An anonymous reader shares this report from Tom's Hardware: When we think about Raspberry Pi, we normally picture single-board computers, but the Raspberry Pi Foundation was started to help kids learn about computers and it wants to help whether or not you own its hardware. The non-profit arm of Raspberry Pi this week released its new, browser-based code editor that's designed for young people (or any people) who are learning.

The Raspberry Pi Code Editor, which is considered to be in beta, is available to everyone for free right now at editor.raspberrypi.org. The editor is currently designed to work with Python only, but the organization says that support for other languages such as HTML, JavaScript and CSS is coming....

The Raspberry Pi Foundation already had a nice set of Python tutorials on its site, but it has adapted some of them to open sample code directly in the online editor....The Pi Foundation says that it plans to add a number of features to the Code Editor, including sharing and collaboration. The organization also plans to release the editor as an open-source project so anyone can modify it.
There's a pane showing your code's output when you click the "Run" button (plus a smaller pane for adding additional files to a project).

Tom's Hardware notes that "Since the entire programming experience takes place online, there's no way (at least right now) to use Python to control local hardware on your PC or your Raspberry Pi." But on the plus side, "If you create a free account on raspberrypi.org, which I did, the system will save all of your projects in the cloud and you can reload them any time you want. You can also download all the files in a project as a .zip file."

Startups and other companies trying to capitalize on the artificial intelligence boom sparked by OpenAI are running into a problem: They can't find enough specialized computers to make their own AI software. The Information: A spike in demand for server chips that can train and run machine-learning software has caused a shortage, prompting major cloud-server providers including Amazon Web Services, Microsoft, Google and Oracle to limit their availability for customers, according to interviews with the cloud companies and their customers. Some customers have reported monthslong wait times to rent the hardware. "All the startups who are trying to get into this space...maybe they can get one [server] but there's no way they're going to get five," said Johnny Dallas, founder and CEO of Zeet, which sells software that makes it easier for engineers to run apps across multiple clouds.

The server chip shortage is a frustrating hangup for software developers trying to build AI tools hinging on recent advancements in machine-learning models. These programmers, at small and big companies alike, are developing large-language models to make personalized writing coaches or search engines that respond to questions with written answers rather than links, similar to OpenAI's ChatGPT. Many others are licensing and augmenting software from OpenAI and its rivals to create specialized customer service chatbots and research tools for corporate employees. For instance, OpenAI software is helping Morgan Stanley bankers find the best locations to auction a work of art, based on the bank's myriad internal reports on art markets.

Multi-brand consumer megacorp Unilever says it has become a "cloud-only enterprise" with the help of Accenture and Microsoft. From a report: One of the largest and most complex cloud migrations in the retail goods industry, according to the company, will give Unilever "resilient, secure and optimised operations" as well as "a platform to drive innovation and growth." The Anglo-Dutch biz owns more than 400 brands, which include everything from ice cream to shampoo to toilet cleaner, and is set to use Microsoft's Azure as its "primary cloud platform."

According to the corporate blurb, the move will see Unilever employ "industrial metaverse technologies" that use real-time data from factory digital twins. It musn't have got the memo from Microsoft, which recently put a bullet in its own industrial metaverse masterplan. The cloud contract is also expected to help "achieve perpetual breakthroughs in research and development," says Unilver. Lastly, through Microsoft's partnership with the controversial GPT maker, it will use "Azure OpenAI Service across Unilever's business to drive increased automation, enabling better customer and employee experiences."

The UK's communications watchdog has called for a probe into Microsoft and Amazon's dominance of the country's cloud computing market in the latest challenge to the tech giants from global regulators. From a report: Ofcom said on Wednesday it was "particularly concerned" by the practices of Amazon Web Services and Microsoft, which together control between 60 and 70 per cent of the UK cloud market. It has proposed referring the sector to the Competition and Markets Authority for further investigation. Cloud computing is dominated by Amazon and Microsoft, and has become a crucial driver of revenue at the tech giants. But growth in demand for these services has slowed this year and customers have sought to cut costs, with some complaining of rising prices and the difficulty of moving between cloud providers. Ofcom's move comes amid growing global scrutiny over the cloud market. Last year, Microsoft changed its cloud licensing policies in Europe in an effort to head off potential antitrust action from regulators in Brussels. The tech companies are already the targets of competition watchdogs in the US, UK and EU on multiple fronts, with investigations into Microsoft's $75bn acquisition of video games maker Activision and Amazon's deal to buy Roomba-maker iRobot. Ofcom said it was concerned that, if unchecked, the concentration of cloud computing supply in the hands of a small number of large US companies could lead to British customers paying more and smaller groups being squeezed out of the market.

Atlassian customers' eleven-year quest for custom domains continues, with the Australian upstart's proposed solution failing to satisfy. The Register: As The Register reported in 2022, Atlassian floated the idea of custom domains for its custom apps in 2011. Yes, 2011. The ticket for the change is called "CLOUD 6999" and has become infamous for the length of time it has remained unresolved. An unidentified wag has even made t-shirts bearing the CLOUD 6999 name. Atlassian promised last year to sort it out some time in 2023, and in February posted an update on its initial designs.

It hasn't gone down well. Atlassian's proposed solution requires "a company-branded domain name, a list of options for the 1st-level subdomain keyword, and a 2nd-level subdomain at your own choice." Atlassian cloud admin experience chap Luke Liu explained that structure as delivering URLs such as internal.support.acme.com or people.knowledge.acme.org. One of Atlassian's stated company values is "Don't #@!% the customer." But plenty of Atlassian customers feel well and truly #@!%ed by the custom domain plan. "The cloud roadmap specifically uses an example of 1 level," wrote one commenter on the 1,445-item thread discussing CLOUD 6999. "The team managing this seems to be completely lost and disconnected from the user base."

The newest startup accelerator from Amazon aims to attract companies building generative AI technologies. From a report: The Amazon Web Services accelerator, revealed Tuesday, is a 10-week program aims to "empower companies applying generative AI to solutions from legal and marketing, to software engineering, green energy, and life sciences, including drug discovery." It also provides up to $300,000 in AWS credits. The hybrid program is open to all startups, with two week-long in-person events in San Francisco. AWS does not take equity from participating companies. The accelerator is a way for Amazon to draw early-stage startups into its cloud ecosystem.

Wired pays a visit to a half-finished Microsoft data center that rises out of the flat North Holland farmland — where the security guard tells a local councillor he's not allowed to visit the site, and "Within minutes, the argument has escalated, and the guard has his hand around Ruiter's throat." The security guard lets go of Ruiter within a few seconds, and the councillor escapes with a red mark across his neck. Back in his car, Ruiter insists he's fine. But his hands shake when he tries to change gears. He says the altercation — which he will later report to the police — shows the fog of secrecy that surrounds the Netherlands' expanding data center business.

"We regret an interaction that took place outside our data center campus, apparently involving one of Microsoft's subcontractors," says Craig Cincotta, general manager at Microsoft, adding that the company would cooperate with the authorities.

The heated exchange between Ruiter and Microsoft's security guard shows how contentious Big Tech's data centers have become in rural parts of the Netherlands. As the Dutch government sets strict environmental targets to cut emissions, industries are being forced to compete for space on Dutch farmland — pitting big tech against the increasingly political population of Dutch farmers.

There are around 200 data centers in the Netherlands, most of them renting out server space to several different companies. But since 2015, the country has also witnessed the arrival of enormous "hyperscalers," buildings that generally span at least 10,000 square feet and are set up to service a single (usually American) tech giant. Lured here by the convergence of European internet cables, temperate climates, and an abundance of green energy, Microsoft and Google have built hyperscalers; Meta has tried and failed.

Against the backdrop of an intensifying Dutch nitrogen crisis, building these hyperscalers is becoming more controversial. Nitrogen, produced by cars, agriculture, and heavy machinery used in construction, can be a dangerous pollutant, damaging ecosystems and endangering people's health. The Netherlands produces four times more nitrogen than the average across the EU. The Dutch government has pledged to halve emissions by 2030, partly by persuading farmers to reduce their livestock herds or leave the industry altogether. Farmers have responded with protests, blockading roads with tractors and manure and dumping slurry outside the nature minister's home.
Farmers object that Microsoft is building its data center before it's even received government permits certifying that it won't worsen the nitrogen problem, according to the article. In response the Farmer Citizen Movement has sprung up, and last month it became the joint-largest party in the Dutch Senate. One party leader tells Wired, "It is a waste of fertile soil to put the data centers boxes here."

And Wired adds that opposition to datacenter development is also growing elsewhere in Europe.

Alphabet's Google Cloud has accused Microsoft of anti-competitive cloud computing practices and criticised imminent deals with several European cloud vendors, saying these do not solve broader concerns about its licensing terms. From a report: In Google Cloud's first public comments on Microsoft and its European deals its Vice President Amit Zavery told Reuters the company has raised the issue with antitrust agencies and urged European Union antitrust regulators to take a closer look.

In response, Microsoft referred to a blogpost in May last year where its president Brad Smith said it 'has a healthy number two position when it comes to cloud services, with just over 20 percent market share of global cloud services revenues'. "We are committed to the European Cloud Community and their success," a Microsoft spokesperson told Reuters on Thursday. There is intense rivalry between the two U.S. tech giants in the fast-growing, multi-billion-dollar cloud computing business, where Google trails market leader Amazon and Microsoft.

Microsoft patched a dangerous security issue in Bing last month just days before it launched a new artificial intelligence-powered version of the search engine. From a report: The problem was discovered by outside researchers at the security firm Wiz. It was created by a mistake in the way that Microsoft configured applications on Azure, its cloud-computing platform, and could be used to gain access to emails and other documents of people who used Bing, the researchers said. Microsoft fixed the problem on Feb. 2, according to Ami Luttwak, Wiz's chief technology officer. Five days later Satya Nadella introduced the new generative AI capabilities to Bing, bringing a renewed interest in Microsoft's 14-year-old search engine. Usage of Bing has jumped, rising to more than 100 million daily active users in the month since the upgrade.

Alphabet's Google is striking a partnership to combine its artificial intelligence language models with software from startup Replit that helps computer programmers write code, a bid to compete with a similar product from Microsoft's GitHub and OpenAI. From a report: Replit's Ghostwriter, which has 20 million users, will rely on Google's language-generation AI to improve its ability to suggest blocks of code, complete programs and answer developer questions. Google Cloud Vice President June Yang declined to specify which language AI products Replit will use, noting that it's a customized combination of systems that address different tasks like chat and code-generation. Previously, Replit built the product with its own AI. Google "has much better technology than most people know," Replit Chief Executive Officer Amjad Masad said in an interview. The startup will also expand its use of Google's cloud services and hopes the relationship with the tech giant will help it win over larger corporate customers -- right now Replit's clients are largely individual developers and startups. Google also will distribute Replit's software as part of the partnership. GitHub, which is wholly owned by Microsoft, last year released a product called Copilot, which suggests blocks of code as a software developer types, speeding up the process and automating rote or finicky coding tasks.

Chinese e-commerce group Alibaba Group Holding will reorganize into six business groups and pursue public listings for five of them, in the most significant governance overhaul since the company was established 24 years ago. From a report: The company announced the move on Tuesday, a day after founder Jack Ma's surprise return to China following a lengthy stint abroad. The six business groups will focus on sectors such as cloud computing, e-commerce and logistics. "This transformation will empower all our businesses to become more agile, enhance decision-making, and enable faster responses to market changes," chief executive Daniel Zhang said in a letter to employees. The six new groups will be: Cloud Intelligence Group, Taobao Tmall Commerce Group, Local Service Group, Cainiao Smart Logistics, Global Digital Commerce Group and Digital Media and Entertainment Group.

Each of the groups will be run by its own CEO and board of directors, with the CEOs assuming full responsibility for company performance. Zhang will remain chairman and CEO of Alibaba Group, which will follow a holding company management model. He will also serve as the CEO of the Cloud Intelligence Group, which will be responsible for the company's cloud and artificial intelligence businesses. Zhang became acting president of Alibaba Cloud Intelligence after its cloud service suffered what the company described as "the longest large-scale outage in more than a decade" in Hong Kong in late December.

A Google-led research paper published in Nature, claiming machine-learning software can design better chips faster than humans, has been called into question after a new study disputed its results. The Register reports: In June 2021, Google made headlines for developing a reinforcement-learning-based system capable of automatically generating optimized microchip floorplans. These plans determine the arrangement of blocks of electronic circuitry within the chip: where things such as the CPU and GPU cores, and memory and peripheral controllers, actually sit on the physical silicon die. Google said it was using this AI software to design its homegrown TPU chips that accelerate AI workloads: it was employing machine learning to make its other machine-learning systems run faster. The research got the attention of the electronic design automation community, which was already moving toward incorporating machine-learning algorithms into their software suites. Now Google's claims of its better-than-humans model has been challenged by a team at the University of California, San Diego (UCSD).

Led by Andrew Kahng, a professor of computer science and engineering, that group spent months reverse engineering the floorplanning pipeline Google described in Nature. The web giant withheld some details of its model's inner workings, citing commercial sensitivity, so the UCSD had to figure out how to make their own complete version to verify the Googlers' findings. Prof Kahng, we note, served as a reviewer for Nature during the peer-review process of Google's paper. The university academics ultimately found their own recreation of the original Google code, referred to as circuit training (CT) in their study, actually performed worse than humans using traditional industry methods and tools.

What could have caused this discrepancy? One might say the recreation was incomplete, though there may be another explanation. Over time, the UCSD team learned Google had used commercial software developed by Synopsys, a major maker of electronic design automation (EDA) suites, to create a starting arrangement of the chip's logic gates that the web giant's reinforcement learning system then optimized. The Google paper did mention that industry-standard software tools and manual tweaking were used after the model had generated a layout, primarily to ensure the processor would work as intended and finalize it for fabrication. The Googlers argued this was a necessary step whether the floorplan was created by a machine-learning algorithm or by humans with standard tools, and thus its model deserved credit for the optimized end product. However, the UCSD team said there was no mention in the Nature paper of EDA tools being used beforehand to prepare a layout for the model to iterate over. It's argued these Synopsys tools may have given the model a decent enough head start that the AI system's true capabilities should be called into question.

The lead authors of Google's paper, Azalia Mirhoseini and Anna Goldie, said the UCSD team's work isn't an accurate implementation of their method. They pointed out (PDF) that Prof Kahng's group obtained worse results since they didn't pre-train their model on any data at all. Prof Kahng's team also did not train their system using the same amount of computing power as Google used, and suggested this step may not have been carried out properly, crippling the model's performance. Mirhoseini and Goldie also said the pre-processing step using EDA applications that was not explicitly described in their Nature paper wasn't important enough to mention. The UCSD group, however, said they didn't pre-train their model because they didn't have access to the Google proprietary data. They claimed, however, their software had been verified by two other engineers at the internet giant, who were also listed as co-authors of the Nature paper. Separately, a fired Google AI researcher claims the internet goliath's research paper was "done in context of a large potential Cloud deal" worth $120 million at the time.

It's one of America's best hospitals — a nonprofit "academic medical center" called the Cleveland Clinic. And this week it installed an IBM-managed quantum computer to accelerate healthcare research (according to an announcement from IBM). IBM is calling it "the first quantum computer in the world to be uniquely dedicated to healthcare research."

The clinic's CEO said the technology "holds tremendous promise in revolutionizing healthcare and expediting progress toward new cares, cures and solutions for patients." IBM's CEO added that "By combining the power of quantum computing, artificial intelligence and other next-generation technologies with Cleveland Clinic's world-renowned leadership in healthcare and life sciences, we hope to ignite a new era of accelerated discovery."

em>Inside HPC points out that "IBM Quantum System One" is part of a larger biomedical research program applying high-performance computing, AI, and quantum computing, with IBM and the Cleveland Clinic "collaborating closely on a robust portfolio of projects with these advanced technologies to generate and analyze massive amounts of data to enhance research." The Cleveland Clinic-IBM Discovery Accelerator has generated multiple projects that leverage the latest in quantum computing, AI and hybrid cloud to help expedite discoveries in biomedical research. These include:

- Development of quantum computing pipelines to screen and optimize drugs targeted to specific proteins;

- Improvement of a quantum-enhanced prediction model for cardiovascular risk following non-cardiac surgery;

- Application of artificial intelligence to search genome sequencing findings and large drug-target databases to find effective, existing drugs that could help patients with Alzheimer's and other diseases.


The Discovery Accelerator also serves as the technology foundation for Cleveland Clinic's Global Center for Pathogen & Human Health Research, part of the Cleveland Innovation District. The center, supported by a $500 million investment from the State of Ohio, Jobs Ohio and Cleveland Clinic, brings together a team focused on studying, preparing and protecting against emerging pathogens and virus-related diseases. Through the Discovery Accelerator, researchers are leveraging advanced computational technology to expedite critical research into treatments and vaccines.

The UK's Competition and Markets Authority (CMA) has now sided with Microsoft over concerns the software giant could remove Call of Duty from PlayStation if its proposed Activision Blizzard deal is approved. From a report: The regulator still has concerns about the deal's impact on the cloud gaming market and will complete its investigation by the end of April. "Having considered the additional evidence provided, we have now provisionally concluded that the merger will not result in a substantial lessening of competition in console gaming services because the cost to Microsoft of withholding Call of Duty from PlayStation would outweigh any gains from taking such action," says Martin Coleman, chair of the independent panel of experts conducting the CMA's investigation. The CMA had originally provisionally concluded that a Microsoft strategy to withhold Call of Duty from PlayStation would be profitable. Microsoft wasn't happy with that conclusion and publicly criticized the regulator's math earlier this month, arguing that the CMA's financial modeling was flawed.

Careful scrutiny of the debris from the impact of NASA's DART mission into Dimorphos has not found any evidence for water-ice on the asteroid, nor the residue of thruster fuel from the spacecraft, new results from the European Southern Observatory (ESO) show. Space.com reports: However, the data from the MUSE (Multi-Unit Spectroscopic Explorer) instrument on ESO's Very Large Telescope in Chile does indicate differences in the size of particles in the debris, and show how the polarization of the light from the asteroid changed. These could both reveal details about the nature of the ejecta excavated by the impact, the recoil from which gave Dimorphos the biggest push. [...] "Before the impact, we were not really sure what to expect," said Cyrielle Opitom of the University of Edinburgh in an interview with Space.com.

Opitom led a team who used MUSE to go in search of any water on Dimorphos. They observed the Didymos-Dimorphos system on 11 occasions, from just before the impact to about a month afterwards. MUSE is able to split the light from the double-asteroid into a spectrum, or rainbow, of colors, to look for emission at specific wavelengths that corresponds to specific molecules. In particular, Opitom's team searched the ejecta for water molecules and for oxygen that could have come from the break-up of water molecules by the impact. However, no evidence of water was detected. Dimorphos, at least, seems to be a dry asteroid.

There was also no evidence in the ejecta of traces of the hydrazine fuel that was on board DART, nor the xenon from its ion engine, although given their small quantities the non-detection is not a surprise. However, MUSE's observations were able to track the evolution of the cloud of ejecta (debris) thrown up by the impact, and in particular they helped determine the size distribution of the dust particles initially in the ejecta cloud and later in the tail streaming away from the asteroid. The research was published in the journal Astronomy & Astrophysics.

The number of victims affected by a mass-ransomware attack, caused by a bug in a popular data transfer tool used by businesses around the world, continues to grow as another organization tells TechCrunch that it was also hacked. From the report: Canadian financing giant Investissement Quebec confirmed to TechCrunch that "some employee personal information" was recently stolen by a ransomware group that claimed to have breached dozens of other companies. Spokesperson Isabelle Fontaine said the incident occurred at Fortra, previously known as HelpSystems, which develops the vulnerable GoAnywhere file transfer tool. Hitachi Energy also confirmed this week that some of its employee data had been stolen in a similar incident involving its GoAnywhere system, but saying the incident happened at Fortra.

Over the past few days, the Russia-linked Clop gang has added several other organizations to its dark web leak site, which it uses to extort companies further by threatening to publish the stolen files unless a financial ransom demand is paid. TechCrunch has learned of dozens of organizations that used the affected GoAnywhere file transfer software at the time of the ransomware attack, suggesting more victims are likely to come forward. However, while the number of victims of the mass-hack is widening, the known impact is murky at best. Since the attack in late January or early February -- the exact date is not known -- Clop has disclosed less than half of the 130 organizations it claimed to have compromised via GoAnywhere, a system that can be hosted in the cloud or on an organization's network that allows companies to securely transfer huge sets of data and other large files.

Adobe is jumping into the generative AI game with the launch of a new family of AI models called Firefly. From a report: Focused on bringing AI into Adobe's suite of apps and services, specifically AI for generating media content, Firefly will be made up of multiple AI models "working across a variety of different use cases,' Adobe VP of generative AI Alexandru Costin told TechCrunch in an email interview. It's an expansion of the generative AI tools Adobe introduced in Photoshop, Express and Lightroom during its annual Max conference last year, which let users create and edit objects, composites and effects by simply describing them. As the fervor around the tech grows, Adobe has raced to maintain pace, for example allowing contributors to sell AI-generated artwork in its content marketplace.

"Firefly is the next step on our AI journey -- bringing together our new 'gentech' models with decades of investment in imaging, typography, illustration and more to produce assets," Costin said. "We'll bring this value to our customers' workflows where content is created across Creative Cloud, Experience Cloud and Document Cloud." Firefly as it exists today, in beta and without firm pricing (Adobe says that's coming), offers a single model designed to generate images and text effects from descriptions. Developed using hundreds of millions of photos, the model can create content across Adobe apps including Express, Photoshop, Illustrator and Adobe Experience Manager given a text prompt.

An anonymous reader writes: Last week, we learned that Microsoft spent hundreds of millions of dollars to buy tens of thousands of Nvidia A100 graphics chips so that partner OpenAI could train the large language models (LLMs) behind Bing's AI chatbot and ChatGPT.

Don't have access to all that capital or space for all that hardware for your own LLM project? Nvidia's DGX Cloud is an attempt to sell remote web access to the very same thing. Announced today at the company's 2023 GPU Technology Conference, the service rents virtual versions of its DGX Server boxes, each containing eight Nvidia H100 or A100 GPUs and 640GB of memory. The service includes interconnects that scale up to the neighborhood of 32,000 GPUs, storage, software, and "direct access to Nvidia AI experts who optimize your code," starting at $36,999 a month for the A100 tier.

Meanwhile, a physical DGX Server box can cost upwards of $200,000 for the same hardware if you're buying it outright, and that doesn't count the efforts companies like Microsoft say they made to build working data centers around the technology.

OpenAI's ChatGPT has enraptured the business world since its November release and OpenAI is signing up customers eager to pay to use its artificial intelligence models in their own products. But the Microsoft-backed startup faces a surprising rival: Microsoft itself. From a report: As part of its multibillion-dollar investment in OpenAI, Microsoft has the rights to sell the startup's software through its Azure cloud business, even as OpenAI licenses its own software directly to customers. Microsoft also gets a share of OpenAI's profits. The offerings cost the same, a fraction of a cent per query. Meanwhile, all of OpenAI's technology runs on Microsoft's Azure cloud infrastructure rent free. The dual offerings mean the companies are at times pitching the same customers on nearly identical products, putting salespeople at Microsoft in the uneasy position of trying to lure customers away from OpenAI while touting its technology.

While the profit-sharing agreement means sales of either offering theoretically benefit both companies, OpenAI pursues direct relationships with big customers, such as Microsoft rival Salesforce, which has licensed ChatGPT for a new suite of customer service software. It's not clear whether the partnership between OpenAI and Microsoft dictates the price each company sets for the models. Microsoft gets 75% of OpenAI's profits until its investment is paid back and 49% of subsequent profits up to a certain cap, The Information previously reported. It's also not clear how much profit Microsoft returns to OpenAI for the models it sells through Azure OpenAI Service. [...] An internal Microsoft document, viewed by The Information, instructs Azure salespeople to tell potential customers that OpenAI's own licenses are "great [for] experimentation" but have "limited enterprise-grade capabilities" and fewer "security/privacy features."

An anonymous reader quotes a report from Ars Technica: An analysis of 2,007 damaged or defective hard disk drives (HDDs) has led a data recovery firm to conclude that "in general, old drives seem more durable and resilient than new drives." The statement comes from a Los Angeles-headquartered HDD, SSD, and RAID data recovery firm aptly named Secure Data Recovery that has been in business since 2007 and claims to have resolved more than 100,000 cases. It studied the HDDs it received in 2022. "Most" of those drives were 40GB to 10TB, according to a blog post by Secure Data Recovery spotted by Blocks & Files on Thursday.

Secure Data Recovery's March 8 post broke down the HDDs it received by engineer-verified "power-on hours," or the total amount of time the drive was functional, starting from when its owner began using it and ending when the device arrived at Secure Data Recovery. The firm also determined the drives' current pending sector count, depicting "the number of damaged or unusable sectors the hard drive developed during routine read-and-write operations." The company's data doesn't include HDDs that endured non-predictable failures or damage by unexpected events, such as electrical surges, malware, natural disasters, and "accidental mishandling," the company said.

Among the sample, 936 drives are from Western Digital, 559 come from Seagate, 211 are Hitachi brand, 151 are Toshiba's, 123 are Samsung's, and there are 27 Maxtor drives. Notably, 74.5 percent of the HDDs came from either Western Digital or Seagate, which Secure Data Recovery noted accounted for 80 percent of hard drive shipments in 2021, citing Digital Storage Technology Newsletter data shared by Forbes. The average time before failure among the sample size was 2 years and 10 months, and the 2,007 defective HDDs had an average of 1,548 bad sectors. "While 1,548 bad sectors out of hundreds of millions or even billions of disk subdivisions might seem minuscule, the rate of development often increases, and the risk of data corruption multiplies," the blog said. "We found that the five most durable and resilient hard drives from each manufacturer were made before 2015," says Secure Data Recovery. "On the other hand, most of the least durable and resilient hard drives from each manufacturer were made after 2015." One of the reasons for this may have to do with HDD manufacturers "pushing the performance envelope," adds Ars. "This includes size limits that cut 'allowance between moving parts, appearing to affect mechanical damage and wear resistance.'"

Secure Data Recovery also believes that shingled magnetic recording (SMR) impacts HDD reliability, as the disks place components under "more stress."

"What this study shows is not the average working life of a hard disk drive," notes Blacks & Files. "Instead it provides the average working life if a failed disk drive. Cloud storage provider Backblaze issues statistics about the working life of its disk drive fleet and its numbers are quite different." A recent report of theirs found that SSDs are more reliable than HDDs.