norm?
• These Students Figured Out their Tests Were Graded by AI -- And The Easy Way to Cheat

Original Submission

Arthur T Knackerbracket has processed the following story:

Chinese hardware giant Lenovo thought it had prepared for a trade war, but its plan proved insufficient once the US started to rapidly change its tax policies in imported goods.

"We are not worried about the tariff," CEO Yuanqing Yang told investors on Thursday during the company's Q4 FY24/25 earnings call. "We are worried about the uncertainty and the quick changes."

Yang explained that Lenovo manufactures its products in many countries, using a mixture of its own facilities and contract manufacturing firms. The company moves production to the optimal location to cope with customer needs and geopolitical conditions, and calls its strategy "China Plus".

The CEO told investors "no other country can replace China" as the Middle Kingdom is "the most competitive manufacturing country with low cost, high efficiency and aggregation of supply chain."

Yang said Lenovo's plans worked when the Trump administration announced universal ten percent tariffs, but not so much when the 25 percent tariff on goods from Mexico and Canada was announced in March and "implemented so suddenly that we didn't even have time to prepare.

"It had a significant impact on our performance last quarter," he said, suggesting an impact of $50 million to $60 million last quarter. He later labelled the overall impact of tariffs as "notable."

That impact didn't stop Lenovo from posting stonking quarterly results [PDF] for the financial year ended March 31, 2025.

[...] Lenovo has long hoped to build an enterprise hardware business to rival the likes of HPE or Dell. Its cloud service provider business is now a profitable $10 billion revenue concern, meaning enterprise sales were around $4.5 billion.

It will come as no surprise to readers that Lenovo attributed its enterprise hardware growth to demand for AI infrastructure, which boosted sales of its servers and liquid cooling kit.

Original Submission

An Anonymous Coward writes:

When you build a server according to your plan and requirements, you want it to run quickly and efficiently, right? But did you know that modern Linux systems, especially those using systemd, often install and run many services by default, even if you don't need them? These unwanted services consume precious system resources and can even become security risks.

In this article, we'll walk through how to identify and disable unnecessary services on systemd-based Linux distributions like Fedora, CentOS, Ubuntu, Debian, and others.

Why Should You Care About Unwanted Services?

When you install Linux, the OS typically enables several services automatically. For example, you might end up with web servers, FTP servers, print servers, or network services running without you asking for them. But if your server doesn't need those, they're just wasting CPU, memory, and opening attack surfaces.

https://www.tecmint.com/remove-unwanted-services-from-linux/

Original Submission

Arthur T Knackerbracket has processed the following story:

Russia's official customs data suggests the country's once-thriving market for US-made processors has nearly disappeared. Figures from the Federal Customs Service (FCS), reported by Russian publication Kommersant, show Intel CPU imports fell by 95 percent last year compared to the previous year. By comparison, AMD shipments dropped by 81 percent. That amounts to just 37,000 CPUs total – a steep decline from 537,000 units in 2023.

Executives in Russia's tech manufacturing sector paint a different picture. Leaders at major domestic assemblers like Lotos Group and Rikor told Kommersant that processor deliveries are not only continuing but increasing. Rikor reports purchasing over 120,000 processors last year – about 30 percent more than the year before. Many Russian tech firms also say chip supplies have improved for the third consecutive year.

Sanctions enforcement is struggling to keep up with a growing number of workarounds. Hong Kong remains a key hub in this network, with one address reportedly managing billions of dollars in smuggled semiconductors. Meanwhile, other chips enter Russia through countries like Malaysia and India, often relabeled or bundled within broader product categories that conceal their true nature from customs officials.

Industry insiders say many processors arrive without being labeled as such. A Russian tech executive told Kommersant that the word "processor" often doesn't appear on delivery sheets. This practice helps explain why the Federal Customs Service's import numbers look so anemic, even though factory shelves remain well-stocked.

It's not all smooth sailing, however. Suppliers warn Russian buyers to expect a 10 to 12 percent price increase in 2025, citing inflation and ongoing tensions in US-China trade relations as key factors. Still, prices for mainstream processors have remained relatively stable for the time being.

Original Submission

Arthur T Knackerbracket has processed the following story:

A team of scientists has unveiled a breakthrough that could one day propel computers to operate at speeds millions of times faster than today's most advanced processors.

The discovery, led by researchers at the University of Arizona and their international collaborators, centers on harnessing ultrafast pulses of light to control the movement of electrons in graphene – a material just one atom thick.

The research, recently published in Nature Communications, demonstrates that electrons can be made to bypass barriers almost instantaneously by firing laser pulses lasting less than a trillionth of a second at graphene. This phenomenon, known as quantum tunneling, has long intrigued physicists, but the team's ability to observe and manipulate it in real time marks a significant milestone.

Mohammed Hassan, an associate professor of physics and optical sciences at the University of Arizona, explained that this advance could usher in processing speeds in the petahertz range – over a thousand times faster than the chips powering today's computers. Such a leap, he said, would transform the landscape of computing, enabling dramatic progress in fields ranging from artificial intelligence and space research to chemistry and health care.

Hassan, who previously led the development of the world's fastest electron microscope, worked alongside colleagues from the University of Arizona, the California Institute of Technology's Jet Propulsion Laboratory, and Ludwig Maximilian University of Munich. Their initial focus was studying how graphene conducts electricity when exposed to laser light. Typically, the symmetrical structure of graphene causes the currents generated on either side to cancel each other out, resulting in no net current.

However, the team made a surprising discovery after modifying the graphene samples. They observed that a single electron could "tunnel" through the material – and that this fleeting event could be captured in real time. This unexpected result prompted further investigation and ultimately led to the creation of what Hassan calls "the world's fastest petahertz quantum transistor."

hubie writes:

Prediction: General-purpose AI could start getting worse:

Opinion: I use AI a lot, but not to write stories. I use AI for search. When it comes to search, AI, especially Perplexity, is simply better than Google.

Ordinary search has gone to the dogs. Maybe as Google goes gaga for AI, its search engine will get better again, but I doubt it. In just the last few months, I've noticed that AI-enabled search, too, has been getting crappier.

In particular, I'm finding that when I search for hard data such as market-share statistics or other business numbers, the results often come from bad sources. Instead of stats from 10-Ks, the US Securities and Exchange Commission's (SEC) mandated annual business financial reports for public companies, I get numbers from sites purporting to be summaries of business reports. These bear some resemblance to reality, but they're never quite right. If I specify I want only 10-K results, it works. If I just ask for financial results, the answers get... interesting,

This isn't just Perplexity. I've done the exact same searches on all the major AI search bots, and they all give me "questionable" results.

Welcome to Garbage In/Garbage Out (GIGO). Formally, in AI circles, this is known as AI model collapse. In an AI model collapse, AI systems, which are trained on their own outputs, gradually lose accuracy, diversity, and reliability. This occurs because errors compound across successive model generations, leading to distorted data distributions and "irreversible defects" in performance. The final result? A Nature 2024 paper stated, "The model becomes poisoned with its own projection of reality."

Model collapse is the result of three different factors. The first is error accumulation, in which each model generation inherits and amplifies flaws from previous versions, causing outputs to drift from original data patterns. Next, there is the loss of tail data: In this, rare events are erased from training data, and eventually, entire concepts are blurred. Finally, feedback loops reinforce narrow patterns, creating repetitive text or biased recommendations.

I like how the AI company Aquant puts it: "In simpler terms, when AI is trained on its own outputs, the results can drift further away from reality."

I'm not the only one seeing AI results starting to go downhill. In a recent Bloomberg Research study of Retrieval-Augmented Generation (RAG), the financial media giant found that 11 leading LLMs, including GPT-4o, Claude-3.5-Sonnet, and Llama-3-8 B, using over 5,000 harmful prompts would produce bad results.

Arthur T Knackerbracket has processed the following story:

Just when you thought you knew all the worlds in the solar system, astronomers go and discover a new object that could rewrite the space map. 

This icy world, temporarily named 2017 OF201, could be a distant cousin of Pluto — and scientists mean "distant" quite literally. At its farthest point, it's more than 1,600 times the distance of Earth from the sun. At its closest, it's still 44.5 times farther than Earth.

What makes 2017 OF201 stand out is its very stretched-out path around the sun, which takes an incredible 25,000 Earth-years to complete. For comparison, Pluto makes a lap around the sun every 248 Earth-years. 

How this world got to the edge of the solar system is a mystery — perhaps the result of close encounters with a giant planet like Jupiter or Neptune that tossed it out into a wide orbit. Or maybe when it was originally ejected, it ended up in the so-called Oort Cloud before returning. The Oort Cloud is thought to be a sphere of ancient, icy objects surrounding the solar system. NASA says the cloud remains a theory because the comets there have been too faint and distant to be directly observed.

The International Astronomical Union’s Minor Planet Center, which catalogs new moons and other small bodies in the solar system, announced the discovery on May 21. At roughly 435 miles wide, 2017 OF201 could qualify as a dwarf planet, the same designation Pluto has had since its demotion from ninth planet in 2006. 

"Even though advances in telescopes have enabled us to explore distant parts of the universe," said Sihao Cheng, the Institute for Advanced Study researcher who led the discovery, in a statement, "there is still a great deal to discover about our own solar system."

fliptop writes:

Unitree director Wang Qixin says the robotics company used AI and motion capture to train the robots on real fight moves:

Four artificial intelligence-enhanced robots have been put through their paces in a Chinese robot fighting competition, duking it out in kickboxing matches until one was declared the champion.

The World Robot Competition Mecha Fighting Series had four human-controlled robots built by China-based firm Unitree compete in three, two-minute rounds with winners crowned through a points system, according to a May 26 report from the China state-owned outlet the Global Times.

[...] The robots reportedly weighed 35 kilograms and stood 132 centimeters tall. Ahead of the boxing rounds, the pint-sized robots were put through tests to demonstrate a variety of kicks and punches and assist the organizers in refining the rules.

The team with the highest points across the three rounds moves on to fight another opponent. A punch to the head was worth one point, and a kick to the head was worth three. Teams lost five points if their robot fell and 10 points if their robot was down for over eight seconds.

[...] Chen Xiyun, a Unitree team member, said the “robots fight in a human-machine collaborative way,” with the machines pre-taught moves, but ultimately, a person controls the bot’s movements.

Original Submission