microsoft

microsoft’s-phi-3-shows-the-surprising-power-of-small,-locally-run-ai-language-models

Microsoft’s Phi-3 shows the surprising power of small, locally run AI language models

AI, Biz & IT, chatgpt, chatgtp, large language models, machine learning, microsoft, Microsoft Phi, Mistral, openai, Phi-2, Phi-3, Simon Willison, small language models / /

small packages —

Microsoft’s 3.8B parameter Phi-3 may rival GPT-3.5, signaling a new era of “small language models.”

–

An illustration of lots of information being compressed into a smartphone with a funnel.

Getty Images

On Tuesday, Microsoft announced a new, freely available lightweight AI language model named Phi-3-mini, which is simpler and less expensive to operate than traditional large language models (LLMs) like OpenAI’s GPT-4 Turbo. Its small size is ideal for running locally, which could bring an AI model of similar capability to the free version of ChatGPT to a smartphone without needing an Internet connection to run it.

The AI field typically measures AI language model size by parameter count. Parameters are numerical values in a neural network that determine how the language model processes and generates text. They are learned during training on large datasets and essentially encode the model’s knowledge into quantified form. More parameters generally allow the model to capture more nuanced and complex language-generation capabilities but also require more computational resources to train and run.

Some of the largest language models today, like Google’s PaLM 2, have hundreds of billions of parameters. OpenAI’s GPT-4 is rumored to have over a trillion parameters but spread over eight 220-billion parameter models in a mixture-of-experts configuration. Both models require heavy-duty data center GPUs (and supporting systems) to run properly.

In contrast, Microsoft aimed small with Phi-3-mini, which contains only 3.8 billion parameters and was trained on 3.3 trillion tokens. That makes it ideal to run on consumer GPU or AI-acceleration hardware that can be found in smartphones and laptops. It’s a follow-up of two previous small language models from Microsoft: Phi-2, released in December, and Phi-1, released in June 2023.

A chart provided by Microsoft showing Phi-3 performance on various benchmarks.

Enlarge / A chart provided by Microsoft showing Phi-3 performance on various benchmarks.

Phi-3-mini features a 4,000-token context window, but Microsoft also introduced a 128K-token version called “phi-3-mini-128K.” Microsoft has also created 7-billion and 14-billion parameter versions of Phi-3 that it plans to release later that it claims are “significantly more capable” than phi-3-mini.

Microsoft says that Phi-3 features overall performance that “rivals that of models such as Mixtral 8x7B and GPT-3.5,” as detailed in a paper titled “Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone.” Mixtral 8x7B, from French AI company Mistral, utilizes a mixture-of-experts model, and GPT-3.5 powers the free version of ChatGPT.

“[Phi-3] looks like it’s going to be a shockingly good small model if their benchmarks are reflective of what it can actually do,” said AI researcher Simon Willison in an interview with Ars. Shortly after providing that quote, Willison downloaded Phi-3 to his Macbook laptop locally and said, “I got it working, and it’s GOOD” in a text message sent to Ars.

A screenshot of Phi-3-mini running locally on Simon Willison's Macbook.

Enlarge / A screenshot of Phi-3-mini running locally on Simon Willison’s Macbook.

Simon Willison

“Most models that run on a local device still need hefty hardware,” says Willison. “Phi-3-mini runs comfortably with less than 8GB of RAM, and can churn out tokens at a reasonable speed even on just a regular CPU. It’s licensed MIT and should work well on a $55 Raspberry Pi—and the quality of results I’ve seen from it so far are comparable to models 4x larger.“

How did Microsoft cram a capability potentially similar to GPT-3.5, which has at least 175 billion parameters, into such a small model? Its researchers found the answer by using carefully curated, high-quality training data they initially pulled from textbooks. “The innovation lies entirely in our dataset for training, a scaled-up version of the one used for phi-2, composed of heavily filtered web data and synthetic data,” writes Microsoft. “The model is also further aligned for robustness, safety, and chat format.”

Much has been written about the potential environmental impact of AI models and datacenters themselves, including on Ars. With new techniques and research, it’s possible that machine learning experts may continue to increase the capability of smaller AI models, replacing the need for larger ones—at least for everyday tasks. That would theoretically not only save money in the long run but also require far less energy in aggregate, dramatically decreasing AI’s environmental footprint. AI models like Phi-3 may be a step toward that future if the benchmark results hold up to scrutiny.

Phi-3 is immediately available on Microsoft’s cloud service platform Azure, as well as through partnerships with machine learning model platform Hugging Face and Ollama, a framework that allows models to run locally on Macs and PCs.

Microsoft’s Phi-3 shows the surprising power of small, locally run AI language models Read More »

windows-vulnerability-reported-by-the-nsa-exploited-to-install-russian-malware

Windows vulnerability reported by the NSA exploited to install Russian malware

Biz & IT, exploits, microsoft, Security, vulnerabilities, Windows / /
Windows vulnerability reported by the NSA exploited to install Russian malware

Getty Images

Kremlin-backed hackers have been exploiting a critical Microsoft vulnerability for four years in attacks that targeted a vast array of organizations with a previously undocumented tool, the software maker disclosed Monday.

When Microsoft patched the vulnerability in October 2022—at least two years after it came under attack by the Russian hackers—the company made no mention that it was under active exploitation. As of publication, the company’s advisory still made no mention of the in-the-wild targeting. Windows users frequently prioritize the installation of patches based on whether a vulnerability is likely to be exploited in real-world attacks.

Exploiting CVE-2022-38028, as the vulnerability is tracked, allows attackers to gain system privileges, the highest available in Windows, when combined with a separate exploit. Exploiting the flaw, which carries a 7.8 severity rating out of a possible 10, requires low existing privileges and little complexity. It resides in the Windows print spooler, a printer-management component that has harbored previous critical zero-days. Microsoft said at the time that it learned of the vulnerability from the US National Security Agency.

On Monday, Microsoft revealed that a hacking group tracked under the name Forest Blizzard has been exploiting CVE-2022-38028 since at least June 2020—and possibly as early as April 2019. The threat group—which is also tracked under names including APT28, Sednit, Sofacy, GRU Unit 26165, and Fancy Bear—has been linked by the US and the UK governments to Unit 26165 of the Main Intelligence Directorate, a Russian military intelligence arm better known as the GRU. Forest Blizzard focuses on intelligence gathering through the hacking of a wide array of organizations, mainly in the US, Europe, and the Middle East.

Since as early as April 2019, Forest Blizzard has been exploiting CVE-2022-38028 in attacks that, once system privileges are acquired, use a previously undocumented tool that Microsoft calls GooseEgg. The post-exploitation malware elevates privileges within a compromised system and goes on to provide a simple interface for installing additional pieces of malware that also run with system privileges. This additional malware, which includes credential stealers and tools for moving laterally through a compromised network, can be customized for each target.

“While a simple launcher application, GooseEgg is capable of spawning other applications specified at the command line with elevated permissions, allowing threat actors to support any follow-on objectives such as remote code execution, installing a backdoor, and moving laterally through compromised networks,” Microsoft officials wrote.

GooseEgg is typically installed using a simple batch script, which is executed following the successful exploitation of CVE-2022-38028 or another vulnerability, such as CVE-2023-23397, which Monday’s advisory said has also been exploited by Forest Blizzard. The script is responsible for installing the GooseEgg binary, often named justice.exe or DefragmentSrv.exe, then ensuring that they run each time the infected machine is rebooted.

Windows vulnerability reported by the NSA exploited to install Russian malware Read More »

after-48-years,-zilog-is-killing-the-classic-standalone-z80-microprocessor-chip

After 48 years, Zilog is killing the classic standalone Z80 microprocessor chip

CP/M, microprocessors, microsoft, nintendo game boy, retrotech, Tech, TI-81, ti-83, Z80, Zilog, zx spectrum / /

rest in silicon —

Z80 powered game consoles, ZX Spectrum, Pac-Man, and a 1970s PC standard based on CP/M.

–

A cropped portion of a ca. 1980 ad for the Microsoft Z80 SoftCard, which allowed Apple II users to run the CP/M operating system.

Enlarge / A cropped portion of a ca. 1980 ad for the Microsoft Z80 SoftCard, which allowed Apple II users to run the CP/M operating system.

Microsoft

Last week, chip manufacturer Zilog announced that after 48 years on the market, its line of standalone DIP (dual inline package) Z80 CPUs is coming to an end, ceasing sales on June 14, 2024. The 8-bit Z80 architecture debuted in 1976 and powered a small-business-PC revolution in conjunction with CP/M, also serving as the heart of the Nintendo Game Boy, Sinclair ZX Spectrum, the Radio Shack TRS-80, the Pac-Man arcade game, and the TI-83 graphing calculator in various forms.

In a letter to customers dated April 15, 2024, Zilog wrote, “Please be advised that our Wafer Foundry Manufacturer will be discontinuing support for the Z80 product and other product lines. Refer to the attached list of the Z84C00 Z80 products affected.”

Designers typically use the Z84C00 chips because of familiarity with the Z80 architecture or to allow legacy system upgrades without needing significant system redesigns. And while many other embedded chip architectures have superseded these Z80 chips in speed, processing power, and capability, they remained go-to solutions for decades in products that didn’t need any extra horsepower.

Zilog will continue to manufacture the eZ80 microcontroller family, which was introduced in 2001 as a faster version of the Z80 series and comes in different physical package configurations (pin layouts).

Powering a microcomputer revolution

The 8-bit Z80 microprocessor was designed in 1974 by Federico Faggin as a binary-compatible, improved version of the Intel 8080 with a higher clock speed, a built-in DRAM refresh controller, and an extended instruction set. It was extensively used in desktop computers of the late 1970s and early 1980s, arcade video game machines, and embedded systems, and it became a cornerstone of several gaming consoles, like the Sega Master System.

The Tandy Radio Shack TRS-80 (1977), which used the Zilog Z80.

Enlarge / The Tandy Radio Shack TRS-80 (1977), which used the Zilog Z80.

SSPL/Getty Images

During the mid-late 1970s, the Z80 became a popular CPU for S-100 bus machines, which were early personal computers with a 100-pin modular bus system that allowed swapping cards to build systems based on parts from various manufacturers. Digital Research targeted the Z80 as a key platform for its CP/M operating system, and the association between Z80 and CP/M stuck, powering dozens of small business computers until the mid-1980s, when IBM PC clones running Microsoft’s MS-DOS became the new industry standard.

Interestingly, Microsoft’s first hardware product, the Z80 SoftCard for the Apple II in 1980, added the famous Zilog CPU to the classic personal computer and allowed users to run CP/M on that machine. In 1982, Bill Gates claimed that SoftCard installations represented the largest single user base of CP/M machines.

Last call in June 2024

Zilog is notably discontinuing several Z84C00 chips that are still available in classic 40-pin DIP packages resembling the classic Z80 CPU chips of the 1970s. (These standalone chips include a CPU and nothing else, unlike a microcontroller, which can include RAM and other accessory devices.) The DIP design features two rows of 20 pins with a plastic package in between that contains the actual embedded silicon chip, resembling the classic Z80 CPU chips of the 1970s.

After June 14, Zilog will stop taking orders, manufacture whatever orders are available if they are sufficient in quantity, then ship the last runs of the chips to resellers like Mouser Electronics and Digikey.

A classic dual inline package (DIP) version of the Z80 from the 1970s. It features two rows of 20 pins in a ceramic package.

Enlarge / A classic dual inline package (DIP) version of the Z80 from the 1970s. It features two rows of 20 pins in a ceramic package.

The discontinuation list provided by Zilog in its letter includes 13 products from the Z84C00 series, which are chips in the Z80 family that run at clock speeds from 6 to 20 MHz and maintain compatibility with the original Z80 architecture. Here’s the full list of part numbers that will be discontinued:

  • Z84C0006VEG
  • Z84C0006PEG
  • Z84C0010PEG
  • Z84C0008AEG
  • Z84C0020VEG
  • Z84C0008PEG
  • Z84C0010AEG
  • Z84C0008VEG
  • Z84C0010VEG
  • Z84C0010VEG00TR
  • Z84C0020AEG
  • Z84C0020PEG
  • Z84C0006AEG

So while the Z80 architecture will stick around in eZ80 form, it appears that this is the last call for newly manufactured standalone 8-bit Z80 CPU chips in the classic DIP form factor. We reached out to Zilog for clarification about its plans for the future of the Z80 platform but did not receive a response by press time.

After 48 years, Zilog is killing the classic standalone Z80 microprocessor chip Read More »

microsoft’s-vasa-1-can-deepfake-a-person-with-one-photo-and-one-audio-track

Microsoft’s VASA-1 can deepfake a person with one photo and one audio track

AI, AI ethics, AI safety, Biz & IT, Deepfakes, image synthesis, machine learning, microsoft, VASA-1, video synthesis / /

pics and it didn’t happen —

YouTube videos of 6K celebrities helped train AI model to animate photos in real time.

–

A sample image from Microsoft for

Enlarge / A sample image from Microsoft for “VASA-1: Lifelike Audio-Driven Talking Faces Generated in Real Time.”

On Tuesday, Microsoft Research Asia unveiled VASA-1, an AI model that can create a synchronized animated video of a person talking or singing from a single photo and an existing audio track. In the future, it could power virtual avatars that render locally and don’t require video feeds—or allow anyone with similar tools to take a photo of a person found online and make them appear to say whatever they want.

“It paves the way for real-time engagements with lifelike avatars that emulate human conversational behaviors,” reads the abstract of the accompanying research paper titled, “VASA-1: Lifelike Audio-Driven Talking Faces Generated in Real Time.” It’s the work of Sicheng Xu, Guojun Chen, Yu-Xiao Guo, Jiaolong Yang, Chong Li, Zhenyu Zang, Yizhong Zhang, Xin Tong, and Baining Guo.

The VASA framework (short for “Visual Affective Skills Animator”) uses machine learning to analyze a static image along with a speech audio clip. It is then able to generate a realistic video with precise facial expressions, head movements, and lip-syncing to the audio. It does not clone or simulate voices (like other Microsoft research) but relies on an existing audio input that could be specially recorded or spoken for a particular purpose.

Microsoft claims the model significantly outperforms previous speech animation methods in terms of realism, expressiveness, and efficiency. To our eyes, it does seem like an improvement over single-image animating models that have come before.

AI research efforts to animate a single photo of a person or character extend back at least a few years, but more recently, researchers have been working on automatically synchronizing a generated video to an audio track. In February, an AI model called EMO: Emote Portrait Alive from Alibaba’s Institute for Intelligent Computing research group made waves with a similar approach to VASA-1 that can automatically sync an animated photo to a provided audio track (they call it “Audio2Video”).

Trained on YouTube clips

Microsoft Researchers trained VASA-1 on the VoxCeleb2 dataset created in 2018 by three researchers from the University of Oxford. That dataset contains “over 1 million utterances for 6,112 celebrities,” according to the VoxCeleb2 website, extracted from videos uploaded to YouTube. VASA-1 can reportedly generate videos of 512×512 pixel resolution at up to 40 frames per second with minimal latency, which means it could potentially be used for realtime applications like video conferencing.

To show off the model, Microsoft created a VASA-1 research page featuring many sample videos of the tool in action, including people singing and speaking in sync with pre-recorded audio tracks. They show how the model can be controlled to express different moods or change its eye gaze. The examples also include some more fanciful generations, such as Mona Lisa rapping to an audio track of Anne Hathaway performing a “Paparazzi” song on Conan O’Brien.

The researchers say that, for privacy reasons, each example photo on their page was AI-generated by StyleGAN2 or DALL-E 3 (aside from the Mona Lisa). But it’s obvious that the technique could equally apply to photos of real people as well, although it’s likely that it will work better if a person appears similar to a celebrity present in the training dataset. Still, the researchers say that deepfaking real humans is not their intention.

“We are exploring visual affective skill generation for virtual, interactive charactors [sic], NOT impersonating any person in the real world. This is only a research demonstration and there’s no product or API release plan,” reads the site.

While the Microsoft researchers tout potential positive applications like enhancing educational equity, improving accessibility, and providing therapeutic companionship, the technology could also easily be misused. For example, it could allow people to fake video chats, make real people appear to say things they never actually said (especially when paired with a cloned voice track), or allow harassment from a single social media photo.

Right now, the generated video still looks imperfect in some ways, but it could be fairly convincing for some people if they did not know to expect an AI-generated animation. The researchers say they are aware of this, which is why they are not openly releasing the code that powers the model.

“We are opposed to any behavior to create misleading or harmful contents of real persons, and are interested in applying our technique for advancing forgery detection,” write the researchers. “Currently, the videos generated by this method still contain identifiable artifacts, and the numerical analysis shows that there’s still a gap to achieve the authenticity of real videos.”

VASA-1 is only a research demonstration, but Microsoft is far from the only group developing similar technology. If the recent history of generative AI is any guide, it’s potentially only a matter of time before similar technology becomes open source and freely available—and they will very likely continue to improve in realism over time.

Microsoft’s VASA-1 can deepfake a person with one photo and one audio track Read More »

llms-keep-leaping-with-llama-3,-meta’s-newest-open-weights-ai-model

LLMs keep leaping with Llama 3, Meta’s newest open-weights AI model

AI, AI assistants, Biz & IT, chatgpt, chatgtp, copilot, Google Gemini, LLaMA, Llama 2, Llama 3, machine learning, Meta, Meta AI Assistant, microsoft, Microsoft Copilot, Open Source, open weights, openai / /

computer-powered word generator —

Zuckerberg says new AI model “was still learning” when Meta stopped training.

–

A group of pink llamas on a pixelated background.

On Thursday, Meta unveiled early versions of its Llama 3 open-weights AI model that can be used to power text composition, code generation, or chatbots. It also announced that its Meta AI Assistant is now available on a website and is going to be integrated into its major social media apps, intensifying the company’s efforts to position its products against other AI assistants like OpenAI’s ChatGPT, Microsoft’s Copilot, and Google’s Gemini.

Like its predecessor, Llama 2, Llama 3 is notable for being a freely available, open-weights large language model (LLM) provided by a major AI company. Llama 3 technically does not quality as “open source” because that term has a specific meaning in software (as we have mentioned in other coverage), and the industry has not yet settled on terminology for AI model releases that ship either code or weights with restrictions (you can read Llama 3’s license here) or that ship without providing training data. We typically call these releases “open weights” instead.

At the moment, Llama 3 is available in two parameter sizes: 8 billion (8B) and 70 billion (70B), both of which are available as free downloads through Meta’s website with a sign-up. Llama 3 comes in two versions: pre-trained (basically the raw, next-token-prediction model) and instruction-tuned (fine-tuned to follow user instructions). Each has a 8,192 token context limit.

A screenshot of the Meta AI Assistant website on April 18, 2024.

Enlarge / A screenshot of the Meta AI Assistant website on April 18, 2024.

Benj Edwards

Meta trained both models on two custom-built, 24,000-GPU clusters. In a podcast interview with Dwarkesh Patel, Meta CEO Mark Zuckerberg said that the company trained the 70B model with around 15 trillion tokens of data. Throughout the process, the model never reached “saturation” (that is, it never hit a wall in terms of capability increases). Eventually, Meta pulled the plug and moved on to training other models.

“I guess our prediction going in was that it was going to asymptote more, but even by the end it was still leaning. We probably could have fed it more tokens, and it would have gotten somewhat better,” Zuckerberg said on the podcast.

Meta also announced that it is currently training a 400B parameter version of Llama 3, which some experts like Nvidia’s Jim Fan think may perform in the same league as GPT-4 Turbo, Claude 3 Opus, and Gemini Ultra on benchmarks like MMLU, GPQA, HumanEval, and MATH.

Speaking of benchmarks, we have devoted many words in the past to explaining how frustratingly imprecise benchmarks can be when applied to large language models due to issues like training contamination (that is, including benchmark test questions in the training dataset), cherry-picking on the part of vendors, and an inability to capture AI’s general usefulness in an interactive session with chat-tuned models.

But, as expected, Meta provided some benchmarks for Llama 3 that list results from MMLU (undergraduate level knowledge), GSM-8K (grade-school math), HumanEval (coding), GPQA (graduate-level questions), and MATH (math word problems). These show the 8B model performing well compared to open-weights models like Google’s Gemma 7B and Mistral 7B Instruct, and the 70B model also held its own against Gemini Pro 1.5 and Claude 3 Sonnet.

A chart of instruction-tuned Llama 3 8B and 70B benchmarks provided by Meta.

Enlarge / A chart of instruction-tuned Llama 3 8B and 70B benchmarks provided by Meta.

Meta says that the Llama 3 model has been enhanced with capabilities to understand coding (like Llama 2) and, for the first time, has been trained with both images and text—though it currently outputs only text. According to Reuters, Meta Chief Product Officer Chris Cox noted in an interview that more complex processing abilities (like executing multi-step plans) are expected in future updates to Llama 3, which will also support multimodal outputs—that is, both text and images.

Meta plans to host the Llama 3 models on a range of cloud platforms, making them accessible through AWS, Databricks, Google Cloud, and other major providers.

Also on Thursday, Meta announced that Llama 3 will become the new basis of the Meta AI virtual assistant, which the company first announced in September. The assistant will appear prominently in search features for Facebook, Instagram, WhatsApp, Messenger, and the aforementioned dedicated website that features a design similar to ChatGPT, including the ability to generate images in the same interface. The company also announced a partnership with Google to integrate real-time search results into the Meta AI assistant, adding to an existing partnership with Microsoft’s Bing.

LLMs keep leaping with Llama 3, Meta’s newest open-weights AI model Read More »

german-state-gov.-ditching-windows-for-linux,-30k-workers-migrating

German state gov. ditching Windows for Linux, 30K workers migrating

Biz & IT, germany, LibreOffice, Linux, microsoft, Software, Windows / /

Open source FTW —

Schleswig-Holstein looks to succeed where Munich failed.

–

many penguins

Schleswig-Holstein, one of Germany’s 16 states, on Wednesday confirmed plans to move tens of thousands of systems from Microsoft Windows to Linux. The announcement follows previously established plans to migrate the state government off Microsoft Office in favor of open source LibreOffice.

As spotted by The Document Foundation, the government has apparently finished its pilot run of LibreOffice and is now announcing plans to expand to more open source offerings.

In 2021, the state government announced plans to move 25,000 computers to LibreOffice by 2026. At the time, Schleswig-Holstein said it had already been testing LibreOffice for two years.

As announced on Minister-President Daniel Gunther’s webpage this week, the state government confirmed that it’s moving all systems to the Linux operating system (OS), too. Per a website-provided translation:

With the cabinet decision, the state government has made the concrete beginning of the switch away from proprietary software and towards free, open-source systems and digitally sovereign IT workplaces for the state administration’s approximately 30,000 employees.

The state government is offering a training program that it said it will update as necessary.

Regarding LibreOffice, the government maintains the possibility that some jobs may use software so specialized that they won’t be able to move to open source software.

In 2021, Jan Philipp Albrecht, then-minister for Energy, Agriculture, the Environment, Nature, and Digitalization of Schleswig-Holstein, discussed interest in moving the state government off of Windows.

“Due to the high hardware requirements of Windows 11, we would have a problem with older computers. With Linux we don’t have that,” Albrecht told Heise magazine, per a Google translation.

This week’s announcement also said that the Schleswig-Holstein government will ditch Microsoft Sharepoint and Exchange/Outlook in favor of open source offerings Nextcloud and Open-Xchange, and Mozilla Thunderbird in conjunction with the Univention active directory connector.

Schleswig-Holstein is also developing an open source directory service to replace Microsoft’s Active Directory and an open source telephony offering.

Digital sovereignty dreams

Explaining the decision, the Schleswig-Holstein government’s announcement named enhanced IT security, cost efficiencies, and collaboration between different systems as its perceived benefits of switching to open source software.

Further, the government is pushing the idea of digital sovereignty, with Schleswig-Holstein Digitalization Minister Dirk Schrödter quoted in the announcement as comparing the concept’s value to that of energy sovereignty. The announcement also quoted Schrödter as saying that digital sovereignty isn’t achievable “with the current standard IT workplace products.”

Schrödter pointed to the state government’s growing reliance on cloud services and said that with related proprietary software, users have no influence on data flow and whether that data makes its way to other countries.

Schrödter also claimed that the move would help with the state’s budget by diverting money from licensing fees to “real programming services from our domestic digital economy” that could also create local jobs.

In 2021, Albrecht said the state was reaching its limits with proprietary software contracts because “license fees have continued to rise in recent years,” per Google’s translation.

“Secondly, regarding our goals for the digitalization of administration, open source simply offers us more flexibility,” he added.

At the time, Albrecht claimed that 90 percent of video conferences in the state government ran on the open source program Jitsi, which was advantageous during the COVID-19 pandemic because the state was able to quickly increase video conferencing capacity.

Additionally, he said that because the school portal was based on (unnamed) open source software, “we can design the interface flexibly and combine services the way we want.”

There are numerous other examples globally of government entities switching to Linux in favor of open source technology. Federal governments with particular interest in avoiding US-based technologies, including North Korea and China, are some examples. The South Korean government has also shared plans to move to Linux by 2026, and the city of Barcelona shared migration plans in 2018.

But some government bodies that have made the move regretted it and ended up crawling back to Windows. Vienna released the Debian-based distribution WIENUX in 2005 but gave up on migration by 2009.

In 2003, Munich announced it would be moving some 14,000 PCs off Windows and to Linux. In 2013, the LiMux project finished, but high associated costs and user dissatisfaction resulted in Munich announcing in 2017 that it would spend the next three years reverting back to Windows.

Albrecht in 2021 addressed this failure when speaking to Heise, saying, per Google’s translation:

The main problem there was that the employees weren’t sufficiently involved. We do that better. We are planning long transition phases with parallel use. And we are introducing open source step by step where the departments are ready for it. This also creates the reason for further rollout because people see that it works.

German state gov. ditching Windows for Linux, 30K workers migrating Read More »

copilot-key-is-based-on-a-button-you-probably-haven’t-seen-since-ibm’s-model-m

Copilot key is based on a button you probably haven’t seen since IBM’s Model M

AI, copilot, generative ai, Keyboards, laptops, microsoft, Tech, Windows, Windows 11 / /

Microsoft chatbot button —

Left-Shift + Windows key + F23

–

A Dell XPS 14 laptop with a Copilot key.

Enlarge / A Dell XPS 14 laptop. The Copilot key is to the right of the right-Alt button.

In January, Microsoft introduced a new key to Windows PC keyboards for the first time in 30 years. The Copilot key, dedicated to launching Microsoft’s eponymous generative AI assistant, is already on some Windows laptops released this year. On Monday, Tom’s Hardware dug into the new addition and determined exactly what pressing the button does, which is actually pretty simple. Pushing a computer’s integrated Copilot button is like pressing left-Shift + Windows key + F23 simultaneously.

Tom’s Hardware confirmed this after wondering if the Copilot key introduced a new scan code to Windows or if it worked differently. Using the scripting program AuthoHotkey with a new laptop with a Copilot button, Tom’s Hardware discovered the keystrokes registered when a user presses the Copilot key. The publication confirmed with Dell that “this key assignment is standard for the Copilot key and done at Microsoft’s direction.”

F23

Surprising to see in that string of keys is F23. Having a computer keyboard with a function row or rows that take you from F1 all the way to F23 is quite rare today. When I try to imagine a keyboard that comes with an F23 button, vintage keyboards come to mind, more specifically buckling spring keyboards from IBM.

IBM’s Model F, which debuted in 1981 and used buckling spring switches over a capacitive PCB, and the Model M, which launched in 1985 and used buckling spring switches over a membrane sheet, both offered layouts with 122 keys. These layouts included not one, but two rows of function keys that would leave today’s 60 percent keyboard fans sweating over the wasted space.

But having 122 keys was helpful for keyboards tied to IBM business terminals. The keyboard layout even included a bank of keys to the left of the primary alpha block of keys for even more forms of input.

An IBM Model M keyboard with an F23 key.

Enlarge / An IBM Model M keyboard with an F23 key.

The 122-key keyboard layout with F23 lives on. Beyond people who still swear by old Model F and M keyboards, Model F Labs and Unicomp both currently sell modern buckling spring keyboards with built-in F23 buttons. Another reason a modern Windows PC user might have access to an F23 key is if they use a macro pad.

But even with those uses in mind, the F23 key remains rare. That helps explain why Microsoft would use the key for launching Copilot; users are unlikely to have F23 programmed for other functions. This was also likely less work than making a key with an entirely new scan code.

The Copilot button is reprogrammable

When I previewed Dell’s 2024 XPS laptops, a Dell representative told me that the integrated Copilot key wasn’t reprogrammable. However, in addition to providing some interesting information about the newest PC key since the Windows button, Tom’s Hardware’s revelation shows why the Copilot key is actually reprogrammable, even if OEMs don’t give users a way to do so out of the box. (If you need help, check out the website’s tutorial for reprogramming the Windows Copilot key.)

I suspect there’s a strong interest in reprogramming that button. For one, generative AI, despite all its hype and potential, is still an emerging technology. Many don’t need or want access to any chatbot—let alone Microsoft’s—instantly or even at all. Those who don’t use their system with a Microsoft account have no use for the button, since being logged in to a Microsoft account is required for the button to launch Copilot.

A rendering of the Copilot button.

Enlarge / A rendering of the Copilot button.

Microsoft

Additionally, there are other easy ways to launch Copilot on a computer that has the program downloaded, like double-clicking an icon or pressing Windows + C, that make a dedicated button unnecessary. (Ars Technica asked Microsoft why the Copilot key doesn’t just register Windows + C, but the company declined to comment. Windows + C has launched other apps in the past, including Cortana, so it’s possible that Microsoft wanted to avoid the Copilot key performing a different function when pressed on computers that use Windows images without Copilot.)

In general, shoehorning the Copilot key into Windows laptops seems premature. Copilot is young and still a preview; just a few months ago, it was called Bing Chat. Further, the future of generative AI, including its popularity and top uses, is still forming and could evolve substantially during the lifetime of a Windows laptop. Microsoft’s generative AI efforts could also flounder over the years. Imagine if Microsoft went all-in on Bing back in the day and made all Windows keyboards have a Bing button, for example. Just because Microsoft wants something to become mainstream doesn’t mean that it will.

This all has made the Copilot button seem more like a way to force the adoption of Microsoft’s chatbot than a way to improve Windows keyboards. Microsoft has also made the Copilot button a requirement for its AI PC certification (which also requires an integrated neural processing unit and having Copilot pre-installed). Microsoft plans to make Copilot keys a requirement for Windows 11 OEM PCs eventually, it told Ars Technica in January.

At least for now, the basic way that the Copilot button works means you can turn the key into something more useful. Now, the tricky part would be finding a replacement keycap to eradicate Copilot’s influence from your keyboard.

Listing image by Microsoft

Copilot key is based on a button you probably haven’t seen since IBM’s Model M Read More »

microsoft-splits-up-the-teams-and-office-apps-worldwide,-following-eu-split

Microsoft splits up the Teams and Office apps worldwide, following EU split

Biz & IT, microsoft, microsoft 365, Microsoft office, Microsoft Teams, office, Teams / /

different teams —

Changes may save a bit of money for people who want Office apps without Teams.

– Updated

Teams is being decoupled from the other Office apps worldwide, six months after Microsoft did the same thing for the EU.

Enlarge / Teams is being decoupled from the other Office apps worldwide, six months after Microsoft did the same thing for the EU.

Microsoft/Andrew Cunningham

Months after unbundling the apps in the European Union, Microsoft is taking the Office and Teams breakup worldwide. Reuters reports that Microsoft will begin selling Teams and the other Microsoft 365 apps to new commercial customers as separate products with separate price tags beginning today.

“To ensure clarity for our customers, we are extending the steps we took last year to unbundle Teams from M365 and O365 in the European Economic Area and Switzerland to customers globally,” a Microsoft spokesperson told Ars. “Doing so also addresses feedback from the European Commission by providing multinational companies more flexibility when they want to standardize their purchasing across geographies.”

The unbundling is a win for other team communication apps like Slack and videoconferencing apps like Zoom, both of which predate Teams but haven’t had the benefits of the Office apps’ huge established user base.

The separation follows an EU regulatory investigation that started in July of 2023, almost exactly three years after Slack initially filed a complaint alleging that Microsoft was “abusing its market dominance to extinguish competition in breach of European Union competition law.”

In August of 2023, Microsoft announced that it would be unbundling the apps in the EU and Switzerland in October. Bloomberg reported in September that Zoom had met with EU and US Federal Trade Commission regulators about Microsoft, further ratcheting up regulatory pressure on Microsoft.

In October, Microsoft European Government Affairs VP Nanna-Louise Linde described the unbundling and other moves as “proactive changes that we hope will start to address these concerns in a meaningful way,” though the EU investigation is ongoing, and the company may yet be fined. Linde also wrote that Microsoft would allow third-party apps like Zoom and Slack to integrate more deeply with the Office apps and that it would “enable third-party solutions to host Office web applications.”

Microsoft has put up a blog post detailing its new pricing structure here—for now, the changes only affect the Microsoft 365 plans for the Business, Enterprise, and Frontline versions of Microsoft 365. Consumer, Academic, US Government, and Nonprofit editions of Microsoft 365 aren’t changing today and will still bundle Teams as they did before.

Current Office/Microsoft 365 Enterprise customers who want to keep using the Office apps and Teams together can continue to subscribe to both at their current prices. New subscribers to the Enterprise versions of Microsoft 365/Office 365 can pay $5.25 per user per month for Teams, whether they’re buying Teams as standalone software or adding it on top of a Teams-free Office/Microsoft 365 subscription.

For the Business and Frontline Microsoft 365 versions, you can either buy the version with Teams included for the same price as before, or choose a new Teams-less option that will save you a couple of dollars per user per month. For example, the Teams-less version of Microsoft 365 Business Standard costs $10.25 per user per month, compared to $12.50 for the version that includes Teams.

Updated April 1, 2024, at 4: 12 pm to add more details about pricing and a link to Microsoft’s official blog post about the announcement; also added a statement from a Microsoft spokesperson.

Microsoft splits up the Teams and Office apps worldwide, following EU split Read More »

intel,-microsoft-discuss-plans-to-run-copilot-locally-on-pcs-instead-of-in-the-cloud

Intel, Microsoft discuss plans to run Copilot locally on PCs instead of in the cloud

AI, chatgpt, copilot, intel, microsoft, Tech / /

the ai pc —

Companies are trying to make the “AI PC” happen with new silicon and software.

–

The basic requirements for an AI PC, at least when it's running Windows.

Enlarge / The basic requirements for an AI PC, at least when it’s running Windows.

Intel

Microsoft said in January that 2024 would be the year of the “AI PC,” and we know that AI PCs will include a few hardware components that most Windows systems currently do not include—namely, a built-in neural processing unit (NPU) and Microsoft’s new Copilot key for keyboards. But so far we haven’t heard a whole lot about what a so-called AI PC will actually do for users.

Microsoft and Intel are starting to talk about a few details as part of an announcement from Intel about a new AI PC developer program that will encourage software developers to leverage local hardware to build AI features into their apps.

The main news comes from Tom’s Hardware, confirming that AI PCs would be able to run “more elements of Copilot,” Microsoft’s AI chatbot assistant, “locally on the client.” Currently, Copilot relies on server-side processing even for small requests, introducing lag that is tolerable if you’re making a broad request for information but less so if all you want to do is change a setting or get basic answers. Running generative AI models locally could also improve user privacy, making it possible to take advantage of AI-infused software without automatically sending information to a company that will use it for further model training.

Right now, Windows doesn’t use local NPUs for much, since most current PCs don’t have them. The Surface Studio webcam features can use NPUs for power-efficient video effects and background replacement, but as of this writing that’s pretty much it. Apple’s and Google’s operating systems both use NPUs for a wider swatch of image and audio processing features, including facial recognition and object recognition, OCR, live transcription and translation, and more.

Intel also said that Microsoft would require NPUs in “next-gen AI PCs” to hit speeds of 40 trillion operations per second (TOPS) to meet its requirements. Intel, AMD, Qualcomm, and others sometimes use TOPS as a high-level performance metric when comparing their NPUs; Intel’s Meteor Lake laptop chips can run 10 TOPS, while AMD’s Ryzen 7040 and 8040 laptop chips hit 10 TOPS and 16 TOPS, respectively.

Unfortunately for Intel, the first company to put out an NPU suitable for powering Copilot locally may come from Qualcomm. The company’s upcoming Snapdragon X processors, long seen as the Windows ecosystem’s answer to Apple’s M-series Mac chips, promise up to 45 TOPS. Rumors suggest that Microsoft will shift the consumer version of its Surface tablet to Qualcomm’s chips after a few years of offering both Intel and Qualcomm options; Microsoft announced a Surface Pro update with Intel’s Meteor Lake chips last week but is only selling it to businesses.

Asus and Intel are offering a NUC with a Meteor Lake CPU and its built-in NPU as an AI development platform.

Enlarge / Asus and Intel are offering a NUC with a Meteor Lake CPU and its built-in NPU as an AI development platform.

Intel

All of that said, TOPS are just one simplified performance metric. As when using FLOPS to compare graphics performance, it’s imprecise and won’t capture variations in how each NPU handles different tasks. And the Arm version of Windows still has software and hardware compatibility issues that could continue to hold it back.

As part of its developer program, Intel is also offering an “AI PC development kit” centered on an Asus NUC Pro 14, a mini PC built around Intel’s Meteor Lake silicon. Intel formally stopped making its NUC mini PCs last year, passing the brand and all of its designs off to Asus. Asus is also handling all remaining warranty service and software support for older NUCs designed and sold by Intel. The NUC Pro 14 is one of the first new NUCs announced since the transition, along with the ROG NUC mini gaming PC.

Intel, Microsoft discuss plans to run Copilot locally on PCs instead of in the cloud Read More »

world’s-first-global-ai-resolution-unanimously-adopted-by-united-nations

World’s first global AI resolution unanimously adopted by United Nations

AI, AI ethics, AI harms, AI policy, AI regulation, AI safety, Biz & IT, Bletchley Declaration, brad smith, Britain, machine learning, microsoft, Policy, Reuters, UK, UN, United Nations, united states, US government / /

We hold these seeds to be self-evident —

Nonbinding agreement seeks to protect personal data and safeguard human rights.

–

The United Nations building in New York.

Enlarge / The United Nations building in New York.

On Thursday, the United Nations General Assembly unanimously consented to adopt what some call the first global resolution on AI, reports Reuters. The resolution aims to foster the protection of personal data, enhance privacy policies, ensure close monitoring of AI for potential risks, and uphold human rights. It emerged from a proposal by the United States and received backing from China and 121 other countries.

Being a nonbinding agreement and thus effectively toothless, the resolution seems broadly popular in the AI industry. On X, Microsoft Vice Chair and President Brad Smith wrote, “We fully support the @UN’s adoption of the comprehensive AI resolution. The consensus reached today marks a critical step towards establishing international guardrails for the ethical and sustainable development of AI, ensuring this technology serves the needs of everyone.”

The resolution, titled “Seizing the opportunities of safe, secure and trustworthy artificial intelligence systems for sustainable development,” resulted from three months of negotiation, and the stakeholders involved seem pleased at the level of international cooperation. “We’re sailing in choppy waters with the fast-changing technology, which means that it’s more important than ever to steer by the light of our values,” one senior US administration official told Reuters, highlighting the significance of this “first-ever truly global consensus document on AI.”

In the UN, adoption by consensus means that all members agree to adopt the resolution without a vote. “Consensus is reached when all Member States agree on a text, but it does not mean that they all agree on every element of a draft document,” writes the UN in a FAQ found online. “They can agree to adopt a draft resolution without a vote, but still have reservations about certain parts of the text.”

The initiative joins a series of efforts by governments worldwide to influence the trajectory of AI development following the launch of ChatGPT and GPT-4, and the enormous hype raised by certain members of the tech industry in a public worldwide campaign waged last year. Critics fear that AI may undermine democratic processes, amplify fraudulent activities, or contribute to significant job displacement, among other issues. The resolution seeks to address the dangers associated with the irresponsible or malicious application of AI systems, which the UN says could jeopardize human rights and fundamental freedoms.

Resistance from nations such as Russia and China was anticipated, and US officials acknowledged the presence of “lots of heated conversations” during the negotiation process, according to Reuters. However, they also emphasized successful engagement with these countries and others typically at odds with the US on various issues, agreeing on a draft resolution that sought to maintain a delicate balance between promoting development and safeguarding human rights.

The new UN agreement may be the first “global” agreement, in the sense of having the participation of every UN country, but it wasn’t the first multi-state international AI agreement. That honor seems to fall to the Bletchley Declaration signed in November by the 28 nations attending the UK’s first AI Summit.

Also in November, the US, Britain, and other nations unveiled an agreement focusing on the creation of AI systems that are “secure by design” to protect against misuse by rogue actors. Europe is slowly moving forward with provisional agreements to regulate AI and is close to implementing the world’s first comprehensive AI regulations. Meanwhile, the US government still lacks consensus on legislative action related to AI regulation, with the Biden administration advocating for measures to mitigate AI risks while enhancing national security.

World’s first global AI resolution unanimously adopted by United Nations Read More »

microsoft-debuts-major-surface-overhauls-that-regular-people-can’t-buy

Microsoft debuts major Surface overhauls that regular people can’t buy

copilot, microsoft, Microsoft Surface, Surface, surface laptop, surface pro, Tech, Windows 11 / /

business time —

Not the first business-exclusive Surfaces, but they’re the most significant.

–

  • Microsoft

  • Yes, both devices launch with Microsoft’s new Copilot key.

    Microsoft

  • The Surface Pro 10. Looks familiar.

    Microsoft

  • An NFC reader supports physical security keys.

    Microsoft

  • The 13.5- and 15-inch Surface Laptop 6.

    Microsoft

  • The 15-inch Laptop 6 can be configured with a security card reader, another business thing.

    Microsoft

Microsoft is debuting major updates to two of its Surface PCs today: both the Surface Pro 10 and the 13.5- and 15-inch Surface Laptop 6 are major internal upgrades to Microsoft’s mainstream Surface devices. Both were last updated nearly a year and a half ago, and they’re both getting new Intel chips with significantly faster integrated GPUs, upgraded webcams, the Copilot key, and better battery life (according to Microsoft’s spec sheets).

The catch is that both of these Surfaces are being sold exclusively to businesses and commercial customers; as of this writing, regular people will not be able to buy one directly from Microsoft, and they won’t show up in most retail stores.

These aren’t the first Surface products released exclusively for businesses. Microsoft introduced a new business-exclusive Surface Go 3 tablet last fall, and a Surface Pro 7+ variant for businesses in early 2021. It is, however, the first time Microsoft has introduced new versions of its flagship tablet and laptop without also making them available to consumers. You can find some of these business-only PCs for sale at some third-party retailers, but usually with extended shipping times and higher prices than consumer systems.

Though this seems like a step back from the consumer PC market, Microsoft is still reportedly planning new consumer Surfaces. The Verge reports that Microsoft is planning a new Surface with Qualcomm’s upcoming Snapdragon X chip, to debut in May. It’s that device, rather than today’s traditional Intel-based Surface Pro 10, that will apparently take over as the flagship consumer Surface PC.

“We absolutely remain committed to consumer devices,” a Microsoft spokesperson told Ars. “Building great devices that people love to use aligns closely with our company mission to empower individuals as well as organizations. We are excited to be bringing devices to market that deliver great AI experiences to our customers. This commercial announcement is only the first part of this effort.”

This would be a big departure for Microsoft, which for a few years now has offered the Intel-based Surface tablets as its primary convertible tablets and the Arm-based Surface Pro X and Surface Pro 9 with 5G as separate niche variants. Older Qualcomm chips’ mediocre performance and lingering software and hardware compatibility issues with the Arm version of Windows have held those devices back, though Snapdragon X at least promises to solve the performance issues. If Microsoft plans to go all-in on Arm for its flagship consumer Surface device, it at least makes a little sense to retain the Intel-based Surface for businesses that will be more sensitive to those performance and compatibility problems.

What’s new in the Surface Pro 10 and Surface Laptop 6?

As for the hardware itself, for people who might be getting them at work or people who go out of their way to find one: The biggest upgrade is that both Surface devices have been updated with Intel Core Ultra CPUs based on the Meteor Lake architecture. While the processor performance improvements in these chips are a bit underwhelming, their Arc-integrated GPUs are significantly faster than the old Iris Xe GPUs. And the chips also include a neural processing unit (NPU) that can accelerate some AI and machine-learning workloads; Microsoft currently uses them mostly for fancy webcam effects, but more software will likely take advantage of them as they become more widely available.

Those new chips (and small battery capacity increases) have also bumped all of Microsoft’s battery life estimates up a bit. The Surface Pro 10 is said to be good for 19 hours of “typical device usage,” up from 15.5 hours from the Intel version of the Surface Pro 9. The 13.5 and 15-inch Surface Laptop 6 gets 18.5 and 19 hours of battery life, respectively, up from 18 and 17 hours for the Surface Laptop 5.

The downside is that the Surface Laptops are a bit heavier than the Laptop 5: 3.06 pounds and 3.7 pounds, compared to 2.86 and 3.44 pounds for the 13.5- and 15-inch models.

Both models also get new webcam hardware to go with those NPU-accelerated video effects. The Surface Pro goes from a 1080p webcam to a 1440p webcam, and the Surface Laptop goes from 720p to 1080p. The Surface Pro 10’s camera also features an “ultrawide field of view.” Both cameras support Windows Hello biometric logins using a scan of your face, and the Surface Pro 10 also has an NFC reader for use with hardware security keys. As business machines, both devices also have dedicated hardware TPM modules to support drive encryption and other features, instead of the firmware TPMs that the Surface Pro 9 and Surface Laptop 5 used. Neither supports Microsoft’s Pluton technology.

A new Type Cover with a brighter backlight and bolder legends was made for users with low vision or those who want to reduce eyestrain.

Enlarge / A new Type Cover with a brighter backlight and bolder legends was made for users with low vision or those who want to reduce eyestrain.

Microsoft

Neither device gets a big screen update, though there are small improvements. Microsoft says the Surface Pro 10’s 13-inch, 2880×1920 touchscreen is 33 percent brighter than before, with a maximum brightness of 600 nits. The screen has a slightly better contrast ratio than before and an anti-reflective coating; it also still supports a 120 Hz refresh rate. The Surface Laptop 6 doesn’t get a brightness bump but does have better contrast and an anti-reflective coating. Both devices are still using regular IPS LCD panels rather than OLED or something fancier.

And finally, some odds and ends. The 15-inch Surface Laptop 6 picks up a second Thunderbolt port and optional support for a smart card reader. The Surface Pro now has a “bold keyset” keyboard option, with an easier-to-read font and brighter backlight for users with low vision. These keyboards should also work with some older Surface devices, if you can find them.

The systems will be available to pre-order “in select markets” on March 21, and they’ll begin shipping on April 9. Microsoft didn’t share any specifics about pricing, though as business machines, we’d generally expect them to cost a little more than equivalent consumer PCs.

Listing image by Microsoft

Microsoft debuts major Surface overhauls that regular people can’t buy Read More »

deepmind-co-founder-mustafa-suleyman-will-run-microsoft’s-new-consumer-ai-unit

DeepMind co-founder Mustafa Suleyman will run Microsoft’s new consumer AI unit

AI, copilot, deepmind, microsoft, syndication / /

Minding deeply —

Most staffers from Suleyman’s startup, Inflection, will join Microsoft as well.

–

Mustafa Suleyman, talks on Day 1 of the AI Safety Summit at Bletchley Park at Bletchley Park on November 1, 2023 in Bletchley, England.

Enlarge / Mustafa Suleyman, talks on Day 1 of the AI Safety Summit at Bletchley Park at Bletchley Park on November 1, 2023 in Bletchley, England.

Microsoft has hired Mustafa Suleyman, the co-founder of Google’s DeepMind and chief executive of artificial intelligence start-up Inflection, to run a new consumer AI unit.

Suleyman, a British entrepreneur who co-founded DeepMind in London in 2010, will report to Microsoft chief executive Satya Nadella, the company announced on Tuesday. He will launch a division of Microsoft that brings consumer-facing products including Microsoft’s Copilot, Bing, Edge, and GenAI under one team called Microsoft AI.

It is the latest move by Microsoft to capitalize on the boom in generative AI. It has invested $13 billion in OpenAI, the maker of ChatGPT, and rapidly integrated its technology into Microsoft products.

Microsoft’s investment in OpenAI has given it an early lead in Silicon Valley’s race to deploy AI, leaving its biggest rival, Google, struggling to catch up. It also has invested in other AI startups, including French developer Mistral.

It has been rolling out an AI assistant in its products such as Windows, Office software, and cyber security tools. Suleyman’s unit will work on projects including integrating an AI version of Copilot into its Windows operating system and enhancing the use of generative AI in its Bing search engine.

Nadella said in a statement on Tuesday: “I’ve known Mustafa for several years and have greatly admired him as a founder of both DeepMind and Inflection, and as a visionary, product maker and builder of pioneering teams that go after bold missions.”

DeepMind was acquired by Google in 2014 for $500 million, one of the first large bets by a big tech company on a startup AI lab. The company faced controversy a few years later over some of its projects, including its work for the UK healthcare sector, which was found by a government watchdog to have been granted inappropriate access to patient records.

Suleyman, who was the main public face for the company, was placed on leave in 2019. DeepMind workers had complained that he had an overly aggressive management style. Addressing staff complaints at the time, Suleyman said: “I really screwed up. I was very demanding and pretty relentless.”

He moved to Google months later, where he led AI product management. In 2022, he joined Silicon Valley venture capital firm Greylock and launched Inflection later that year.

Microsoft will also hire most of Inflection’s staff, including Karén Simonyan, cofounder and chief scientist of Inflection, who will be chief scientist of the AI group. Microsoft did not clarify the number of employees moving over but said it included AI engineers, researchers, and large language model builders who have designed and co-authored “many of the most important contributions in advancing AI over the last five years.”

Inflection, a rival to OpenAI, will switch its focus from its consumer chatbot, Pi, and instead move to sell enterprise AI software to businesses, according to a statement on its website. Sean White, who has held various technology roles, has joined as its new chief executive.

Inflection’s third cofounder, Reid Hoffman, the founder and executive chair of LinkedIn, will remain on Inflection’s board. Inflection had raised $1.3 billion in June, valuing the group at about $4 billion, in one of the largest fundraisings by an AI start-up amid an explosion of interest in the sector.

The new unit marks a big organizational shift at Microsoft. Mikhail Parakhin, its president of web services, will move along with his entire team to report to Suleyman.

“We have a real shot to build technology that was once thought impossible and that lives up to our mission to ensure the benefits of AI reach every person and organization on the planet, safely and responsibly,” Nadella said.

Competition regulators in the US and Europe have been scrutinising the relationship between Microsoft and OpenAI amid a broader inquiry into AI investments.

© 2024 The Financial Times Ltd. All rights reserved. Not to be redistributed, copied, or modified in any way.

DeepMind co-founder Mustafa Suleyman will run Microsoft’s new consumer AI unit Read More »