Microsoft is releasing a new version of Office this week, designed for people that don't want to subscribe to Microsoft 365. From a report: The standalone Microsoft Office 2024 release is now available for both consumers and small businesses, and includes locked-in-time versions of Word, Excel, PowerPoint, OneNote, and Outlook across both Mac and PC. Office 2024 includes a lot of the updates that Microsoft has been delivering to Microsoft 365 subscribers over the past few years.

Microsoft last released a standalone version of Office in 2021, and this new Office 2024 release includes improvements to the core apps, as well as accessibility and UI changes. Office 2024 has a new default theme, with Microsoft's latest Fluent Design principles that match the visual changes to Windows 11. Microsoft has also added accessibility-focused improvements to help Office users find potential accessibility issues in documents, slideshows, workbooks, and emails.

Microsoft is bringing some new AI-powered Paint and Photos features to Copilot Plus PCs that could make creatives less reliant on more powerful image editing software. From a report: Generative Fill and Generative Erase -- which appear to be heavily inspired by similar AI tools in Adobe Photoshop -- are being introduced to Paint, allowing users to precisely add or remove objects in their images.

Both tools utilize a size-adjustable brush to "paint" over specific areas of an image to edit. Generative Erase will remove unwanted figures, objects like background clutter, and other distractions, similar to the Magic Eraser feature on Google's Pixel phones. Generative Fill allows Paint users to add new AI-generated assets to an image using a text description and select precisely where they should be placed -- much like the Photoshop tool that shares the same name. These build on the Cocreator tool for Paint announced for Copilot Plus PCs earlier this year that can generate images using a combination of text prompts and reference sketches. The company says the diffusion-based model powering these features has been updated to improve output quality and speed and now includes "built-in moderation" to help prevent it from being abused.

Microsoft won't impose a new return-to-office mandate unless management concludes that productivity has dropped, a high-level exec has reportedly told workers. From a report: The software and cloud-computing giant currently allows employees to work remotely, with many new hires promised the flexibility of working from home at least half the week. But that isn't written in stone. According to two anonymous sources that spoke with Business Insider, executive vice president Scott Guthrie recently told staff at his Microsoft's Cloud and AI group, which includes Azure, that a policy change isn't on the cards at present -- so long as workers stay productive.

While no statement has been provided as of press time, Microsoft told Business Insiderthat the company's work policies have not changed. Amazon CEO Andy Jassy's bombshell decree has roiled tech employees across the sector, many of whom dread a return to hours wasted in traffic jams on the long daily commute.

Microsoft launched its annual Windows 11 update today, introducing significant changes to the operating system. The Windows 11 2024 Update, or 24H2, will roll out gradually, starting with PCs running versions 22H2 or 23H2 that have opted for faster feature updates. Key additions include an Energy Saver feature, Wi-Fi 7 support, and 80Gbps USB4 Version 2.0 compatibility. Select high-end PCs meeting Copilot+ requirements will gain access to enhanced features like an improved Recall function and generative AI tools in Paint.

This update marks the most substantial overhaul of Windows 11 since its 2021 release, with major changes to the compiler, kernel, and scheduler. Microsoft has also improved the Arm-to-x86 app translation layer, now dubbed "Prism." While stable, users may encounter occasional issues. The update maintains Windows 11's existing hardware requirements but raises the bar for unsupported installations.

An anonymous reader shares a report: HoloLens 2 production has ended, Microsoft confirmed to UploadVR. Now is the last time to buy the device before stock runs out, the company has been telling its partners and customers. HoloLens 2 will continue to receive "updates to address critical security issues and software regressions" until December 31 2027. As soon as 2028 starts, software support for HoloLens 2 will end. For the original HoloLens headset from 2016, software support will end after December 10 of this year, just over two months from now. Production of it ended back in 2018. HoloLens 2 launched in 2019, three years after the original, with upgrades to almost every aspect: a wider field of view, higher resolution, eye tracking, vastly improved hand tracking, and more powerful compute housed in the rear of the strap to deliver a balanced comfortable design.

Microsoft has unveiled new features for its Copilot AI assistant, including screen analysis and voice interaction capabilities. Copilot Vision, available to Copilot Pro subscribers, can analyze web content in Microsoft Edge and answer queries about on-screen information. The company said processed data is immediately deleted and not used for model training.

A new Think Deeper function aims to tackle complex problems using advanced reasoning models. Copilot Voice introduces synthetic speech output and voice input in select English-speaking countries. Microsoft also announced personalization features, leveraging user history to tailor Copilot recommendations. This functionality will be limited initially, with the company evaluating options for European Economic Area users due to regulatory considerations.

Cerebras Systems, an AI chip startup, filed (PDF) for an IPO and plans to trade under the ticker "CBRS" on Nasdaq. CNBC reports: Cerebras competes with Nvidia, whose graphics processing units are the industry's choice for training and running AI models. Cerebras says on its website that its WSE-3 chip comes with more cores and memory than Nvidia's popular H100. It's also a physically larger chip. In addition to selling chips, Cerebras offers cloud-based services that rely on its own computing clusters. [...] In addition to Nvidia, Cerebras cites AMD, Intel, Microsoft and Google as competitors, "as well as internally developed custom application-specific integrated circuits and a variety of private companies." Taiwan Semiconductor Manufacturing Company makes the Cerebras chips. Cerebrus warned investors that any possible supply chain disruptions may hurt the company.

Cerebras was founded in 2016 and is based in Sunnyvale, California. Andrew Feldman, the startup's co-founder and CEO, sold server startup SeaMicro to AMD for $355 million in 2012. The company said in 2021 that it was valued at over $4 billion in a $250 million funding round.In May, G42 committed to purchasing $1.43 billion in orders from Cerebras before March 2025, according to the filing. G42 currently owns under 5% of Cerebras' Class A shares, and the firm has an option to purchase more depending on how much Cerebras product it buys.

Apple has withdrawn from discussions to invest in OpenAI's $6.5 billion funding round, though reasons for the decision remain unclear. The company still plans to proceed with integrating ChatGPT into Siri. MacRumors reports: The development comes just a month after WSJ reported that Apple was considering an investment in OpenAI as part of a fundraising effort that could value the AI company at over $100 billion. The high valuation reflects the intense competition in the artificial intelligence sector that OpenAI helped ignite with ChatGPT's launch in late 2022. While Apple has stepped away, other major tech companies remain involved. Microsoft, which has already invested $13 billion in OpenAI, is expected to contribute about $1 billion to this latest round. Nvidia is also reportedly in talks to participate. OpenAI's transition into a for-profit structure may have factored into Apple's decision. Last week, Reuters reported on OpenAI's plan to restructure its core business into a for-profit benefit corporation that will no longer be controlled by its non-profit board. "Chief executive Sam Altman will also receive equity for the first time in the for-profit company, which could be worth $150 billion after the restructuring as it also tries to remove the cap on returns for investors," reported Reuters.

U.S. real estate developers "are having a hard time keeping up with demand," reports the Los Angeles Times, "as businesses in search of secure spots for their servers rent nearly every square foot that becomes available..." Construction of new data centers is at "extraordinary levels" driven by "insatiable demand," a recent report on the industry by real estate brokerage JLL found. "Never in my career of 25 years in real estate have I seen demand like this on a global scale," said JLL real estate broker Darren Eades, who specializes in data centers...

The biggest drivers are AI and cloud service providers that include some of the biggest names in tech, such as Amazon, Microsoft, Google and Oracle. With occupancy in conventional office buildings still down sharply following the impact of the COVID-19 pandemic and property values falling, data centers represent a rare ripe opportunity for real estate developers, who are pursuing opportunities in major markets like Los Angeles and less urban locales that are served by plentiful and preferably cheap power needed to run data centers. "If you can find a cluster of power to build a site, they'll come," Eades said of developers. Construction is taking place at an "extraordinary" pace nationwide and still not keeping up, the JLL data center report said. [Data center] "Vacancy declined to a record low of 3% at midyear due to insatiable demand and despite rampant construction."

Development increased more than sevenfold in two years, with the pipeline of new projects leveling off in the first half of 2024, a potential signal that the U.S. power grid cannot support development at a faster pace. But when projects currently under construction or planned are complete, the U.S. colocation market, in which businesses rent space in a data center owned by another company for their servers and other computing hardware, will triple in size from current levels... Real estate investors and landlords are being drawn into the market because demand from tenants is high and they are likely to renew their leases after shouldering the costs of setting up data centers. "They invest in their space and in your space and they tend to stick around longer," said Mark Messana, president of Downtown Properties, which owns offices in Los Angeles and San Francisco. "As we all know, the office market is struggling a little bit, so it's nice to be able to have some data customers in the mix..."

Power demand for computing is growing so intense that it threatens to strain the nation's electrical grid, sending users to remote locations where power is plentiful and preferably cheap. Data center developers are working in Alabama, the Dakotas and Indiana, "traditionally states that wouldn't have data centers," Eades said.
The article includes "the mother of all data centers" in the western U.S. — a 30-story building where "thousands of miles of undersea fiber-optic cables disappear into an ordinary-looking office tower." Once a prestigious location for businesses, "The recent departure of a law firm that had been in the building more than 50 years cleared out five floors that will quickly be re-leased to data tenants, said Eades, who represents the landlord..."

To retrofit the building for data centers, "two elevators were removed so the empty shafts could hold water pipes used to help keep the temperature cool enough for the heat-producing servers" — and developers are happy rents "can be double what they are at newer downtown office high-rises, according to real estate data provider CoStar...

"By 2030, data centers could account for as much as 11% of U.S. power demand — up from 3% now, according to analysts at Goldman Sachs."

Uplevel provides insights from coding and collaboration data, according to a recent report from CIO magazine — and recently they measured "the time to merge code into a repository [and] the number of pull requests merged" for about 800 developers over a three-month period (comparing the statistics to the previous three months).

Their study "found no significant improvements for developers" using Microsoft's AI-powered coding assistant tool Copilot, according to the article (shared by Slashdot reader snydeq): Use of GitHub Copilot also introduced 41% more bugs, according to the study...

In addition to measuring productivity, the Uplevel study looked at factors in developer burnout, and it found that GitHub Copilot hasn't helped there, either. The amount of working time spent outside of standard hours decreased for both the control group and the test group using the coding tool, but it decreased more when the developers weren't using Copilot.
An Uplevel product manager/data analyst acknowledged to the magazine that there may be other ways to measure developer productivity — but they still consider their metrics solid. "We heard that people are ending up being more reviewers for this code than in the past... You just have to keep a close eye on what is being generated; does it do the thing that you're expecting it to do?"

The article also quotes the CEO of software development firm Gehtsoft, who says they didn't see major productivity gains from LLM-based coding assistants — but did see them introducing errors into code. With different prompts generating different code sections, "It becomes increasingly more challenging to understand and debug the AI-generated code, and troubleshooting becomes so resource-intensive that it is easier to rewrite the code from scratch than fix it."

On the other hand, cloud services provider Innovative Solutions saw significant productivity gains from coding assistants like Claude Dev and GitHub Copilot. And Slashdot reader destined2fail1990 says that while large/complex code bases may not see big gains, "I have seen a notable increase in productivity from using Cursor, the AI powered IDE." Yes, you have to review all the code that it generates, why wouldn't you? But often times it just works. It removes the tedious tasks like querying databases, writing model code, writing forms and processing forms, and a lot more. Some forms can have hundreds of fields and processing those fields along with doing checks for valid input is time consuming, but can be automated effectively using AI.
This prompted an interesting discussion on the original story submission. Slashdot reader bleedingobvious responded: Cursor/Claude are great BUT the code produced is almost never great quality. Even given these tools, the junior/intern teams still cannot outpace the senior devs. Great for learning, maybe, but the productivity angle not quite there.... yet.

It's damned close, though. GIve it 3-6 months.
And Slashdot reader abEeyore posted: I suspect that the results are quite a bit more nuanced than that. I expect that it is, even outside of the mentioned code review, a shift in where and how the time is spent, and not necessarily in how much time is spent.
Agree? Disagree? Share your own experiences in the comments.

And are developers really saving time with AI coding assistants?

wiredmikey writes: Three months after pulling previews of the controversial Windows Recall feature due to public backlash, Microsoft says it has completely overhauled the security architecture with proof-of-presence encryption, anti-tampering and DLP checks, and screenshot data managed in secure enclaves outside the main operating system.

In an interview with SecurityWeek, Microsoft vice president David Weston said the company's engineers rewrote the security model of Windows Recall to reduce attack surface on Copilot+ PCs and minimize the risk of malware attackers targeting the screenshot data store.

OpenAI's chief technology officer Mira Murati has announced her departure from the company, marking the latest high-profile exit from the Microsoft-backed AI firm. Murati, who briefly served as interim CEO during last year's leadership turmoil, cited a desire for personal exploration after six and a half years at OpenAI.

Her resignation follows the departures of founders Ilya Sutskever and John Schulman earlier this year. The startup, creator of ChatGPT, is currently in talks to raise over $6 billion at a $150 billion valuation, according to media reports.

Alphabet unit Google filed a complaint to the European Commission on Wednesday against what it said were Microsoft's anti-competitive practices to lock customers into Microsoft's cloud platform Azure. From a report: Google, whose biggest cloud computing rivals are Microsoft and Amazon Web Services, said Microsoft was exploiting its dominant Windows Server operating system to prevent competition. Google Cloud Vice President Amit Zavery told a briefing that Microsoft made customers pay a 400% mark-up to keep running Windows Server on rival cloud computing operators. This did not apply if they used Azure. Users of rival cloud systems would also get later and more limited security updates, Zavery said.

Google pointed to a 2023 study by cloud services organization CISPE which found that European businesses and public sector bodies were paying up to 1 billion euros ($1.12 billion) per year on Microsoft licensing penalties. Microsoft in July clinched a 20-million-euro deal to settle an antitrust complaint about its cloud computing licensing practices with CISPE, averting an EU investigation. However, the settlement did not include Amazon Web Services, Google Cloud Platform and AliCloud, prompting criticism from the first two companies.

Microsoft giveth and Microsoft taketh away, as administrators using Windows Server Update Services (WSUS) will soon find out. From a report: Windows Server 2025 remains in preview, but Microsoft has been busy letting users know what is set for removal and what will be deprecated in the release. WSUS fits into the latter category -- still there for now, but no longer under active development. This is a big deal for many administrators who rely on the feature to deploy and manage the distribution of updates and features in an enterprise environment.

It'll even work on a network disconnected from the internet -- download the patches to a connected computer, stick them on some removable media, import the patches to a WSUS server on the disconnected network, and away you go. A tame administrator told El Reg: "We are migrating to Intune. It's a lot more complicated than WSUS, and it takes a lot longer to get set up."

"Such is progress!" he sighed. Microsoft's advice is, unsurprisingly, to migrate to cloud tools. As well as the aforementioned Intune, there is also Windows Autopatch for client update management or Azure Update Manager for server update management. And there are plenty of third-party tools out there too, such as Ansible. Microsoft's announcement has attracted comment. One user said: "Congratulations, you just made centralized automated patching subject to internal politics and budget constraints. "I survived the era of Melissa, SQL Slammer, and other things that were solved when we no longer had to choose between paid patch management or trusting admins of every server to do the right thing. For those of you that did not live through that, buckle up!"

An anonymous reader quotes a report from TechCrunch: Microsoft today revealed Correction, a service that attempts to automatically revise AI-generated text that's factually wrong. Correction first flags text that may be erroneous -- say, a summary of a company's quarterly earnings call that possibly has misattributed quotes -- then fact-checks it by comparing the text with a source of truth (e.g. uploaded transcripts). Correction, available as part of Microsoft's Azure AI Content Safety API (in preview for now), can be used with any text-generating AI model, including Meta's Llama and OpenAI's GPT-4o.

"Correction is powered by a new process of utilizing small language models and large language models to align outputs with grounding documents," a Microsoft spokesperson told TechCrunch. "We hope this new feature supports builders and users of generative AI in fields such as medicine, where application developers determine the accuracy of responses to be of significant importance." Experts caution that this tool doesn't address the root cause of hallucinations. "Microsoft's solution is a pair of cross-referencing, copy-editor-esque meta models designed to highlight and rewrite hallucinations," reports TechCrunch. "A classifier model looks for possibly incorrect, fabricated, or irrelevant snippets of AI-generated text (hallucinations). If it detects hallucinations, the classifier ropes in a second model, a language model, that tries to correct for the hallucinations in accordance with specified 'grounding documents.'"

Os Keyes, a PhD candidate at the University of Washington who studies the ethical impact of emerging tech, has doubts about this. "It might reduce some problems," they said, "But it's also going to generate new ones. After all, Correction's hallucination detection library is also presumably capable of hallucinating." Mike Cook, a research fellow at Queen Mary University specializing in AI, added that the tool threatens to compound the trust and explainability issues around AI. "Microsoft, like OpenAI and Google, have created this issue where models are being relied upon in scenarios where they are frequently wrong," he said. "What Microsoft is doing now is repeating the mistake at a higher level. Let's say this takes us from 90% safety to 99% safety -- the issue was never really in that 9%. It's always going to be in the 1% of mistakes we're not yet detecting."

An anonymous reader quotes a report from Ars Technica: California Gov. Gavin Newsom vetoed a bill that would have required makers of web browsers and mobile operating systems to let consumers send opt-out preference signals that could limit businesses' use of personal information. The bill approved by the State Legislature last month would have required an opt-out signal "that communicates the consumer's choice to opt out of the sale and sharing of the consumer's personal information or to limit the use of the consumer's sensitive personal information." It would have made it illegal for a business to offer a web browser or mobile operating system without a setting that lets consumers "send an opt-out preference signal to businesses with which the consumer interacts."

In a veto message (PDF) sent to the Legislature Friday, Newsom said he would not sign the bill. Newsom wrote that he shares the "desire to enhance consumer privacy," noting that he previously signed a bill "requir[ing] the California Privacy Protection Agency to establish an accessible deletion mechanism allowing consumers to request that data brokers delete all of their personal information." But Newsom said he is opposed to the new bill's mandate on operating systems. "I am concerned, however, about placing a mandate on operating system (OS) developers at this time," the governor wrote. "No major mobile OS incorporates an option for an opt-out signal. By contrast, most Internet browsers either include such an option or, if users choose, they can download a plug-in with the same functionality. To ensure the ongoing usability of mobile devices, it's best if design questions are first addressed by developers, rather than by regulators. For this reason, I cannot sign this bill." Vetoes can be overridden with a two-thirds vote in each chamber. The bill was approved 59-12 in the Assembly and 31-7 in the Senate. But the State Legislature hasn't overridden a veto in decades. "It's troubling the power that companies such as Google appear to have over the governor's office," said Justin Kloczko, tech and privacy advocate for Consumer Watchdog, a nonprofit group in California. "What the governor didn't mention is that Google Chrome, Apple Safari and Microsoft Edge don't offer a global opt-out and they make up for nearly 90 percent of the browser market share. That's what matters. And people don't want to install plug-ins. Safari, which is the default browsers on iPhones, doesn't even accept a plug-in."

joshuark shares a report: Microsoft has officially announced that Windows Server Update Services (WSUS) is now deprecated, but plans to maintain current functionality and continue publishing updates through the channel. This move isn't surprising, as Microsoft first listed WSUS as one of the "features removed or no longer developed starting with Windows Server 2025" on August 13. In June, the company also revealed that it would also soon deprecate WSUS driver synchronization.

While new features and development for WSUS will cease, Microsoft said today that it plans to continue supporting the service's existing functionality and updates, which will still be distributed, even after deprecation. "Specifically, this means that we are no longer investing in new capabilities, nor are we accepting new feature requests for WSUS," Microsoft's Nir Froimovici said on Friday. "However, we are preserving current functionality and will continue to publish updates through the WSUS channel. We will also support any content already published through the WSUS channel."

Microsoft unveiled detailed security reforms Monday, five months after CEO Satya Nadella pledged to prioritize cybersecurity following major breaches. The 25-page Secure Future Initiative report [PDF] outlines technical and governance changes addressing criticisms in an April 2024 Cyber Safety Review Board report that deemed Microsoft's security culture "inadequate."

Microsoft said it implemented significant security upgrades to its Entra ID and Microsoft Account systems, introducing Azure-managed hardware security modules for access token signing keys. The company has also purged 5.75 million inactive tenants to minimize potential attack vectors and adopted a new testing system with secure defaults to prevent legacy-related security issues. Concurrently, Microsoft has enhanced its network tracking capabilities, now monitoring over 99 percent of its physical network through a centralized inventory system, which aids in firmware compliance and logging.

Internal security measures have been tightened, with engineering teams facing stricter access controls. Personal access tokens are now limited to seven days, SSH access has been disabled for internal engineering repositories, and access to critical engineering systems has been restricted to fewer groups. Additionally, Microsoft has extended its audit log retention period to a minimum of two years, bolstering its ability to investigate and respond to potential security incidents.

Marc Benioff said Microsoft's Copilot AI hasn't lived up to the hype. The Salesforce CEO said on the company's second-quarter earnings call that its own AI is nothing like Copilot, which he said was unimpressive. From a report: "So many customers are so disappointed in what they bought from Microsoft Copilot because they're not getting the accuracy and the response that they want," Benioff said. "Microsoft has disappointed so many customers with AI."

Microsoft Copilot integrates OpenAI's ChatGPT tech into the company's existing suite of business software like Word, Excel, and PowerPoint that comes with Microsoft 365. Launched last year, Copilot is meant to help companies boost productivity by responding to employee prompts and helping them with daily tasks like scheduling meetings, writing up product announcements, and creating presentations. In response to Benioff's comments, Jared Spataro, Microsoft's corporate vice president for AI at work, said in a statement to Fortune that the company was "hearing something quite different" from its customers.

Nature published an introduction to running an LLM locally, starting with the example of a bioinformatician who's using AI to generate readable summaries for his database of immune-system protein structures. "But he doesn't use ChatGPT, or any other web-based LLM." He just runs the AI on his Mac... Two more recent trends have blossomed. First, organizations are making 'open weights' versions of LLMs, in which the weights and biases used to train a model are publicly available, so that users can download and run them locally, if they have the computing power. Second, technology firms are making scaled-down versions that can be run on consumer hardware — and that rival the performance of older, larger models. Researchers might use such tools to save money, protect the confidentiality of patients or corporations, or ensure reproducibility... As computers get faster and models become more efficient, people will increasingly have AIs running on their laptops or mobile devices for all but the most intensive needs. Scientists will finally have AI assistants at their fingertips — but the actual algorithms, not just remote access to them.
The article's list of small open-weights models includes Meta's Llama, Google DeepMind's Gemma, Alibaba's Qwen, Apple's DCLM, Mistral's NeMo, and OLMo from the Allen Institute for AI. And then there's Microsoft: Although the California tech firm OpenAI hasn't open-weighted its current GPT models, its partner Microsoft in Redmond, Washington, has been on a spree, releasing the small language models Phi-1, Phi-1.5 and Phi-2 in 2023, then four versions of Phi-3 and three versions of Phi-3.5 this year. The Phi-3 and Phi-3.5 models have between 3.8 billion and 14 billion active parameters, and two models (Phi-3-vision and Phi-3.5-vision) handle images1. By some benchmarks, even the smallest Phi model outperforms OpenAI's GPT-3.5 Turbo from 2023, rumoured to have 20 billion parameters... Microsoft used LLMs to write millions of short stories and textbooks in which one thing builds on another. The result of training on this text, says Sébastien Bubeck, Microsoft's vice-president for generative AI, is a model that fits on a mobile phone but has the power of the initial 2022 version of ChatGPT. "If you are able to craft a data set that is very rich in those reasoning tokens, then the signal will be much richer," he says...

Sharon Machlis, a former editor at the website InfoWorld, who lives in Framingham, Massachusetts, wrote a guide to using LLMs locally, covering a dozen options.
The bioinformatician shares another benefit: you don't have to worry about the company updating their models (leading to different outputs). "In most of science, you want things that are reproducible. And it's always a worry if you're not in control of the reproducibility of what you're generating."

And finally, the article reminds readers that "Researchers can build on these tools to create custom applications..." Whichever approach you choose, local LLMs should soon be good enough for most applications, says Stephen Hood, who heads open-source AI at the tech firm Mozilla in San Francisco. "The rate of progress on those over the past year has been astounding," he says. As for what those applications might be, that's for users to decide. "Don't be afraid to get your hands dirty," Zakka says. "You might be pleasantly surprised by the results."