Cerebras Systems and G42 have introduced the Condor Galaxy project, a network of nine interconnected supercomputers designed for AI model training with a combined performance of 36 FP16 ExaFLOPs. The first supercomputer, CG-1, located in California, offers 4 ExaFLOPs of FP16 performance and 54 million cores, focusing on Large Language Models and Generative AI without the need for complex distributed programming languages. AnandTech reports: CG-2 and CG-3 will be located in the U.S. and will follow in 2024. The remaining systems will be located across the globe and the total cost of the project will be over $900 million. The CG-1 supercomputer, situated in Santa Clara, California, combines 64 Cerebras CS-2 systems into a single user-friendly AI supercomputer, capable of providing 4 ExaFLOPs of dense, systolic FP16 compute for AI training. Based around Cerebras's 2.6 trillion transistor second-generation wafer scale engine processors, the machine is designed specifically for Large Language Models and Generative AI. It supports up to 600 billion parameter models, with configurations that can be expanded to support up to 100 trillion parameter models. Its 54 million AI-optimized compute cores and massivefabric network bandwidth of 388 Tb/s allow for nearly linear performance scaling from 1 to 64 CS-2 systems, according to Cerebras. The CG-1 supercomputer also offers inherent support for long sequence length training (up to 50,000 tokens) and does not require any complex distributed programming languages, which is common in case of GPU clusters.

This supercomputer is provided as a cloud service by Cerebras and G42 and since it is located in the U.S., Cerebras and G42 assert that it will not be used by hostile states. CG-1 is the first of three 4 FP16 ExaFLOP AI supercomputers (CG-1, CG-2, and CG-3) created by Cerebras and G42 in collaboration and located in the U.S. Once connected, these three AI supercomputers will form a 12 FP16 ExaFLOP, 162 million core distributed AI supercomputer, though it remains to be seen how efficient this network will be. In 2024, G42 and Cerebras plan to launch six additional Condor Galaxy supercomputers across the world, which will increase the total compute power to 36 FP16 ExaFLOPs delivered by 576 CS-2 systems. The Condor Galaxy project aims to democratize AI by offering sophisticated AI compute technology in the cloud. "Delivering 4 exaFLOPs of AI compute at FP16, CG-1 dramatically reduces AI training timelines while eliminating the pain of distributed compute," said Andrew Feldman, CEO of Cerebras Systems. "Many cloud companies have announced massive GPU clusters that cost billions of dollars to build, but that are extremely difficult to use. Distributing a single model over thousands of tiny GPUs takes months of time from dozens of people with rare expertise. CG-1 eliminates this challenge. Setting up a generative AI model takes minutes, not months and can be done by a single person. CG-1 is the first of three 4 ExaFLOP AI supercomputers to be deployed across the U.S. Over the next year, together with G42, we plan to expand this deployment and stand up a staggering 36 exaFLOPs of efficient, purpose-built AI compute."

An anonymous reader quotes a report from Apple Insider: Apple's processor manufacturer TSMC says that it can't find enough skilled workers to open its Arizona facility on time, and mass chip production will have to wait until 2025. The Taiwan Semiconductor Manufacturing Company (TSMC) began work on a first factory in Arizona in 2021. Since then, the plant has seen safety concerns, complaints from TSMC about US taxation, and a claim that US staff don't work hard enough. Most recently, the company announced that it was sending more Taiwanese workers to the US to manage the final stages of making the plant operational. Now according to Nikkei Asia, that move has proven insufficient.

"We are encountering certain challenges, as there is an insufficient amount of skilled workers with the specialized expertise required for equipment installation in a semiconductor-grade facility," said TSMC chair Mark Liu. "Consequently we expect the production schedule of N4 [4-nanometer] process technology to be pushed out to 2025," continued Liu. The news comes alongside TSMC's latest earnings report, which shows that the firm's profits have fallen, though they are expected to recover when the iPhone 15 range launches. TSMC blames the results on a slow economic recover in China, and a downturn in the consumer electronics market.

An anonymous reader quotes a report from Engadget: Fervo Energy says it has achieved a breakthrough in geothermal technology. It carried out a 30-day well test at its site in northern Nevada and says it was able to achieve a "flowrate of 63 liters per second at high temperature that enables 3.5 megawatts of electric production." The company says the test resulted in flow and power output records for an enhanced geothermal system (EGS) and that it was completed without incident. A megawatt can power around 750 homes at once. Fervo is expected to connect its Project Red site to the grid this year. It will be used to power Google data centers and some of the company's other Nevada infrastructure. Google and Fervo signed an agreement in 2021 to develop a "next-generation geothermal power project."

This is the first time an energy company has shown that an EGS can work on a commercial scale, according to Bloomberg. It's been a long road to reach this point, as scientists have been trying to make EGS a reality since the 1970s. [...] Fervo says it's the first company to "successfully drill a horizontal well pair for commercial geothermal production, achieving lateral lengths of 3,250 feet, reaching a temperature of 191C, and proving controlled flow through rigorous tracer testing." The company is hoping to replicate its success at a site in Utah. If Fervo sees similar results there and it successfully implements design upgrades to maximize output, the site is expected to generate enough electricity to power 300,000 homes simultaneously, Latimer said. That's around a quarter of all homes in Utah.

Today, Nissan announced it's adopting Tesla's North American Charging Standard (NACS) in its electric vehicles, following in the footsteps of Ford, GM, Rivian, Mercedes-Benz, Volvo, and Polestar. Ars Technica reports: "Adopting the NACS standard underlines Nissan's commitment to making electric mobility even more accessible as we follow our Ambition 2030 long-term vision of greater electrification," said Jeremie Papin, chairperson of Nissan Americas. "We are happy to provide access to thousands more fast chargers for Nissan EV drivers, adding confidence and convenience when planning long-distance journeys."

This is actually Nissan's second time changing its DC fast-charging plugs. An early pioneer of EVs with the first- and then second-generation Leaf, it chose the CHAdeMO standard for those models, which is popular in Japan but never really caught on elsewhere. But when Nissan built the Ariya crossover as its third-generation EV, it dropped CHAdeMO for CCS, which appeared like it was going to win the charging standard war by dint of having every OEM onboard other than Tesla. CCS may have had the power of numbers in terms of OEMs, but EVs from all those makes are still heavily outnumbered on the road by the sheer mass of Models 3 and Y, and it's hard to argue with the superiority of Tesla's Supercharger network, either in terms of reliability or number of deployed chargers.

Intel has announced ASUS as the company's first partner for its Next Unit of Compute (NUC) mini PC business. From a report: The two companies have entered a non-binding agreement that will see ASUS manufacture, sell and support the 10th- to 13th-generation products in Intel's NUC line. ASUS will also develop future NUC designs. Based on the business' current lineup, ASUS could be developing future NUC mini PCs, DIY kits for mini PCs, DIY kits for laptops, customizable boards, chassis and other assembly elements.

If you'll recall, Intel recently told Engadget that it's ending its "direct investment" in its NUC business and will no longer produce first-party NUC products. It didn't elaborate on its reasoning, but working with partners for a non-essential business will free up resources it could use to concentrate on making chips. Intel previously said its first quarter earnings exceeded expectations, but its revenue was still down 36 percent year-over-year when compared to its results in the same period for 2022. The company also said that it remains cautious in this economy.

An anonymous reader quotes a report from Electrek: Tesla has officially launched its new 'Charge on Solar' feature to let its electric car owners charge their vehicles with sunshine. Tesla describes the feature: "With Charge on Solar, your Tesla vehicle can charge using only excess solar energy produced by your Tesla solar system. Using excess energy to charge your electric vehicle maximizes the value of your home's solar system. Use the Tesla app to set Charge on Solar limits and have your vehicle charge using extra solar energy."

If you have all the required hardware and software, you can go to your Tesla app, and under "Charge on Solar," you will be able to set your charge limit using excess solar energy. Tesla writes about the charge limit: "Your vehicle will charge from solar and the grid when your current charge level is below the left sun slider. After your vehicle's charge level passes the sun slider, your vehicle automatically switches to only charge on excess solar up to your charge limit. Solar power and home loads are variables so if you ever want to charge faster, you can simply increase the lower charge limit to a desired range." There's also a scheduling feature that stops charging with solar after a specific time if you know that your energy consumption will increase at a specific time.

Dale Vince, the green energy tycoon and founder of Ecotricity, is planning to launch Britain's first electric airline called Ecojet. The Guardian reports: Ecojet, styled as a "flag carrier for green Britain," will launch early next year with a 19-seater plane traveling on a route between Edinburgh and Southampton. The planes will run initially on kerosene-based fuel for the first year, before being retrofitted with engines that convert green hydrogen into electricity. The airline will launch with several green-striped 19-seater planes capable of traveling for 300 miles. Vince hopes to expand the number of routes out to cover all of Britain's big cities. Staff will wear environmentally friendly uniforms, and serve plant-based meals.

A second phase, 18 months later, will result in 70-seater planes capable of flying to Europe being introduced. The company is in the process of applying for a license from the Civil Aviation Authority and securing takeoff and landing slots at airports. However, the process of launching an airline is regarded as slow, and Ecojet will not launch as an electric plane operator, starting by using kerosene-based fuel instead. [...] Vince said Ecojet would "price match" existing airlines on air fares and was intended to attract a mass market, beyond environment-conscious consumers. He said he would invest one million pounds initially but plans to raise further funds next year.

Long-time Slashdot reader hackingbear quotes the South China Morning Post: After a 13-year legal tussle, semiconductor equipment giant Advanced Micro-Fabrication Equipment of China (AMEC) has won an intellectual property infringement case against US competitor Lam Research Corp in a Shanghai court, as US-China technology rivalry in the semiconductor field rages on. The Shanghai People's High Court gave a final ruling requiring Lam Research to destroy "one technical document and two photographs" relating to an AMEC plasma etching machine that Lam illegally obtained, according to a statement by AMEC on Tuesday.

The court has also banned two individual defendants from Lam from using AMEC's proprietary trade secrets. The court ordered Lam Research to pay damages and legal fees to AMEC for the infringement.

"There seem to be no limits to the ingenious ways that designers are devising to harvest energy or take existing approaches and exploit and enhance them," writes the site Electronic Design: A research team at the Pacific Northwest National Laboratory (PNNL) has developed a contact-separation mode triboelectric nanogenerator (TENG) with a simple structure for harvesting wave energy and powering marine sensors and transmitters.Although this isn't the first cylindrical TENG (C-TENG) — several models are already in use — the PNNL team maintains that this design overcomes weaknesses of those existing ones.
Long-time Slashdot reader RoccamOccam writes that "triboelectrification is the process by which two originally uncharged bodies become charged when brought into contact and then separated." The key is a new mechanism for wave-driven energy-harvesting "TENGs" that can convert the low-amplitude, low-frequency ocean waves into high-frequency mechanical motion for more effective power generation. This new TENG must be able to operate and be triggered by any wave conditions, even in the middle of the ocean where waves have uniform or random low amplitude and frequency.
The researchers tested their process in a 12-meter-long water tank (with "adjustable wave height and frequency," according to the article). The patent-pending device "sustainably powered up an array of 27 LEDs and was able to charge up a capacitor up to 1.8 V for driving an acoustic transmitter."

In mid-June, the European Parliament voted in favor of new legislation that would, among other things, require batteries in consumer devices like smartphones to be easily removable and replaceable. This week, the European Council officially agreed to the new regulation. Now, when the European Council and Parliament sign on the dotted line, the clock will start ticking for manufacturers to ensure their devices have replaceable batteries by 2027 -- that is, if they want to sell their devices in the EU. Android Authority reports: Now, the only step left is for the European Council and Parliament to sign on the dotted line. Once they do, the clock starts ticking: any manufacturer wanting to sell phones in the EU must ensure those phones have replaceable batteries by 2027. [...] The grace period from now until 2027 is to give OEMs enough time to redesign their products. This new law states, specifically, that users should be able to replace a battery in their phone without any special expertise or tools. Being that almost all smartphones today are designed like a "glass sandwich" that relies on extensive use of adhesives, the very fundamentals of how companies design phones will need to change. It's too early to say yet how this law will change iPhones, Galaxy S phones, Pixels, etc. However, they will change in response to this law, which is huge news.

Here are some other rules this new law covers related to phones with replaceable batteries:

- Collection of waste: OEMs will need to collect 63% of portable batteries that would normally go to a landfill by the end of 2027. By the end of 2030, that number should be at 73%.
- Recovery of waste: Lithium recovery from waste batteries will need to be at 50% by 2027. By the end of 2031, it should be at 80%, meaning 80% of the lithium inside a battery can be recovered and repurposed for new batteries.
- Recycling minimums: Industrial, SLI, and EV batteries will need to be made up of certain percentages of recycled content. Initially, this will be 16% for cobalt, 85% for lead, 6% for lithium, and 6% for nickel.
- Early recycling efficiency target: Nickel-cadmium batteries should have a recycling efficiency target of 80% by the end of 2025. All other batteries should be at a 50% efficiency target by 2025.

Chipotle has developed a robot that can cut the 50-minute process of making guacamole in half. "The fast-casual chain developed the collaborative robot, or cobot, in partnership with Vebu Labs, a California-based robotics startup," reports CNBC. "Chipotle also announced Wednesday that its $50 million venture arm, Cultivate Next, is investing in Vebu. Financial terms weren't disclosed." From the report: To prepare avocados using the Autocado, Chipotle employees load up the device with a full case of the ripe fruit. The Autocado can hold up to 25 pounds at one time. Then, the machine vertically orients the avocados, slices them in half and removes their cores and skin. A bowl at the bottom collects the fruit, which employees can then hand mash and mix with the rest of the guacamole ingredients.

Chipotle still wants employees to have a hand in making their guacamole. "There's no plan to test automated guac made in our restaurant," Curt Garner, Chipotle's chief technology officer, told CNBC. Employees don't have to monitor the Autocado while it prepares the avocados and can even use the top of the device as more counter space to prepare other ingredients. The prototype is "very close" to design for manufacture, according to Garner. Chipotle expects to test the Autocado in restaurants later this year.

Eventually, Vebu plans to add machine learning capabilities and sensors to the Autocado that will help it evaluate the quality of avocados. Preparing avocados for guacamole routinely ranks as one of employees' least favorite tasks, Garner said. It's also one of the most dangerous duties in Chipotle kitchens, sometimes resulting in knife injuries. On top of saving time and labor costs, the robot could also cut food waste. If the chain deploys the Autocado across its footprint of more than 3,200 locations, it could help save millions of dollars on avocados annually, the company said. Despite those savings, guacamole will probably still cost customers extra. "It's worth it," Garner said.

Intel has decided to stop making its Next Unit of Computing (NUC), but the company will encourage partners to keep making the small form-factor (SFF) PCs, the company said Tuesday. From a report: Intel's NUC championed compact PCs, while leaving larger chassis options to partners like Dell and HP. But Intel's decision seems like a natural one, given that Intel has refocused on its core businesses during a period in which it also invested heavily in its own manufacturing operations and foundry business.

An Intel spokesman confirmed an initial report by Serve The Home, saying that Intel will continue to support the existing NUCs it has already shipped into the market. "We have decided to stop direct investment in the Next Unit of Compute (NUC) Business and pivot our strategy to enable our ecosystem partners to continue NUC innovation and growth," the Intel spokesman said in an email.

An anonymous reader quotes a report from the Washington Post: If you're printing something on actual paper, there's a good chance it's important, like a tax form or a job contract. But popular printing products and services won't promise not to read it. In fact, they won't even promise not to share it with outside marketing firms. The spread of digital file-sharing -- along with obnoxious business practices by printing manufacturers -- has pushed many U.S. households to give up at-home printers and rely on nearby printing services instead. At the same time, major printer manufacturers have adopted mobile apps and cloud-based storage, creating new opportunities to collect personal data from customers. Whether you're walking to the corner store or sending your files to the cloud, it's tough to figure out whether you're printing in private.

Ideally, printing services should avoid storing the content of your files, or at least delete daily. Print services should also communicate clearly upfront what information they're collecting and why. Some services, like the New York Public Library and PrintWithMe, do both. Others dodged our questions about what data they collect, how long they store it and whom they share it with. Some -- including Canon, FedEx and Staples -- declined to answer basic questions about their privacy practices. Wondering whether your printer app or printing service stores the content of your documents? Here's The Washington Post Help Desk's at-a-glance guide to printer privacy. Here's a summary of each company's privacy policy as it pertains to storing the content of your files:

HP: HP's privacy policy states that it does not store the content of files when using their printers or HP Smart app, providing reassurance that they do not invade privacy by snooping into print jobs.
Canon: Canon's privacy policy indicates that it can collect personal data, including files and content, which may be used for marketing purposes. However, Canon did not disclose whether they store, use, or share the content of printed documents.
FedEx: FedEx's privacy policy states that it collects user-uploaded information, including the contents of documents uploaded for printing services, leaving room for potential advertising or sharing with third parties. Although FedEx prioritizes customer privacy, it did not specify the extent of encryption or whether document content is included.
UPS: While the UPS Store, a subsidiary of UPS, can store the contents of printed documents, it does not use this information for marketing or advertising without user consent. The storage duration is undisclosed, but UPS honors customer requests for data deletion.
Staples: According to Staples' privacy policy, the company can store personal data such as copy/print materials, driver's license numbers, passport numbers, and mail contents. They may also use copy/print materials for advertising. The duration of data storage is not disclosed.
PrintWithMe: PrintWithMe, a company placing printers in shared spaces, temporarily stores printed documents with a third-party cloud provider for 24 hours. CEO Jonathan Treble assures that the data is never used for advertising.
Your local library: The New York Public Library, one of the largest library systems, does not store the contents of printed documents. Their computers only retain file names and delete them at the end of the day. However, privacy policies may vary among different libraries, so it is advisable to inquire beforehand.

Long-time Slashdot reader beforewisdom shared this report from the Independent: South Koreans have begun to hoard excessive amounts of sea salt and other items as Japan prepares to dump treated radioactive water from the Fukushima power plant into the ocean... Tokyo has repeatedly assured that the water is safe and has been filtered to remove most isotopes though it does contain traces of tritium, an isotope of hydrogen hard to separate from water.

Although Japan has not set a date for the release, the announcement has made fishermen and shoppers across the region apprehensive. South Korea's fisheries authorities have vowed to ramp up efforts to monitor natural salt farms for any rise in radioactive substances and maintain a ban on seafood from the waters near Fukushima... The panic buying has led to a 27 per cent rise in the price of salt in South Korea in June from two months ago, though officials say the weather and lower production were also to blame. The Korean government in response has decided to release about 50 metric tons of salt a day from stocks, at a 20 per cent discount from market prices, until 11 July...

More than 85 per cent of the South Korean public oppose Japan's plan, according to a survey last month by local pollster Research View. Seven in 10 people reportedly said that they would consume less seafood if the waste water release goes ahead.

"The obsession with EV range is all wrong," argues a new article in the Washington Post's Climate section. "This year, one EV on the market — the sleek $140,000 Lucid Air Grand Touring — boasts a whopping 516-mile range. Toyota recently announced that it had achieved a breakthrough with solid-state battery technology, saying it will soon be able to produce electric cars that can go 746 miles on a single charge.

"But some analysts say that all that range — and all that battery — misses the point, and wastes resources." Only 5% of trips in the U.S. are longer than 30 miles. The vast majority of big batteries will never be used — particularly if the owner has a place to plug in their car every day... Those batteries are massive, in every sense of the word: the battery on the electric F-150 Lightning, which allows the car to go more than 300 miles on a single charge, weighs a whopping 1,800 pounds.

But is all that necessary? Americans drive a lot, but most of our trips are not very long. According to data from the U.S. Department of Transportation, 95.1% of trips taken in personal vehicles are less than 31 miles; almost 60% of all trips are less than 6 miles. In total, the average U.S. driver only covers about 37 miles per day. And there is evidence that much smaller batteries could do the lion's share of the work. In a study published in 2016, researchers at MIT found that a car with a 73-mile range (like an early version of the Nissan Leaf), charged only at night, could satisfy 87% of all driving days in the United States. Providing Nissan Leafs to everyone whose driving fit that pattern, the researchers found, would cut 61% of U.S. gasoline consumption by personal vehicles...

So most of the time, drivers are lugging around giant batteries but only using 10 to 15% of their actual power. And those big batteries require mining a lot of metals, damaging the environment and workers' health... In a report by researchers at the University of California at Davis, the Climate and Community Project, and Providence College, experts found that simply switching to smaller EV batteries — batteries that could give a small car a range of 125 miles or so — could cut lithium demand by 42%...


The article notes that the upcoming Dodge Ram 1500 REV, with a range of about 500 miles, will need a battery "roughly equivalent in terms of resources to 16 batteries for the Prius Prime plug-in hybrid..."

"For those who need to take frequent long road trips and don't want to have to plug in, a plug-in hybrid can be a good option. But for most Americans, an EV with medium range will do just fine."

We've just had the world's first press conference with AI-enabled, humanoid social robots. Click here to jump straight to Slashdot's transcript of all the robots' answers during the press conference, or watch the 40-minute video here.

It all happened as the United Nations held an "AI for Good" summit in Geneva, where the Guardian reports that the foyer was "humming with robotic voices, the whirring of automated wheels and limbs, and Desdemona, the 'rock star' humanoid, who is chanting 'the singularity will not be centralised' on stage backed by a human band, Jam Galaxy."

But the Associated Press describes how one UN agency had "assembled a group of robots that physically resembled humans at a news conference Friday, inviting reporters to ask them questions in an event meant to spark discussion about the future of artificial intelligence. "The nine robots were seated and posed upright along with some of the people who helped make them at a podium in a Geneva conference center... Among them: Sophia, the first robot innovation ambassador for the U.N. Development Program, or UNDP; Grace, described as a health care robot; and Desdemona, a rock star robot."

"I'm terrified by all of this," said one local newscaster, noting that the robots also said they "had no intention of rebelling against their creators."

But the Associated Press points out an important caveat: While the robots vocalized strong statements - that robots could be more efficient leaders than humans, but wouldn't take anyone's job away or stage a rebellion - organizers didn't specify to what extent the answers were scripted or programmed by people. The summit was meant to showcase "human-machine collaboration," and some of the robots are capable of producing preprogrammed responses, according to their documentation.
Two of the robots seemed to disagree on whether AI-powered robots should submit to stricter regulation. (Although since they're only synthesizing sentences from large-language models, can they really be said to "agree" or "disagree"?)

There were unintentionally humorous moments, starting right from the beginning. Click here to start reading Slashdot's transcript of the robots' answers:

Remember those researchers who generated electricity from the energy in air humidity?

"To be frank, it was an accident," the study's lead author, Prof Jun Yao, tells the Guardian: "We were actually interested in making a simple sensor for humidity in the air. But for whatever reason, the student who was working on that forgot to plug in the power." The UMass Amherst team were surprised to find that the device, which comprised an array of microscopic tubes, or nanowires, was producing an electrical signal regardless. Each nanowire was less than one-thousandth the diameter of a human hair, wide enough that an airborne water molecule could enter, but so narrow it would bump around inside the tube. Each bump, the team realised, lent the material a small charge, and as the frequency of bumps increased, one end of the tube became differently charged from the other. "So it's really like a battery," says Yao. "You have a positive pull and a negative pull, and when you connect them the charge is going to flow..."

"The beauty is that the air is everywhere," says Yao. "Even though a thin sheet of the device gives out a very tiny amount of electricity or power, in principle, we can stack multiple layers in vertical space to increase the power." That's exactly what another team, Prof Svitlana Lyubchyk and her twin sons, Profs Andriy and Sergiy Lyubchyk, are trying to do. Svitlana Lyubchyk and Andriy are part of the Lisbon-based Catcher project, whose aim is "changing atmospheric humidity into renewable power", and along with Sergiy they have founded CascataChuva, a startup intended to commercialise the research... Catcher and related projects [received] nearly €5.5m (£4.7m) in funding from the European Innovation Council. The result is a thin grey disc measuring 4cm (1.5in) across. According to the Lyubchyks, one of these devices can generate a relatively modest 1.5 volts and 10 milliamps. However, 20,000 of them stacked into a washing machine-sized cube, they say, could generate 10 kilowatt hours of energy a day — roughly the consumption of an average UK household. Even more impressive: they plan to have a prototype ready for demonstration in 2024...

The Lyubchyks estimate that the levelised cost of energy — the average net present cost of electricity generation for a generator over its lifetime — from these devices will indeed be high at first, but by moving into mass production, they hope to lower it significantly, ultimately making this hygroelectric power competitive with solar and wind... The team accept that it may take years to optimise a prototype and scale up production, but if they're successful, the benefits are clear. Unlike solar or wind, hygroelectric generators could work day and night, indoors and out, and in many places.
Yao explains to the Guardian that "Lots of energy is stored in water molecules in the air. That's where we get the lightning effect during a thunderstorm.

"The existence of this type of energy isn't in doubt. It's about how we collect it."

Thanks to Slashdot reader j3x0n for sharing the article.

Mercedes-Benz is the latest automaker to commit to adopting Tesla's electric vehicle charging connectors for its vehicles. The Verge reports: Starting in 2024, the German automaker will offer adapters to its customers so they can access Tesla Supercharger stations, which use the company's North American Charging Standard (NACS) plug and outlet. And in 2025, Mercedes will begin to manufacture EVs with Tesla's charging port, obviating the need for an adapter. The company joins Ford, GM, Volvo, Polestar, and Rivian in adopting Tesla's connector, which is rapidly becoming the de facto charging standard in North America and Europe. Mercedes is also the first German automaker to jump on the Tesla bandwagon, putting pressure on Volkswagen and BMW to follow suit. VW, along with Hyundai, Kia, and Stellantis, has previously confirmed it is in talks with Musk's company about adopting NACS.

Mercedes also announced it plans on expanding its network of EV charging stations in North America to 400 hubs with more than 2,500 "high-power chargers," which will also grow to include Tesla's NACS connector. The company plans on installing 2,000 hubs with 10,000 chargers worldwide. Mercedes announced earlier this year that it was working with ChargePoint and MN8, a solar company, to install chargers in key cities and along major highways. The first stations, which will be accessible to Mercedes and non-Mercedes EV owners, are scheduled to open in the fourth quarter of 2023.

Apple plans to launch its Vision Pro mixed-reality headset in select US markets early next year, with designated sections in Apple stores offering demos, seating, and tools for sizing accessories. Bloomberg reports: The company will designate special areas in the stores with seating, headset demo units and tools to size accessories for buyers. While the device will be sold at all of Apple's roughly 270 US locations, the company is planning the sections for the Vision Pro initially at stores in major areas -- such as New York and Los Angeles -- before rolling them out nationwide, according to people with knowledge of the plans.

Apple said it will offer the headset in other countries at the end of 2024. The company is discussing the UK and Canada as two of its first international markets with Asia and Europe soon after, although a final decision hasn't been made, according to the people, who asked not to be named discussing internal matters. Apple engineers are working to localize the device for France, Germany, Australia, China, Hong Kong, Japan and Korea, the people said. Apple will also sell the Vision Pro through its US web store in early 2024 before expanding online elsewhere.

Inflection AI, an AI startup company, has built a cutting-edge supercomputer equipped with 22,000 NVIDIA H100 GPUs. Wccftech reports: For those unfamiliar with Inflection AI, it is a business that aims at creating "personal AI for everyone." The company is widely known for its recently introduced Inflection-1 AI model, which powers the Pi chatbot. Although the AI model hasn't yet reached the level of ChatGPT or Google's LaMDA models, reports suggest that Inflection-1 performs well on "common sense" tasks, making it much more suitable for applications such as personal assistance.
>
Coming back, Inflection announced that it is building one of the world's largest AI-based supercomputers, and it looks like we finally have a glimpse of what it would be. It is reported that the Inflection supercomputer is equipped with 22,000 H100 GPUs, and based on analysis, it would contain almost 700 four-node racks of Intel Xeon CPUs. The supercomputer will utilize an astounding 31 Mega-Watts of power.

The surprising fact about the supercomputer is the acquisition of 22,000 NVIDIA H100 GPUs. We all are well aware that, in recent times, it has been challenging to acquire even a single unit of the H100s since they are in immense demand, and NVIDIA cannot cope with the influx of orders. In the case of Inflection AI, NVIDIA is considering being an investor in the company, which is why in their case, it is easier to get their hands on such a massive number of GPUs.