Meta’s Generative AI Ambitions, Microsoft & Siemens Bring OpenAI to IIoT, More News
It's been a tumultuous week for Meta. First, it announced the release of its AI-based Segment Anything Model (SAM), which can identify and separate objects in videos and images. But the item that made a bigger splash was an interview Meta’s CTO Andrew Bosworth gave to Nikkei Asia magazine.
Bosworth spoke at length about generative AI and Meta’s ambitions to release a commercialized version of the tech by the end of the year. "[I] expect we'll start seeing some of them [commercialization of the tech] this year,” Bosworth told Nikkei.
The interview received a good deal of coverage, with some commentators questioning whether Mark Zuckerberg was already abandoning his metaverse ambitions. (The short answer is no, we will exploring that further soon).
Meta plans to use generative AI in its metaverse, and the release of SAM indicates what might be on the way. What is also striking about the announcement is that Meta is releasing it under open source, as well as providing full details on its 1 billion mask dataset, which according to the company is the largest segmentation dataset ever created.
Segmentation, according to a Meta post about the release, identifies which pixels in an image belong to a photo of a given object and can be used in the analysis of photos and other kinds of images. What Meta has done is to build a segmentation mode for specific tasks, something that typically would require specialized work conducted by experts with access to AI training infrastructure, not to mention huge quantities of clean data in the enterprise.
The complexities involved in this process help in part explain why Bosworth was talking up generative AI instead of the metaverse.
The blog explained the future applications of SAM in image-dependent work such as web design, video editing and art work, as well as its role in broader AI deployments in helping create a "multimodal understanding of the world." It also made the implications for the metaverse explicit: "In the AR/VR domain, SAM could enable selecting an object based on a user’s gaze and then 'lifting' it into 3D."
Viewing Bosworth's comments in this context, it becomes clear that Meta will use generative AI to help build its vision for the metaverse. The company stated as much at the beginning of March when it launched what Zuckerberg described as a new “top-level group” focused on the development of generative AI.
Keep in mind too that despite massive losses last year for Reality Labs — it posted a $4.28 billion operating loss in the fourth quarter, bringing the total for 2022 to $13.72 billion – Meta is pressing ahead with its metaverse ambitions.
Microsoft, Siemens Bring Generative AI to Factories
The generative AI gold rush continues this week with Redmond, Wash.-based Microsoft announcing an extended partnership with Munich-based industrial manufacturing firm Siemens.
The partnership will combine Microsoft Teams with Siemens’ Teamcenter software for product lifecycle management (PLM), was well as bring the large language models available through Azure OpenAI Service as well as other Azure capabilities to Simens customers.s.
The new capabilities will be on display at the Hannover Messe show, which begins on April 17 in Germany. The companies will demonstrate how OpenAI’s ChatGPT and other Azure AI services can augment Siemens’ industrial automation engineering solutions.
According to a post about the extended partnership, Siemens and Microsoft will demonstrate how generative AI can enhance factory automation and operations through AI-powered software development, problem reporting and visual quality inspection.
“The integration of AI into technology platforms will profoundly change how we work and how every business operates,” wrote Scott Guthrie, executive vice president for Cloud and AI at Microsoft. "With Siemens, we are bringing the power of AI to more industrial organizations, enabling them to simplify workflows, overcome silos and collaborate in more inclusive ways to accelerate customer-centric innovation."
It is hard to hear the word industrial in the context of technology without thinking industrial internet of things (IIoT), which Siemens has invested heavily in over the years. In fact, explaining IIoT on its website, Siemens described it is as "combining information technology (IT) with operational technology (OT) enables maximum transparency in a digital enterprise to meaningfully use the data generated across disciplines within and beyond company borders."
The alignment with generative AI is clear. It also provides a commercial context for its development and use.
The two companies will also be collaborating to help software developers and automation engineers accelerate the code generation for Programmable Logic Controllers (PLC). PLCs are the industrial computers that control most machines across the world’s factories.
Attendees of the Hannover Messe show will also be able to see how the engineering teams behind many of the automated workplaces in large plants can significantly reduce time and the probability of errors by generating PLC code through natural language inputs.
Siemens and Microsoft have been working together for a long time — 35 years, according to the post. Some of those collaborations include several IIoT specific releases. In 2019, for example, the two released MindSphere on Azure, a cloud-based, open IoT operating system, with another join release coming last year following Siemens purchase of predictive maintenance solution provider Senseye.
Microsoft Shuts Down Microsoft Teams Free (Classic)
Staying with Microsoft, this week also saw the closure of Microsoft Teams Free (classic). Those using the legacy free version will now have to move to a newer free version or upgrade to a paid plan.
Microsoft announced its plans to do this in February. For those using classic who haven't taken action, this means you will not longer be able to access chats, files or any other content stored in the free version.
The Essentials version, Microsoft's recommended alternative, costs $4/user/month, which can add up for small companies. Those that do opt for Essentials get unlimited group meetings up to 30 hours long, as many as 300 participants per meeting, and 10GB of cloud storage per user.
Learning Opportunities
There is also the Microsoft 365 Business Basic tier which costs $6/user/month which offers everything you get with Essentials, plus 1TB of cloud storage per user, as well as meetings recordings with transcripts.
A third option Microsoft had offered was the possibility of saving classic version files and then uploading them into the new free version. However, with the cut-off date now past, classic files are no longer accessible.
Generative AI Comes to China’s Alibaba
Generative AI development isn't just happening in the US. Since late December 2022, some of the biggest companies outside of the US have also been talking up their application of generative AI in their technologies.
For example, Baidu in China has already rolled out a ChatGPT equivalent. And this week the Chinese tech giant Alibaba unveiled its new generative AI offering.
Much like Microsoft’s Copilot, Alibaba is bringing its large language model (LLM), called Tongyi Qianwen, into its entire portfolio of apps, making it possible for users interrogate the model using NLP.
To start, Tongyi Qianwen will be integrated into DingTalk, Alibaba's workplace messaging app. It can already be used to summarize meting notes, draft proposals and write emails. It will also be added into Alibaba’s voice assistant.
Registration for the technology to enable customized large language models began last week and came days before the publication of the Cyberspace Administration of China’s new draft rules around the use of these new AI technologies.
China’s Cyberspace Administration (CAC) is China's regulatory authority and is responsible for monitoring and managing the internet and related technologies. The proposed rules are open for public comment until May 10.
It may appear ironic that China is pushing to regulate the emergence of a modern technology. However, concerns about the capabilities of generative AI are so pervasive — particularly with the possibility that generative AI could evolve into general AI soon — that many companies and countries are exploring how it might be regulated.
A group of artificial intelligence experts and industry executives, including Elon Musk, a founding member of OpenAI, have also called for a six-month pause in developing systems more powerful than OpenAI's newly launched GPT-4, in an open letter citing potential risks to society.
Learning Management System Developer LM365 Raises $20M
Finally this week, Aarhus, Denmark-based LMS365 has raised $20 million in its first funding round.
The company provides a learning management system (LMS) for use inside Microsoft 365 tools, notably Teams and SharePoint, which organizations can use to create customized courses from scratch or, alternatively, deliver courses via tools like LinkedIn Learning or Go1.
Embedding a learning management system directly into workers’ everyday tools is one of LMS365’s core selling points and distinguishes it from the like of Workday and Eloomi.
“LMS365 does not invent a technology problem, but meets the professionals where they are,” CEO Rasmus Holst explained to TechCrunch.
LMS365 became a LinkedIn Learning partner in March. According to a statement at the time, the partnership will empower learners to build up their skills by finding LinkedIn Learning training content directly from the LMS365 learning platform.
Have a tip to share with our editorial team? Drop us a line:
About the Author
David is a European-based journalist of 35 years who has spent the last 15 following the development of workplace technologies, from the early days of document management, enterprise content management and content services. Now, with the development of new remote and hybrid work models, he covers the evolution of technologies that enable collaboration, communications and work and has recently spent a great deal of time exploring the far reaches of AI, generative AI and General AI.