Author name: Paul Patrick

the-future-of-arrakis-is-at-stake-in-latest-trailer-for-dune:-part-two

The future of Arrakis is at stake in latest trailer for Dune: Part Two

culture, denis villeneuve, Dune: part 2, Entertainment, film trailers, Frank Herbert, Timothée Chalamet, Trailers, Warner Bros. / /

“You are not prepared for what is to come” —

“This is a form of power that our world has not yet seen.”

–

Dune: Part Two is the next chapter in director Denis Villeneuve’s adaptation of Frank Herbert’s celebrated novel.

We didn’t get to see Dune: Part Two—the second film in director Denis Villeneuve’s stunning adaptation of Frank Herbert’s sci-fi classic—last month as originally planned since the film’s November release was delayed until next March due to the Hollywood strikes. But Warner Bros. doesn’t want us to completely forget about Dune in the meantime, so it dropped another trailer for the holiday season.

(Spoilers for Dune: Part One below.)

As reported previously (also here and here), Herbert’s novel Dune is set in the distant future and follows the fortunes of various noble houses in what amounts to a feudal interstellar society. Much of the action takes place on the planet Arrakis, where the economy is driven largely by a rare, life-extending drug called melange (“the spice”). Melange also conveys a kind of prescience and makes faster-than-light travel practical. There’s betrayal, a prophecy concerning a messianic figure, giant sandworms, and battle upon battle as protagonist Paul Atreides (Timothée Chalamet) contends with rival House Harkonnen and strives to defeat the forces of Shaddam IV, Emperor of the Known Universe.

Part One‘s finale left Paul and his mother, Lady Jessica (Rebecca Ferguson), presumed dead in the harsh desert of Arrakis, having fled their home when Baron Vladimir Harkonnen (Stellan Skarsgård) betrayed the Atreides family and killed Paul’s father, Leto (Oscar Isaac). They were taken in by the Fremen, the planet’s native inhabitants, who include Chani (Zendaya), a girl appearing in Paul’s dreams/visions.

All the surviving principles from Part 1 reprise their roles in Part 2: Chalamet, Zendaya, Ferguson, Skarsgård, Javier Bardem as Stilgar, Josh Brolin as Gurney Halleck, Dave Bautista as Glossu Rabban Harkonnen, Charlotte Rampling as the Reverend Mother Mohiam, and Stephen McKinley Henderson as Thufir Hawat. New cast members include Christopher Walken as Shaddam IV, emperor of House Corrine; Florence Pugh as his daughter, Princess Irulan; Austin Butler as Harkonnen’s younger nephew, Feyd-Rautha, the presumed heir on Arrakis; Lea Seydoux as Lady Margot, a Bene Gesserit who is close with the Emperor; and Souheila Yacoub as a Fremen warrior named Shishakli.

  • Love blooms between Paul and Chani in the midst of pending war.

    YouTube/Warner Bros.

  • Paul is having recurrent nightmares.

    YouTube/Warner Bros.

  • Christopher Walken plays Shaddam IV, Padishah Emperor of the Known Universe and head of House Corrino.

    YouTube/Warner Bros.

  • Feyd-Rautha Harkonnen does love his knives.

    YouTube/Warner Bros.

  • Florence Pugh plays the Emperor’s daughter, Princess Irulan.

    YouTube/Warner Bros.

  • Lady Jessica (Rebecca Ferguson), Paul’s mother.

    YouTube/Warner Bros.

  • “Silence!” Paul is starting to come into his power.

    YouTube/Warner Bros.

  • Beware of sandworms!

    YouTube/Warner Bros.

The first trailer dropped in May after being unveiled in an exclusive sneak peek during CinemaCon in Las Vegas. The highlight was a sequence showing Paul’s first ride on a sandworm. It’s a major rite of passage in Fremen culture, and the scene demonstrates that, in Part 2, Paul is well on his way to becoming Muad’Dib, prophet of the Fremen. A second trailer arrived in June, showing Paul offering to fight with the Fremen against their common enemy, though not everyone welcomes his inclusion. We also saw a reunion with Halleck; Shaddam IV learning that Paul is still alive; Feyd-Rautha’s lethal knife-fighting skills; and love blooming between Paul and Chani.

That love story is a major focus of this latest trailer after two that mostly highlighted the war for the future of Arrakis. The trailer opens with Paul having one of his recurring nightmares and Chani comforting him. He can only remember fragments but later tells Chani that he sees “possible futures all at once. And in so many futures, our enemies prevail.” He said, “There is a narrow way through.” Meanwhile, the Emperor orders assassins to “deal with this prophet.” One person who might get the job done is Feyd-Rautha, described as psychotic as we see him staring someone down while licking a sharp curved blade and brutally stabbing an opponent in an arena while a crowd cheers wildly.

There’s a fantastic battle scene involving Fremen warriors riding sandworms, and we catch a glimpse of the darker side of Paul when he screams “Silence!” after Mother Mohian asks him to carefully consider his planned course of action. Despite the war, he vows to love Chani “as long as I breathe.” She claims he will never lose her “as long as you stay who you are.” But fans of the books know that the romance has its complications, and given one new cast member in particular, we can expect to see the beginnings of those complications.

Dune: Part Two hits theaters on March 1, 2024.

Listing image by YouTube/Warner Bros.

The future of Arrakis is at stake in latest trailer for Dune: Part Two Read More »

broadcom-ends-vmware-perpetual-license-sales,-testing-customers-and-partners

Broadcom ends VMware perpetual license sales, testing customers and partners

Biz & IT, Broadcom, Software, Tech, virtualization, vmware / /

saas —

Already-purchased licenses can still be used but will eventually lose support.

–

The logo of American cloud computing and virtualization technology company VMware is seen at the Mobile World Congress (MWC), the telecom industry's biggest annual gathering, in Barcelona on March 2, 2023.

Broadcom has moved forward with plans to transition VMware, a virtualization and cloud computing company, into a subscription-based business. As of December 11, it no longer sells perpetual licenses with VMware products. VMware, whose $61 billion acquisition by Broadcom closed in November, also announced on Monday that it will no longer sell support and subscription (SnS) for VMware products with perpetual licenses. Moving forward, VMware will only offer term licenses or subscriptions, according to its VMware blog post.

VMware customers with perpetual licenses and active support contracts can continue using them. VMware “will continue to provide support as defined in contractual commitments,” Krish Prasad, senior vice president and general manager for VMware’s Cloud Foundation Division, wrote. But when customers’ SnS terms end, they won’t have any support.

Broadcom hopes this will force customers into subscriptions, and it’s offering “upgrade pricing incentives” that weren’t detailed in the blog for customers who switch from perpetual licensing to a subscription.

These are the products affected, per Prasad’s blog:

  • VMware Aria Automation
  • VMware Aria Suite
  • VMware Aria Operations
  • VMware Aria Operations for Logs
  • VMware Aria Operations for Networks
  • VMware Aria Universal
  • VMware Cloud Foundation
  • VMware HCX
  • VMware NSX
  • VMware Site Recovery Manager
  • VMware vCloud Suite
  • VMware vSAN
  • VMware vSphere

Subscription-based future

Broadcom is looking to grow VMware’s EBITDA (earnings before interest, taxes, depreciation, and amortization) from about $4.7 billion to about $8.5 billion in three years, largely through shifting the company’s business model to subscriptions, Tom Krause, president of the Broadcom Software Group, said during a December 7 earnings call, per Forbes.

“This shift is the natural next step in our multi-year strategy to make it easier for customers to consume both our existing offerings and new innovations. VMware believes that a subscription model supports our customers with the innovation and flexibility they need as they undertake their digital transformations,” VMware’s blog said.

With changes effective immediately upon announcement, the news might sound abrupt. However, in May, soon after announcing its plans to acquire VMware, Broadcom CEO Hock Tan signaled a “rapid transition” to subscriptions.

At the time, Tan pointed to the importance of maintaining current VMware customers’ happiness, as well as leveraging the VMware sales team already in place. However, after less than a month of the deal’s close, reports point to concern among VMWare customers and partners.

Customer and partner concerns

VMware’s blog said “the industry has already embraced subscription as the standard for cloud consumption.” For years, software and even hardware vendors and investors have been pushing IT solution provider partners and customers toward recurring revenue models. However, VMware built much of its business on the perpetual license model. As noted by The Stack, VMware in February noted that perpetual licensing was the company’s “most renowned model.”

VMware’s blog this week listed “continuous innovation” and “faster time to value” as customer benefits for subscription models but didn’t detail how it came to those conclusions.

“Predictable investments” is also listed, but it’s hard to imagine a more predictable expense than paying for something once and having supported access to it indefinitely (assuming you continue paying any support costs). Now, VMware and its partners will be left convincing customers that their finances can afford a new monthly expense for something they thought was paid for. For Broadcom, though, it’s easier to see the benefits of turning VMware into more of a reliable and recurring revenue stream.

Additionally, Broadcom’s layoffs of at least 2,837 VMware employees have brought uncertainty to the VMware brand. A CRN report in late November pointed to VMware partners hearing customer concern about potential price raises and a lack of support. C.R. Howdyshell, CEO of Advizex, which reportedly made $30 million in VMware-tied revenue in 2022, told the publication that partners and customers were experiencing “significant concern and chaos” around VMware sales. Another channel partner noted to CRN the layoff of a close VMware sales contact.

But Broadcom has made it clear that it wants to “complete the transition of all VMware by Broadcom solutions to subscription licenses,” per Prasad’s blog.

The company hopes to convince skeptical channel partners that they’ll see the way, too. VMware, like many tech companies urging subscription models, pointed to “many partners” having success with subscription models already and “opportunity for partners to engage more strategically with customers and deliver higher-value services that drive customer success.”

However, because there’s no immediate customer benefit to the end of perpetual licenses, those impacted by VMware’s change in business strategy have to assess how much they’re willing to pay to access VMware products moving forward.

Broadcom ends VMware perpetual license sales, testing customers and partners Read More »

ted-cruz-wants-to-stop-the-fcc-from-updating-data-breach-notification-rules

Ted Cruz wants to stop the FCC from updating data-breach notification rules

data breach, federal communications commission, Policy, ted cruz / /
Sen. Ted Cruz speaks at a Senate committee hearing while holding up three fingers.

Enlarge / Sen. Ted Cruz (R-Texas) at a Senate Judiciary Committee hearing on Thursday, November 30, 2023.

Getty Images | Bill Clark

Sen. Ted Cruz (R-Texas) and other Republican senators are fighting a Federal Communications Commission plan to impose new data-breach notification requirements on telecom providers. In a letter sent to FCC Chairwoman Jessica Rosenworcel today, the senators claim the pending FCC action would violate a congressional order.

The letter was sent by Cruz, Sen. Minority Leader Mitch McConnell (R-Ky.), Sen. John Thune (R-S.D.), and Sen. Marsha Blackburn (R-Tenn.). They say the proposed data-breach notification rules are preempted by an action Congress took in 2017 to kill an assortment of privacy and security rules issued by the FCC.

The Congressional Review Act (CRA) was used in 2017 by Congress and then-President Donald Trump to throw out rules that would have required home Internet and mobile broadband providers to get consumers’ opt-in consent before using, sharing, or selling Web browsing history, app usage history, and other private information.

The invalidated FCC rules also included data-breach notification requirements that are similar to those the current FCC now plans to impose. The FCC already enforces data-breach notification requirements, but the pending proposal would expand the scope of those rules.

Rosenworcel’s data-breach proposal is scheduled for a vote at tomorrow’s commission meeting, and it may ultimately be up to the courts to decide whether it violates the 2017 congressional resolution. The Republican senators urged the FCC to rescind the draft plan and remove it from the meeting agenda.

Cruz also protested a recent FCC vote to enforce rules that prohibit discrimination in access to broadband services, calling it “government-mandated affirmative action and race-based pricing.”

Republicans: FCC plan “clearly unlawful”

When an agency-issued rule is nullified by a Congressional Review Act resolution, that rule “may not be reissued in substantially the same form” without authorization from Congress. The key legal question seems to be whether the FCC can re-implement one portion of the nullified rules as long as it doesn’t bring back the entire privacy order.

Cruz and fellow Republicans say that Rosenworcel’s plan would “resurrect a portion of the 2016 Broadband Privacy Order pertaining to data security.”

“This is clearly unlawful: the FCC’s proposed rules in the Report and Order are clearly ‘substantially similar’ to the nullified 2016 rules,” they wrote. “Specifically, the requirements in the Report and Order governing notification to the FCC, law enforcement, and consumers, as well as the recordkeeping requirements with respect to breaches and notifications, are substantially similar to the notification and recordkeeping requirements disapproved by Congress.”

The FCC proposal anticipates this argument but says the agency believes it can re-implement part of the Obama-era privacy order:

We conclude that it would be erroneous to construe the resolution of disapproval as applying to anything other than all of the rule revisions, as a whole, adopted as part of the 2016 Privacy Order. That resolution had the effect of nullifying each and every provision of the 2016 Privacy Order—each part being, under the APA [Administrative Procedure Act], “a rule”—but not “the rule” specified in the resolution of disapproval. By its terms, the CRA does not prohibit the adoption of a rule that is merely substantially similar to a limited portion of the disapproved rule or one that is the same as individual pieces of the disapproved rule.

Thus, according to the FCC proposal, the resolution “does not prohibit the Commission from revising its breach notification rules in ways that are similar to, or even the same as, some of the revisions that were adopted in the 2016 Privacy Order, unless the revisions adopted are the same, in substance, as the 2016 Privacy Order as a whole.”

Ted Cruz wants to stop the FCC from updating data-breach notification rules Read More »

a-new-essential-guide-to-electronics-by-naomi-wu-details-a-different-shenzen

A New Essential Guide to Electronics by Naomi Wu details a different Shenzen

andrew "bunnie" huang, andrew huang, Biz & IT, bunnie huang, china, electronics manufacturing, manufacturing, naomi wu, shenzen, Tech / /

Crystal clear, super-bright, and short leads —

Eating, tipping, LGBTQ+ advice, and Mandarin for “Self-Flashing” and “RGB.”

–

Point to translate guide in the New Essential Guide to Electronics in Shenzen

Enlarge / The New Essential Guide to Electronics in Shenzen is made to be pointed at, rapidly, in a crowded environment.

Machinery Enchantress / Crowd Supply

“Hong Kong has better food, Shanghai has better nightlife. But when it comes to making things—no one can beat Shenzen.”

Many things about the Hua Qiang market in Shenzen, China, are different than they were in 2016, when Andrew “bunnie” Huang’s Essential Guide to Electronics in Shenzen was first published. But the importance of the world’s premiere electronics market, and the need for help navigating it, are a constant. That’s why the book is getting an authorized, crowdfunded revision, the New Essential Guide, written by noted maker and Shenzen native Naomi Wu and due to ship in April 2024.

Naomi Wu’s narrated introduction to the New Essential Guide to Electronics in Shenzen.

Huang notes on the crowdfunding page that Wu’s “strengths round out my weaknesses.” Wu speaks Mandarin, lives in Shenzen, and is more familiar with Shenzen, and China, as it is today. Shenzen has grown by more than 2 million people, the central Huaqiangbei Road has been replaced by a car-free boulevard, and the city’s metro system has more than 100 new kilometers with dozens of new stations. As happens anywhere, market vendors have also changed locations, payment and communications systems have modernized, and customs have shifted.

The updated guide’s contents are set to include typical visitor guide items, like “Taxis,” “Tipping,” and, new to this edition, “LGBTQ+ Visitors.” Then there are the more Shenzen-specific guides: “Is It Fake?,” “Do Not Burn Your Contacts,” and “Type It, Don’t Say It.” The original guide had plastic business card pockets, but “They are anachronistic now,” Wu writes; removing them has allowed the 2023 guide to be sold for the same price as the original.

Machinery Enchantress / Crowd Supply

Both the original and updated guide are ring-bound and focus on quick-flipping and “Point to Translate” guides, with clearly defined boxes of English and Mandarin characters for things like “RGB,” “Common anode,” and “LED tape.” “When sourcing components, speed is critical, and it’s quicker to flip through physical pages,” Wu writes. “The market is full of visitors struggling to navigate mobile interfaces in order to make their needs known to busy vendors. It simply doesn’t work as well as walking up and pointing to large, clearly written Chinese of exactly what you want.”

Then there is the other notable thing that’s different about the two guides. Wu, a Chinese national, accomplished hardware maker, and former tech influencer, has gone quiet since the summer of 2023, following interactions with state security actors. The guide’s crowdfunding page notes that “offering an app or download specifically for English-speaking hardware engineers to install on their phones would be… iffy.” Wu adds, “If at some point ‘I’ do offer you such a thing, I’d suggest you not use it.”

Huang, who previously helped sue the government over DRM rules, designed and sold the Chumby, and was one of the first major Xbox hackers, released the original Essential Guide on the rights-friendly Crowd Supply under a Creative Commons license (BY-NC-SA 4.0) that restricted commercial derivatives without explicit permission, which he granted to Wu. The book costs $30, with roughly $8 shipping costs to the US. It is dedicated to Gavin Zhao, whom Huang considered a mentor and who furthered his ambition to print the original guide.

Listing image by Machinery Enchantress/Crowd Supply

A New Essential Guide to Electronics by Naomi Wu details a different Shenzen Read More »

cvs,-rite-aid,-walgreens-hand-out-medical-records-to-cops-without-warrants

CVS, Rite Aid, Walgreens hand out medical records to cops without warrants

Amazon, cops, cvs, health, health privacy, HHS, HIPAA, law enforcement, Medical records, pharmacy, Policy, prescription drugs, rite aid, Science, Walgreens, warrant / /

prescription for privacy —

Lawmakers want HHS to revise health privacy law to require warrants.

–

CVS, Rite Aid, Walgreens hand out medical records to cops without warrants

All of the big pharmacy chains in the US hand over sensitive medical records to law enforcement without a warrant—and some will do so without even running the requests by a legal professional, according to a congressional investigation.

The revelation raises grave medical privacy concerns, particularly in a post-Dobbs era in which many states are working to criminalize reproductive health care. Even if people in states with restrictive laws cross state lines for care, pharmacists in massive chains, such as CVS, can access records across borders.

Lawmakers noted the pharmacies’ policies for releasing medical records in a letter dated Tuesday to the Department of Health and Human Services (HHS) Secretary Xavier Becerra. The letter—signed by Sen. Ron Wyden (D-Ore.), Rep. Pramila Jayapal (D-Wash.), and Rep. Sara Jacobs (D-Calif.)—said their investigation pulled information from briefings with eight big prescription drug suppliers.

They include the seven largest pharmacy chains in the country: CVS Health, Walgreens Boots Alliance, Cigna, Optum Rx, Walmart Stores, Inc., The Kroger Company, and Rite Aid Corporation. The lawmakers also spoke with Amazon Pharmacy.

All eight of the pharmacies said they do not require law enforcement to have a warrant prior to sharing private and sensitive medical records, which can include the prescription drugs a person used or uses and their medical conditions. Instead, all the pharmacies hand over such information with nothing more than a subpoena, which can be issued by government agencies and does not require review or approval by a judge.

Three pharmacies—CVS Health, The Kroger Company, and Rite Aid Corporation—told lawmakers they didn’t even require their pharmacy staff to consult legal professionals before responding to law enforcement requests at pharmacy counters. According to the lawmakers, CVS, Kroger, and Rite Aid said that “their pharmacy staff face extreme pressure to immediately respond to law enforcement demands and, as such, the companies instruct their staff to process those requests in store.”

The rest of the pharmacies—Amazon, Cigna, Optum Rx, Walmart, and Walgreens Boots Alliance—at least require that law enforcement requests be reviewed by legal professionals before pharmacists respond. But, only Amazon said it had a policy of notifying customers of law enforcement demands for pharmacy records unless there were legal prohibitions to doing so, such as a gag order.

HIPAA and transparency

The lawmakers note that the pharmacies aren’t violating regulations under the Health Insurance Portability and Accountability Act (HIPAA). The pharmacies pointed to language in HIPAA regulations that allow health care providers, including pharmacists, to provide medical records if required by law, with subpoenas being a sufficient legal process for such a request. However, the lawmakers note that the HHS has discretion in determining the legal standard here—that is, it has the power to strengthen the regulation to require a warrant, which the lawmakers say it should do.

“We urge HHS to consider further strengthening its HIPAA regulations to more closely align them with Americans’ reasonable expectations of privacy and Constitutional principles,” the three lawmakers wrote.

They also pushed for pharmacies to do better, encouraging them to follow the lead of tech companies. “Pharmacies can and should insist on a warrant, and invite law enforcement agencies that insist on demanding patient medical records with solely a subpoena to go to court to enforce that demand. The requirement for a warrant is exactly the approach taken by tech companies to protect customer privacy.” The trio noted that Google, Microsoft, and Yahoo have since 2010 required law enforcement to have a warrant to obtain customers’ emails.

Also noting tech companies’ lead, the lawmakers encouraged pharmacies to publish annual transparency reports. In the course of the investigation, only CVS Health said it planned to do so.

“Americans deserve to have their private medical information protected at the pharmacy counter and a full picture of pharmacies’ privacy practices, so they can make informed choices about where to get their prescriptions filled,” the lawmakers wrote.

For now, HIPAA regulations grant patients the right to know who is accessing their health records. But, to do so, patients have to specifically request that information—and almost no one does that. “Last year, CVS Health, the largest pharmacy in the nation by total prescription revenue, only received a single-digit number of such consumer requests,” the lawmakers noted.

“The average American is likely unaware that this is even a problem,” the lawmakers said.

CVS, Rite Aid, Walgreens hand out medical records to cops without warrants Read More »

everybody’s-talking-about-mistral,-an-upstart-french-challenger-to-openai

Everybody’s talking about Mistral, an upstart French challenger to OpenAI

AI, Anthropic, Arthur Mensch, Biz & IT, chatgpt, chatgtp, france, GPT-3, GPT-3.5, GPT-4, large language models, LLaMA, Llama 2, machine learning, Meta, Mistral, Mistral AI, Mixtral, Mixtral 8x7B, openai, Paris / /

A challenger appears —

“Mixture of experts” Mixtral 8x7B helps open-weights AI punch above its weight class.

–

An illustrated robot holding a French flag.

Enlarge / An illustration of a robot holding a French flag, figuratively reflecting the rise of AI in France due to Mistral. It’s hard to draw a picture of an LLM, so a robot will have to do.

On Monday, Mistral AI announced a new AI language model called Mixtral 8x7B, a “mixture of experts” (MoE) model with open weights that reportedly truly matches OpenAI’s GPT-3.5 in performance—an achievement that has been claimed by others in the past but is being taken seriously by AI heavyweights such as OpenAI’s Andrej Karpathy and Jim Fan. That means we’re closer to having a ChatGPT-3.5-level AI assistant that can run freely and locally on our devices, given the right implementation.

Mistral, based in Paris and founded by Arthur Mensch, Guillaume Lample, and Timothée Lacroix, has seen a rapid rise in the AI space recently. It has been quickly raising venture capital to become a sort of French anti-OpenAI, championing smaller models with eye-catching performance. Most notably, Mistral’s models run locally with open weights that can be downloaded and used with fewer restrictions than closed AI models from OpenAI, Anthropic, or Google. (In this context “weights” are the computer files that represent a trained neural network.)

Mixtral 8x7B can process a 32K token context window and works in French, German, Spanish, Italian, and English. It works much like ChatGPT in that it can assist with compositional tasks, analyze data, troubleshoot software, and write programs. Mistral claims that it outperforms Meta’s much larger LLaMA 2 70B (70 billion parameter) large language model and that it matches or exceeds OpenAI’s GPT-3.5 on certain benchmarks, as seen in the chart below.

A chart of Mixtral 8x7B performance vs. LLaMA 2 70B and GPT-3.5, provided by Mistral.

Enlarge / A chart of Mixtral 8x7B performance vs. LLaMA 2 70B and GPT-3.5, provided by Mistral.

Mistral

The speed at which open-weights AI models have caught up with OpenAI’s top offering a year ago has taken many by surprise. Pietro Schirano, the founder of EverArt, wrote on X, “Just incredible. I am running Mistral 8x7B instruct at 27 tokens per second, completely locally thanks to @LMStudioAI. A model that scores better than GPT-3.5, locally. Imagine where we will be 1 year from now.”

LexicaArt founder Sharif Shameem tweeted, “The Mixtral MoE model genuinely feels like an inflection point — a true GPT-3.5 level model that can run at 30 tokens/sec on an M1. Imagine all the products now possible when inference is 100% free and your data stays on your device.” To which Andrej Karpathy replied, “Agree. It feels like the capability / reasoning power has made major strides, lagging behind is more the UI/UX of the whole thing, maybe some tool use finetuning, maybe some RAG databases, etc.”

Mixture of experts

So what does mixture of experts mean? As this excellent Hugging Face guide explains, it refers to a machine-learning model architecture where a gate network routes input data to different specialized neural network components, known as “experts,” for processing. The advantage of this is that it enables more efficient and scalable model training and inference, as only a subset of experts are activated for each input, reducing the computational load compared to monolithic models with equivalent parameter counts.

In layperson’s terms, a MoE is like having a team of specialized workers (the “experts”) in a factory, where a smart system (the “gate network”) decides which worker is best suited to handle each specific task. This setup makes the whole process more efficient and faster, as each task is done by an expert in that area, and not every worker needs to be involved in every task, unlike in a traditional factory where every worker might have to do a bit of everything.

OpenAI has been rumored to use a MoE system with GPT-4, accounting for some of its performance. In the case of Mixtral 8x7B, the name implies that the model is a mixture of eight 7 billion-parameter neural networks, but as Karpathy pointed out in a tweet, the name is slightly misleading because, “it is not all 7B params that are being 8x’d, only the FeedForward blocks in the Transformer are 8x’d, everything else stays the same. Hence also why total number of params is not 56B but only 46.7B.”

Mixtral is not the first “open” mixture of experts model, but it is notable for its relatively small size in parameter count and performance. It’s out now, available on Hugging Face and BitTorrent under the Apache 2.0 license. People have been running it locally using an app called LM Studio. Also, Mistral began offering beta access to an API for three levels of Mistral models on Monday.

Everybody’s talking about Mistral, an upstart French challenger to OpenAI Read More »

after-15-months-blue-origin’s-new-shepard-spacecraft-will-finally-fly-again

After 15 months Blue Origin’s New Shepard spacecraft will finally fly again

blue origin, new shepard, Science, Space / /

Blue is back —

Taking some science and some postcards for a ride.

–

Photos from New Shepard launch day.

Enlarge / Blue Origin’s New Shepard launch system consists of a booster and a capsule.

Blue Origin is finally returning to flight.

On Tuesday the company announced, via the social media site X, that its New Shepard spacecraft would launch no earlier than next Monday.

“We’re targeting a launch window that opens on Dec. 18 for our next New Shepard payload mission,” the company stated. “#NS24 will carry 33 science and research payloads as well as 38,000 @clubforfuture postcards to space.”

The uncrewed New Shepard 24 test flight will refly the science payloads that were aboard the New Shepard 23 flight, which experienced an engine nozzle failure at 1 minute and 4 seconds following liftoff in September 2022. The capsule’s emergency escape system performed as intended, rapidly pulling the spacecraft away from the disintegrating rocket and allowing Blue Origin to recover the payloads flown for NASA and other customers.

Blue Origin finished its accident analysis this spring and implemented a fix to the problem, including design changes to the BE-3 engine combustion chamber. In May, the company said it planned to return to flight “soon.” Then, in September, the Federal Aviation Administration closed its mishap investigation.

The company originally targeted an uncrewed return-to-flight mission in early October; however, two sources told Ars that the additional two-month delay was caused by an issue with certifying an engine part intended for flight.

A new rocket?

Blue Origin has not specified which rocket and spacecraft will be flying next week from its launch site in West Texas, near the town of Van Horn. The company’s first New Shepard rocket, Booster 1, was lost during an April 2015 flight. Booster 2 was retired in October 2016 after performing a successful test of the launch escape system on its fifth and final flight. Booster 3, which was lost during the NS-23 mission in September, was the company’s oldest operational rocket, making its debut in December 2017.

The company has used its newest rocket, Booster 4, exclusively for human launches on New Shepard. This rocket has some modifications from Booster 3 to qualify it as a human-rated rocket. The company has also built a fifth booster that may be making next Monday’s flight.

Tuesday’s announcement came amid a tidal wave of changes in leadership at Blue Origin this month, with several high-profile retirements and the arrival of its new chief executive, who has come to the company from Amazon, Dave Limp. He replaced Bob Smith, who had an uneven tenure as leader of Blue Origin. As Ars reported last month, Limp is likely evaluating the long-term prospects of New Shepard, which remains far from breaking even financially.

The company may pivot toward its larger projects, including the New Glenn rocket and lunar lander for the Artemis program, which have a greater chance of raising significant revenue. Amazon founder Jeff Bezos has funded Blue Origin out of his pocket, providing as much as $2 billion a year in operating expenses.

However, given the announcement of New Shepard’s return to flight, it’s clear that Blue Origin isn’t moving entirely away from New Shepard just yet.

After 15 months Blue Origin’s New Shepard spacecraft will finally fly again Read More »

a-locally-grown-solution-for-period-poverty

A locally grown solution for period poverty

agave, agriculture, materials science, Science, Sustainability / /

Absorbant agave —

A Kenyan tinkerer and Stanford engineer team up to make maxi pads from agave fibers.

–

Image of rows of succulents with long spiky leaves and large flower stalks.

Enlarge / Sisal is an invasive species that is also grown agriculturally.

Women and girls across much of the developing world lack access to menstrual products. This means that for at least a week or so every month, many girls don’t go to school, so they fall behind educationally and often never catch up economically. 

Many conventional menstrual products have traditionally been made of hydrogels made from toxic petrochemicals, so there has been a push to make them out of biomaterials. But this usually means cellulose from wood, which is in high demand for other purposes and isn’t readily available in many parts of the globe. So Alex Odundo found a way to solve both of these problems: making maxi pads out of sisal, a drought-tolerant agave plant that grows readily in semi-arid climates like his native Kenya.

Putting an invasive species to work

Sisal is an invasive plant in rural Kenya, where it is often planted as livestock fencing and feedstock. It doesn’t require fertilizer, and its leaves can be harvested all year long over a five- to seven-year span. Odundo and his partners in Manu Prakash’s lab at Stanford University developed a process to generate soft, absorbent material from the sisal leaves. It relies on treatment with dilute peroxyformic acid (1 percent) to increase its porosity, followed by washing in sodium hydroxide (4 percent) and then spinning in a tabletop blender to enhance porosity and make it softer. 

They tested their fibers with a mixture of water mixed with glycerol—to make it thicker, like blood—and found that it is as absorbent as the cotton used in commercially available maxi pads. It was also as absorbent as wood pulp and more absorbent than fibers prepared from other biomaterials, including hemp and flax. Moreover, their process is less energy-intensive than conventional processing procedures, which are typically performed at higher temperatures and pressures. 

In a cradle-to-gate carbon footprint life cycle analysis, including sisal cultivation, harvesting, manufacturing, and transportation, sisal cellulose microfiber production fared roughly the same as production of cellulose microfiber from wood and much better than that from cotton in terms of both carbon footprint and water consumption, possibly because cotton requires so much upstream fertilizer. Much of the footprint comes from transportation, highlighting how useful it can be to make products like this in the same communities that need them.

Science for the greater good

This is not Odundo’s first foray into utilizing sisal; at Olex Techno Enterprises in Kisumu, Kenya, he has been making machines to turn sisal leaves into rope for over 10 years. This benefits local farmers since sisal rope and even sisal fibers sell for ten times as much as sisal leaves. In addition to making maxi pads, Odundo also built a stove that burns sawdust, rice husks, and other biodegradable waste products. 

By reducing wood stoves, he is reducing deforestation and improving the health of the women who breathe in the smoke of the cookfires. Adoption of such stoves have been a goal of environmentalists for years, and although a number of prototypes have been developed by mostly male engineers in developed countries, they have not been widely used because they are not that practical or appealing to the mostly female cooks in developing countries—the people who actually need to cook with them, yet were not consulted in their design.

Manu Prakash’s lab’s website proclaims that “we are dedicated toward inventing and distributing ‘frugal science’ tools to democratize access to science.” Partnering with Alex Odundo to manufacture menstrual products in the low-income rural communities that most need them seems like the apotheosis of that goal.

Communications Engineering, 2023. DOI:  10.1038/s44172-023-00130-y

A locally grown solution for period poverty Read More »

the-quest-to-turn-basalt-dust-into-a-viable-climate-solution

The quest to turn basalt dust into a viable climate solution

carbon capture, climate adaptation, climate adaptation strategies, Science, syndication / /
The quest to turn basalt dust into a viable climate solution

Mary Yap has spent the last year and a half trying to get farmers to fall in love with basalt. The volcanic rock is chock full of nutrients, captured as its crystal structure forms from cooling magma, and can make soil less acidic. In that way it’s like limestone, which farmers often use to improve their soil. It’s a little more finicky to apply, and certainly less familiar. But basalt also comes with an important side benefit: It can naturally capture carbon from the atmosphere.

Yap’s pitch is part of a decades-long effort to scale up that natural weathering process and prove that it can lock carbon away for long enough to make a different to the climate. “The bottleneck is getting farmers to want to do this,” Yap says.

On Thursday, Yap’s young startup, Lithos Carbon, got a $57.1 million boost for its quest to turn basalt dust into a viable climate solution. It came from Frontier, a benefit corporation backed by a consortium of companies aiming to finance promising approaches to carbon dioxide removal, or CDR. Lithos says it will use the funds to soak up 154,000 tons of CO2 by 2028, by sprinkling basalt dust on thousands of acres of US farmland. The average US car emits about 4 tons of CO2 each year.

The carbon removal purchase is the largest yet by Frontier, which was formed last year with nearly $1 billion from its tech-dominated members. Many of those companies, which include Meta, Alphabet, and payments processor Stripe, which owns Frontier, have made climate pledges that require not only reducing the emissions from their operations and supply chains but also “negative emissions”—sucking up carbon from the atmosphere to cancel out other emissions.

That accounting trick has been easier to prove out on paper than in practice. Many companies would have once turned to buying carbon offsets from activities like protecting forests that would otherwise be felled. But some have been trying to move away from those scandal-plagued and often short-lived approaches and into more durable techniques for carbon removal.

The current options for companies seeking negative emissions are limited. Frontier’s purchases are essentially down payments on ideas that are still in their infancy—generally too hard to verify or too expensive, or both, to attract a significant customer base. “What we’re trying to evaluate the field on is whether it’s on the trajectory to get to climate-relevant scale,” says Nan Ransohoff, who leads Frontier and also climate work at Stripe. The group starts with small “prepurchases” meant to help promising startups, and then moves on to “offtake” agreements for larger amounts of carbon that its members can count toward their emissions goals.

The Lithos purchase is one of those larger deals. It prices carbon removals at $370 per ton, about a quarter of which will pay for field monitoring and modeling to verify that carbon is being sequestered away from the atmosphere for the long term. Ransohoff says Frontier believes that Lithos is on a path to its goal of removing CO2 for customers at a cost of less than $100 per ton, and at a rate of at least a half a billion tons per year.

“Most promising” approach

Lithos, founded in 2022, is developing a technology called enhanced rock weathering. It involves spreading a fine dust of basalt across fields before planting. As the rock further weathers from rainfall, it reacts with CO2 in the air. That forms bicarbonate, which locks away the carbon by combining it with hydrogen and oxygen atoms. Ultimately, the compound is washed into the ocean, where the carbon should stay put.

The strategy has the benefit of piggybacking on things that humans already do, Yap says. That’s in contrast with techniques like direct air capture, which involves building industrial plants that suck carbon out of the atmosphere. It’s easy to measure carbon removed that way—it’s all captured there onsite—but critics say it will be difficult to scale up because removing enough carbon to make a difference will require thousands of dedicate, resource-intensive facilities.

Using basalt dust to capture carbon should be more easily scaled up. There are plenty of fields to dump rock dust onto, and plenty of water for carbon to end up in. But the distributed nature of the process also makes measuring how much carbon was actually removed from the atmosphere more difficult.

The quest to turn basalt dust into a viable climate solution Read More »

hubble-back-in-service-after-gyro-scare—nasa-still-studying-reboost-options

Hubble back in service after gyro scare—NASA still studying reboost options

Hubble Space Telescope, NASA, Science, Space, spacex / /
The Hubble Space Telescope viewed from Space Shuttle Atlantis during a servicing mission in 2009.

Enlarge / The Hubble Space Telescope viewed from Space Shuttle Atlantis during a servicing mission in 2009.

NASA

The Hubble Space Telescope resumed science observations on Friday after ground teams spent most of the last three weeks assessing the performance of a finicky gyroscope, NASA said.

The troublesome gyroscope is a critical part of the observatory’s pointing system. Hubble’s gyros measure how fast the spacecraft is turning, helping the telescope aim its aperture toward distant cosmic wonders.

Hubble still provides valuable scientific data for astronomers nearly 34 years since its launch aboard NASA’s Space Shuttle Discovery in 1990. Five more shuttle servicing missions repaired Hubble, upgraded its science instruments, and replaced hardware degraded from long-term use in space. Among other tasks, astronauts on the last of the shuttle repair flights in 2009 installed six new gyroscopes on Hubble.

Moving parts sometimes break

The gyros have long been one of the parts of Hubble that require the most upkeep. A wheel inside each gyro spins at a constant rate of 19,200 revolutions per minute, and the wheel is, in turn, sealed inside a cylinder suspended in a thick fluid, according to NASA. Electronics within each gyro detect very small movements of the axis of the wheel, which supply Hubble’s central computer with information about the spacecraft’s turn rate. Hair-thin wires route signals from the gyroscopes, and these wires can degrade over time.

Three of the six gyros installed on Hubble in 2009 have failed, and three others remain operational. The three still-functioning gyros are based on a newer design for longer life, but one of these units has shown signs of wear in the last few months. This gyroscope, designated Gyro 3, has always exhibited “consistent noisy behavior,” said Pat Crouse, Hubble project manager at NASA’s Goddard Space Flight Center.

Hubble typically needs three gyros to operate normally, so ground controllers shut down Gyro 3 for roughly seven years until Hubble needed it in 2018, when another gyroscope failed, leaving only three of the devices still working.

“Back in August, we saw issues,” Crouse told Ars this week. “It would sort of sporadically output some rate information that was not consistent with the observed spacecraft body rates, but it was short-lived, and we were characterizing what that performance was like and how much we could tolerate.”

The gyro’s performance worsened in November when it fed Hubble’s control system erroneous data. The gyroscope sensed that the spacecraft was changing its orientation when it really wasn’t moving. “That, then, contributed to an error in attitude that was kind of causing a little bit of drift,” Crouse said.

Automated software on Hubble detected the errors and put the spacecraft into “safe mode” two times last month. Hubble quickly resumed science observations each time but then went into safe mode again on November 23. Hubble managers took some extra time to gather data on the gyro’s health. Engineers commanded Hubble to move back and forth, and the suspect gyro consistently seemed to work well.

Hubble back in service after gyro scare—NASA still studying reboost options Read More »

eu-agrees-to-landmark-rules-on-artificial-intelligence

EU agrees to landmark rules on artificial intelligence

AI, AI law, Biz & IT, European Union, syndication / /

Get ready for some restrictions, Big Tech —

Legislation lays out restrictive regime for emerging technology.

–

EU Commissioner Thierry Breton talks to media during a press conference in June.

Enlarge / EU Commissioner Thierry Breton talks to media during a press conference in June.

Thierry Monasse | Getty Images

European Union lawmakers have agreed on the terms for landmark legislation to regulate artificial intelligence, pushing ahead with enacting the world’s most restrictive regime on the development of the technology.

Thierry Breton, EU commissioner, confirmed in a post on X that a deal had been reached.

He called it a historic agreement. “The EU becomes the very first continent to set clear rules for the use of AI,” he wrote. “The AIAct is much more than a rulebook—it’s a launchpad for EU start-ups and researchers to lead the global AI race.”

The deal followed years of discussions among member states and politicians on the ways AI should be curbed to have humanity’s interest at the heart of the legislation. It came after marathon discussions that started on Wednesday this week.

Members of the European Parliament have spent years arguing over their position before it was put forward to member states and the European Commission, the executive body of the EU. All three—countries, politicians, and the commission—must agree on the final text before it becomes law.

European companies have expressed their concern that overly restrictive rules on the technology, which is rapidly evolving and gained traction after the popularisation of OpenAI’s ChatGPT, will hamper innovation. Last June, dozens of some of the largest European companies, such as France’s Airbus and Germany’s Siemens, said the rules were looking too tough to nurture innovation and help local industries.

Last month, the UK hosted a summit on AI safety, leading to broad commitments from 28 nations to work together to tackle the existential risks stemming from advanced AI. That event attracted leading tech figures such as OpenAI’s Sam Altman, who has previously been critical of the EU’s plans to regulate the technology.

© 2023 The Financial Times Ltd. All rights reserved. Please do not copy and paste FT articles and redistribute by email or post to the web.

EU agrees to landmark rules on artificial intelligence Read More »

revisiting-the-ford-mustang-mach-e—how’s-the-pony-ev-doing-3-years-later?

Revisiting the Ford Mustang Mach-E—how’s the pony EV doing 3 years later?

car review, Cars, electric car review, Ford Mustang Mach-E / /

yay or neigh —

This midsize crossover EV has a lot more competition than when it debuted in 2021.

–

A Ford Mustang Mach-E, head-on

Enlarge / The Ford Mustang Mach-E is now in its third year of production, so it felt like a good idea to see how it’s maturing.

Jonathan Gitlin

When Ars first drove the then-new Ford Mustang Mach-E back in early 2021, the car was an attention magnet. Now, almost three years later, the Mustang Mach-E is a much more common sight on our roads, but so are other electric crossovers from most of Ford’s usual rivals, including the sales juggernaut that is the Tesla Model Y. We decided to book a few days with a Mustang Mach-E to see how (or if) this equine EV has matured since launch.

Originally, Ford had been working on a much more boring battery electric car until Tesla started delivering its Model 3s, at which point a hastily convened “Team Edison” set to work adding some much-needed brio to the design, rethinking Ford’s EV strategy in the process.

Giving this midsize crossover EV a Mustang name tag remains divisive—I expect a reasonable percentage of comments to this story will be people showing up to complain, “It ain’t no real Mustang.” The crossover’s name is what it is, and there are plenty of Mustang styling cues, but even with the designers’ trick of using black trim to make you ignore the bits they don’t want you to see, there’s no denying the proportions are pretty far from coupe-like.

The Mustang Mach-E has good angles and bad angles.

Enlarge / The Mustang Mach-E has good angles and bad angles.

Jonathan Gitlin

It’s cheaper now

Our test car was a Mustang Mach-E Premium eAWD model with just a single option ticked, the 91 kWh (useable capacity) extended range battery. This increases the car’s EPA range estimate from 224 miles (360 km) to 290 miles (467 km) but costs $8,600, which, combined with the delivery charge, bumps the sticker price to $67,575.

At least, that’s what this Mustang Mach-E cost when it arrived on the press fleet some 5,500 miles ago. Ford had to respond to Tesla’s string of price cuts, dropping the MSRP by almost $7,000 and cutting the cost of the extended range battery to $7,000—when I configure the same spec on Ford’s online car builder, it tells me the total price should be $59,940 with all the various fees. (A further price cut came to most other Mustang Mach-E variants in May, but not for the extended range Premium eAWD.)

And until the end of this year, the EV is still eligible for half of the clean vehicle tax credit. However, Ford believes that $3,750 credit will no longer be available to Mach-E buyers from next year as new rules regarding batteries made by “foreign entities of concern” go into effect. These remove eligibility from EVs batteries made in China or by Chinese-owned companies from January 1, 2024.

Taillights say Mustang, but the car's width and height say crossover.

Enlarge / Taillights say Mustang, but the car’s width and height say crossover.

Jonathan Gitlin

It’s a hard life

The fact that the Mustang Mach-E’s trip computer hadn’t been reset in 3,572 miles (5,749 km) provides an illustrating insight into both the life of a press fleet vehicle as well as the long-term efficiency of this EV. Collectively, the car had been driven very unsympathetically over that time, grading the drivers at 1 percent for deceleration and 2 percent for both acceleration and speed. Despite the lead foot treatment, the average of 2.7 miles/kWh (23 kWh/100 km) matches the EPA efficiency estimate (expressed as 37 kWh/100 miles).

I’m guessing this particular car spent most of those miles in Unbridled, which is what the Mustang Mach-E calls its sport mode. That or Engage, which is the middle of the three settings and the one used to calculate the car’s official efficiency.

In Whisper (think eco mode), you don’t get quite all the 346 hp (258 kW) or the full 428 lb-ft (580 Nm), and the 0–60 time feels between a second or two slower than the 4.8 seconds that’s possible if all the electric horses are harnessed at the same time, at least 3.1 miles/kWh (20 kWh/100 km) should be possible.

Revisiting the Ford Mustang Mach-E—how’s the pony EV doing 3 years later? Read More »