Home Blog

Google challenges NVIDIA’s Vera Rubin supremacy with new 8th-Gen TPUs

By
Pradeep Viswanathan
-

When it comes to powering AI infrastructure, NVIDIA is the undisputed leader right now. While the market expected AMD and others to challenge NVIDIA’s dominance, Google and AWS surprised many by consistently rolling out their own AI chips. These chips are now powering not only their own products and services but also third-party customers, including Anthropic, one of the fastest-growing AI startups in the world.

At Google Cloud Next today, Google announced two new eighth-generation TPUs: TPU 8t and TPU 8i. Both of these TPUs are purpose-built architectures for training and inference, respectively. With these new chips, Google is positioning its latest AI infrastructure as a serious alternative to NVIDIA for both frontier model development and production AI services.

Google claims that the new TPU 8t is designed for training frontier models faster by cutting development cycles from months to weeks, offering nearly 3x more compute performance per pod than the previous generation. The company claims that a single TPU 8t superpod can now scale to 9,600 chips and 2 petabytes of shared high-bandwidth memory to offer 121 exaflops of compute, doubling the interchip bandwidth compared to the previous generation.

Google also highlighted that TPU 8t comes with 10x faster storage access and support for TPUDirect, which allows data to be pulled directly into the TPU. The company claims that its new Virgo Network, along with JAX and Pathways, enables near-linear scaling for up to one million chips in a single logical cluster.

The new TPU 8i is built for low-latency, reasoning-heavy inference workloads and can deliver 80% better performance-per-dollar than the previous generation. The TPU 8i chip combines 288GB of high-bandwidth memory with 384MB of on-chip SRAM to keep a model’s active working set closer to the processor.

Google is also ditching x86 CPUs for its own custom Axion Arm-based CPUs for better overall efficiency. For modern Mixture of Experts (MoE) models, Google has doubled interconnect bandwidth to 19.2Tb/s, and the new Boardfly architecture reduces maximum network diameter by more than 50% to reduce system latency. The new on-chip Collectives Acceleration Engine (CAE) offloads global operations and reduces on-chip latency by up to 5x, according to Google.

Both TPU 8t and TPU 8i chips will be generally available later this year.

AMD Ryzen AI PRO 300 Series brings Copilot+ PCs to the enterprise

By
Pradeep Viswanathan
-
NVIDIA Grace Blackwell Superchip

NVIDIA Grace Blackwell Superchip

Back in June, AMD first launched the Ryzen AI 300 series CPUs. Recently, AMD launched the commercial version of these processors, the Ryzen AI PRO 300 series, specifically targeting enterprise PCs. These new CPUs bring AMD-powered Copilot+ PC functionality to the enterprise for the first time and deliver up to three times the AI performance compared to the previous generation.

To meet the needs of enterprise organizations, the Ryzen AI PRO 300 series incorporates AMD PRO Technologies, offering several security and manageability features. These include AMD Secure Processor, AMD Shadow Stack, AMD Platform Secure Boot, and a new Cloud Bare Metal Recovery feature that allows enterprise IT teams to recover systems seamlessly via the cloud. Additionally, the new Supply Chain Security (AMD Device Identity) feature enables traceability across the supply chain, while the Watch Dog Timer enables additional detection and recovery processes. Notably, some of the new security features in the Ryzen AI PRO 300 series now utilize the integrated NPU to run AI-based security workloads without impacting CPU performance.

Highlights of Ryzen AI PRO 300 Series processors:

  • AMD “Zen 5” architecture.
  • The Ryzen AI 9 HX PRO 375 offers up to 40% higher overall performance and up to 14% faster productivity performance compared to Intel’s Core Ultra 7 165U.
  • With the addition of XDNA 2 architecture powering the integrated NPU, these CPUs offer a cutting-edge 50+ NPU TOPS of AI processing power, exceeding Microsoft’s Copilot+ AI PC requirements.
  • Built on a 4nm process with innovative power management for improved battery life.

Jack Huynh, senior vice president and general manager, Computing and Graphics Group at AMD, said:

“Enterprises are increasingly demanding more compute power and efficiency to drive their everyday tasks and most taxing workloads. We are excited to add the Ryzen AI PRO 300 Series, the most powerful AI processor built for business PCs, to our portfolio of mobile processors. Our third generation AI-enabled processors for business PCs deliver unprecedented AI processing capabilities with incredible battery life and seamless compatibility for the applications users depend on.”

AMD expects that more than 100 Ryzen AI PRO PCs will be available in the market by 2025.

Google Gemini now allows users to create Gems, their own custom AI experts

By
Pradeep Viswanathan
-
Google Gems

Google Gems

During Google I/O’s keynote event this past May, Google announced Gems, customizable versions of Gemini that users can create. Users can create a Gem for various scenarios, such as a gym buddy, sous chef, coding partner, creative writing guide, and more. Google announced today that Gemini Advanced, Business, and Enterprise subscribers can start creating and chatting with Gems in the coming days.

Creating a Gem is a simple process. Users just write simple instructions for their Gem, provide a name for it, and then start chatting with it. Gems can remember a detailed set of instructions, allowing users to save time on repetitive tasks.

To get started, Google is also launching a set of pre-made Gems for the following scenarios:

  • Learning coach: Helps you break down complex topics, making them easier to understand.
  • Brainstormer: Gives you easy inspiration — from fresh ideas for a themed party to the perfect gift for an upcoming birthday.
  • Career guide: Unlocks your career potential with detailed plans to refine your skills and achieve your career goals.
  • Writing editor: Can elevate your writing through clear, constructive feedback on everything from grammar to structure.
  • Coding partner: Levels up your coding skills and can help you build projects and learn as you go.

Gems will be available on both desktop and mobile devices in the coming days to all Gemini Advanced, Gemini Business, and Gemini Enterprise users in 150 countries, in the various languages that Gemini supports.

Microsoft Copilot had a similar feature called Copilot GPT Builder. Unfortunately, Microsoft ended the Copilot GPT Builder feature for general consumers last month. In the retirement announcement, Microsoft mentioned that they want to focus on GPTs for commercial and enterprise scenarios.

Whether Gems will achieve mainstream adoption and prove to be a successful venture for Google remains to be seen. It will be interesting to observe if Google can tap into the consumer market’s desire for personalized AI experiences and establish Gems as a valuable tool for everyday tasks and interactions.

Stack Overflow’s Annual Developer Survey reveals that AI tools are here to stay

By
Pradeep Viswanathan
-
Stackoverflow 2024 Developer Survey

Stackoverflow 2024 Developer Survey

In May of this year, the 14th annual Stack Overflow Developer Survey began. Last week, Stack Overflow published the survey results based on responses from over 65,000 developers. Participating developers provided info about coding, technologies, tools they use, and more.

Some of the results were not surprising, such as JavaScript remaining the most popular programming language in the world. Another interesting result from the survey concerns the usage of AI tools in development. Over 76% of all respondents use or plan to use AI tools in their development process this year, an increase from last year. Also, 70% of professional developers do not perceive AI as a threat to their jobs.

2024 Stack Overflow Developer Survey Highlights:

  • PostgreSQL is the most popular database for the second year in a row, with over 49% of respondents using it. Microsoft SQL Server was ranked 4th in the survey.
  • AWS is the most used cloud platform, followed by Azure and Google Cloud with 27.8% and 25.1% usage, respectively.
  • Node.js is the most popular web framework, with React close behind. Microsoft’s ASP.NET Core was ranked 7th.
  • Raspberry Pi is the most popular platform in embedded development.
  • .NET is the most used among other frameworks.
  • Docker is the most popular tool among developers for compiling, building, and testing.
  • Visual Studio Code is the most popular IDE, with over 73.6% usage, followed by Visual Studio.
  • Microsoft Teams is the most used chat platform.
  • Windows remained the most popular OS for developers, across both personal and professional use.
  • ChatGPT is the most popular AI tool among developers. GitHub Copilot is ranked 2nd on the list.
  • API and SDK documents are the documentation source of choice for 90% of developers.

You can review the full results of the 2024 Stack Overflow Developer Survey here.

Microsoft Surface Takes UEFI In-House to Bolster Device Security and Performance

By
Pradeep Viswanathan
-
Microsoft Surface UEFI

In response to the evolving landscape of digital threats, Microsoft’s Surface team is proactively enhancing device security through the development of their Unified Extensible Firmware Interface (UEFI). This initiative aims to deliver a more secure and personalized experience for Surface users, distinguishing these devices in the competitive market.

The UEFI development for Surface is rooted in Project Mu, an open-source endeavor focused on creating robust firmware solutions. Leveraging the benefits of Rust, a programming language renowned for its performance, safety features, and ease of maintenance, further fortifies the firmware against vulnerabilities and attacks, demonstrating a steadfast commitment to security.

By bolstering their UEFI stack, Surface gains comprehensive control over firmware management. This capability facilitates swift responses to emerging threats and seamless integration with Windows update processes. Internally developing this technology reduces reliance on external providers, thereby minimizing potential security risks.

Surface UEFI incorporates several pivotal security features:

  • Root of Trust: Ensures secure device boot-up from reset, safeguarding against unauthorized firmware alterations and validated by hardware.
  • Secured Core PC: Employs technologies like Dynamic Root of Trust for Measurement (DRTM) and Firmware Attack Surface Reduction (FASR) to defend against sophisticated attacks, crucial for high-security environments.
  • Standalone Management Mode: Enhances isolation for Secured Core PCs, bolstering resistance to potential threats.

For corporate and enterprise users, Surface UEFI offers advanced capabilities such as Device Firmware Configuration Interface (DFCI) and Surface Enterprise Management Mode (SEMM). These tools enable remote management of UEFI settings, secure configurations, selective hardware deactivation, and control over peripheral access. Dynamic USB-C disablement adds an additional layer of security by preventing unauthorized device connections.

Surface UEFI prioritizes fast boot times and seamless device servicing. It optimizes wake-up speeds across various scenarios and supports on-field upgrades and repairs, minimizing downtime and maintenance costs.

Looking ahead, Surface UEFI is poised for continued evolution with plans to enhance capabilities and elevate security measures. Through collaboration with industry partners, adoption of advanced technologies, and participation in UEFI-related initiatives, Surface aims to lead the way in firmware innovation.

This proactive approach underscores Surface’s commitment to delivering a secure and high-performance computing environment, setting a new standard for device security within the industry.

Microsoft Designer Unveils AI-Powered ‘Restyle Image’ Feature, Transforming Photos with Artistic Flair

By
Pradeep Viswanathan
-
Microsoft Designer Restyle

Microsoft has a fresh development in digital image creation called “Restyle Image,” found within Microsoft Designer. This new tool uses AI to let users easily convert regular photos into attractive artworks, offering different styles like pop art or stylized 3D.

Microsoft Designer, which is already liked as a graphic design app, has added more creative power with Restyle Image. This new feature uses AI to make many artistic versions of photos users upload. They can then select the style that matches their idea best.

Restyle Image is useful for creating different avatars on social media account, improving visuals in blogs, and giving life to materials in education.

Key Features and Benefits:

Different Types of Art: Pick from various styles such as pop art, flat illustration, papercraft and other types.

Customizable Prompts: Guide the AI’s creative process by adding specific descriptions and keywords.

User-Friendly Inclusion: Utilize a Microsoft account to access Restyle Image directly from the Microsoft Designer home screen.

Varying Results: You may produce one, two, three or four choices for each uploaded picture.

Edit and Download: Further refine generated images or directly download them for use.

Restyle Image is available only in English for now. Microsoft has plans to add additional language support in the future. They are also aware of the limitations that exist and are working on solutions for them like tackling occasional misspelling in the generated images.

Quantum Leap in Scientific Discovery: Microsoft Introduces Generative Chemistry & Accelerated DFT in Azure Quantum Elements

By
Pradeep Viswanathan
-
Microsoft Quantum Generative Chemistry

Microsoft recently announced its new capabilities offered by Azure Quantum Elements, a platform that aims to revolutionize scientific discovery. The platform introduces Generative Chemistry and Accelerated DFT to help researchers design new molecules and carry out quantum mechanical simulations at unprecedented speeds.

The Generative Chemistry feature is meant to offer a faster and more efficient alternative to the traditional process of discovering new molecules. Users start by providing details about the characteristics of a molecule they’re looking for (e.g. a substance that can break down rapidly in the environment). Generative Chemistry then uses a combination of AI models trained on huge datasets of existing compounds to generate a list of potential candidate molecules.

But Generative Chemistry doesn’t stop there. The platform can suggest steps for how to create the molecules in a lab, and provide insights on how to best evaluate the molecules in real-world experiments. And to ensure that the molecules it suggests are truly novel, Generative Chemistry incorporates a “feedback loop” that evaluates the molecules relative to known compounds, Microsoft says.

Azure Quantum Elements is also getting a new capability called Accelerated DFT, which allows researchers to perform quantum mechanical simulations — in this case, simulations of the electronic properties of molecules — in a matter of hours as opposed to days. Microsoft says that Accelerated DFT, which is based on technology developed by its Azure Quantum and Microsoft Research teams, was built with input from organizations like AspenTech, Unilever and DTU Energy University of Denmark.

Microsoft says that AspenTech, for example, plans to use Accelerated DFT to train AI models for tasks like characterizing new molecules and materials. And Unilever, which has been using other Azure Quantum Elements features in its R&D efforts, will use Accelerated DFT to explore new types of materials for personal care and home products, Microsoft says.

Speaking of Unilever, the company is also one of the first customers to have access to Generative Chemistry and Accelerated DFT; the features are now in private preview for select Azure Quantum Elements customers.

Microsoft Completes Activision Blizzard merger: A Game-Changing Alliance in the Gaming Industry

By
MSD
-
activision blizzard Microsoft

It’s a deal that’s been two years in the making, and it’s finally come to fruition. On October 13, 2023, Phil Spencer, CEO of Microsoft Gaming, officially announced the integration of Activision Blizzard King into the Xbox team. This marks the biggest acquisition in video game history, forever altering the landscape of the gaming industry.

Here’s what you need to know about this groundbreaking merger and its potential implications.

Overcoming Regulatory Hurdles

The road to this acquisition was far from smooth. Microsoft had to navigate a complex labyrinth of regulatory challenges, both in Europe and the United States. Initial objections from the UK’s Competition and Markets Authority (CMA) centered on concerns about Microsoft’s potential monopoly in the cloud gaming sector. Similar reservations were expressed by the European Union, leading to demands for cloud-related concessions.

To address these issues, the deal was restructured to allow some of Activision’s cloud streaming rights to be sold to rival Ubisoft Entertainment. This amendment not only appeased the CMA but also led to last month’s preliminary approval, thereby overcoming the final obstacle to the acquisition’s completion.

Market Impact: A Tale of Two Stocks

The announcement of the merger had an immediate and significant impact on the stock market. While Activision’s shares skyrocketed by as much as 36%, Microsoft experienced a decrease, ending the day down by 2.5% from its opening price. This divergence likely reflects investors’ differing opinions on the immediate benefits of the acquisition for each company.

A Strategic Power Play for Microsoft

The strategic implications of this acquisition are immense for Microsoft, particularly for its gaming division, Xbox. Microsoft gains access to some of the world’s most popular franchises, ranging from “Call of Duty” to “World of Warcraft.” Phil Spencer’s announcement suggests a broad, future-looking strategy. While Microsoft initially focused on the metaverse, the acquisition allows them to diversify and strengthen their gaming portfolio substantially. This move is expected to bolster Microsoft’s competitiveness against market leaders like Tencent and Sony.

Embracing a Player-Centric Philosophy

In his announcement, Phil Spencer emphasized a shared, player-centric philosophy, stating that “Players have always been at the center of everything we do.” This approach is expected to continue as the two giants merge, with plans to bring favorite games to more platforms, engage with players in innovative ways, and offer more games in more places, starting with cloud streaming in the European Economic Area.

“When Everyone Plays, We All Win”

Microsoft Xbox Activision Blizzard

One of the most striking aspects of the announcement was its focus on inclusivity and community. Spencer assured fans of Activision, Blizzard, and King games that they would continue to be welcome, regardless of their preferred gaming platform. This community-driven focus is a cornerstone of both Microsoft’s and Activision Blizzard’s strategies, and it’s poised to be a defining feature of their joint efforts going forward.

Anticipating the Road Ahead

Though the acquisition has cleared its final regulatory hurdles, questions remain. How will this deal affect competition within the industry? What will be its impact on subscription costs and the broader cloud gaming landscape? These are questions that will be closely watched in the months and years to come.

Final Thoughts

The Microsoft-Activision Blizzard merger is more than just a business deal; it’s a seismic shift in the gaming industry that will set precedents for years to come. With the promise of creating new worlds, expanding the reach of iconic franchises, and fostering a player-centric community, this union heralds an exciting new chapter in interactive entertainment.

SpaceX Aims to Transform Mobile Connectivity with Starlink Direct to Cell

By
MSD
-

As we hurdle toward a future where global connectivity becomes a non-negotiable, SpaceX is planning to take a significant leap in this direction with its Starlink Direct to Cell service. The ambitious project aims to provide ubiquitous coverage from cell phone towers in space, radically transforming how we think about mobile connectivity.

The Starlink Plan: Bringing Space Closer to Your Smartphone

SpaceX recently updated its Starlink website, revealing more details about its satellite-delivered cell phone service. The company promises to start with text services in 2024, followed by voice and data services in 2025. Internet of Things (IoT) services are also planned for the same year. What sets this apart from existing satellite phone services is the compatibility with standard LTE phones without requiring any special hardware or firmware changes.

Starlink’s approach differs fundamentally from what’s currently available in satellite connectivity. Existing networks like Globalstar and Iridium operate at altitudes of 1,400 km and 781 km above Earth, respectively. Starlink plans to operate its satellites much closer to Earth, at around 550 km, allowing for more effective and efficient connections.

Technical Innovations: Bigger Rockets, Bigger Satellites, Better Connectivity

SpaceX has the advantage of using its own rockets for deploying satellites. The company is developing the world’s largest rocket, Starship, capable of launching bigger satellites equipped with more sensitive antennas. The closer proximity and larger antennas make it easier for regular smartphones to connect with these space-based cell towers, a feat not previously possible.

SpaceX’s Falcon 9 rocket will initially carry these advanced satellites into orbit, with plans to transition to the larger Starship rocket for future launches. The shift to Starship is crucial for Starlink’s more ambitious plans, as the full-sized “V2” satellites can’t fit into the Falcon 9.

Partnerships and Rollout

Starlink has already partnered with several traditional cell phone companies to sell the service, including T-Mobile in the US, Rodgers in Canada, and KDDI in Japan, among others. The company is actively seeking additional cellular partners to expand its global reach. However, like all SpaceX projects, the rollout timeline should be considered cautiously. Initial plans to begin beta service this year have already been delayed.

Competing Technologies: How Does iPhone’s Emergency Satellite Service Stack Up?

Apple recently introduced an emergency satellite service available on its latest iPhone models, enabling text to emergency services in areas without cellular or Wi-Fi coverage. While it’s a significant step for emergency communications, it is not designed to replace regular phone service. It also has limitations such as slower transmission speed and lower image quality. Unlike Starlink, it requires you to be inside a connectivity window and follow a signal-targeting app.

Final Thoughts: A Network Above and Beyond

SpaceX’s Starlink Direct to Cell service promises a future where one can stay connected from virtually anywhere on Earth, as long as the sky is visible. This level of global coverage has far-reaching implications, not just for everyday communication but also for emergency services, IoT devices, and other applications that require reliable connectivity.

From eliminating dead zones to providing peace of mind in remote regions, the service could be a game-changer. The successful execution of this plan could redefine mobile connectivity, making the sky — quite literally — the limit. With technological advancements and strategic partnerships, SpaceX is all set to make this vision a reality. But as with all groundbreaking initiatives, only time will tell how smoothly this journey to ubiquitous connectivity will unfold.

Bard Catches Up with Bing Chat with Generative Art Support

By
MSD
-

Generative AI is increasingly becoming a focal point in search and communication platforms. As Bing Chat integrates OpenAI’s DALL-E 3 for text-to-image capabilities, Bard, another conversational AI, has also jumped into the generative art arena. This development follows a broader trend in which AI is not just a tool for information retrieval but also a creative assistant that can generate images and even write drafts based on user input.

Google’s Search Gets a Creative Twist

Google has been experimenting with generative AI capabilities in its Search function. Known as Search Generative Experience (SGE), this new feature allows users to generate images based on specific queries. For example, if a user searches for “draw a picture of a capybara wearing a chef’s hat and cooking breakfast,” SGE will display up to four generated images in line with the search query. Users can even modify these images by editing the descriptions to add more details.

A search for “draw an image of a capybara wearing a chefs hat and cooking breakfast” leads to four generated images, all depicting a capybara in a chefs hat cooking different types of breakfast foods like bacon and eggs.

This feature is not just limited to Google Search. If you’re opted into the SGE experiment, you may also find an option to create AI-generated images directly in Google Images. This comes handy when searching for inspiration, say, for “minimalist Halloween table settings” or “spooky dog house ideas.”

A Google Image search for “minimalist halloween table settings” shows various images, then taps into an option to “Create something new, see brand news images generated with AI,” which generates an image of a table with a black tablecloth, white plates and napkins and spider decor.

Bing Chat and DALL-E 3: A Symbiotic Relationship

On the other hand, Bing Chat has integrated DALL-E 3, OpenAI’s most advanced text-to-image model. Since the launch of Bing Image Creator, the platform has generated over 1 billion images, serving various creative needs like social media thumbnails, design inspirations, and more. DALL-E 3 enhances this feature by offering more precise, reliable, and aesthetically pleasing images based on textual prompts.

The safety features are also well-thought-out in Bing Chat. Every AI-generated image comes with an invisible digital watermark to confirm its AI-generated provenance. Moreover, a content moderation system removes any images that are harmful or inappropriate.

Comparing Bard and Bing Chat in Generative Art

While Bing Chat focuses primarily on image generation, Bard takes a more holistic approach by combining text and image generation. It helps not just in creating images but also in writing drafts. For example, if you are searching for how to convert your garage into a home office, Bard can help you write a note to a contractor asking for a quote. This draft can then be easily exported to Google Docs or Gmail.

In terms of safety and ethics, both platforms seem to be on the same page. Just like Bing Chat, Bard also ensures responsible AI usage by incorporating metadata labeling and watermarking in every generated image. It also plans to introduce an ‘About this image’ tool to help users assess the context and credibility of images.

User Experience and Accessibility

Bard’s generative art feature is currently available only to those opted into the SGE experiment and is limited to English speakers in the United States. Bing Chat’s DALL-E 3 integration is generally available to everyone and offers a free experience, making it accessible to a broader user base.

Conclusion: The Future of Generative Art in AI Platforms

As generative AI capabilities become more sophisticated, their integration into search and communication platforms is likely to become more prevalent. Both Bard and Bing Chat offer compelling but slightly different approaches to how generative art can be used, each with its own set of advantages and limitations. As these platforms continue to evolve, the way we search, communicate, and even create could undergo a significant transformation.

The generative art race between Bard and Bing Chat is indeed heating up, and it will be fascinating to see how each evolves to serve user needs better while also adhering to ethical considerations. With user feedback and ongoing testing, generative AI has the potential to redefine our interaction with technology, offering a more interactive, creative, and efficient experience.

123...207Page 1 of 207

EDITOR PICKS

POPULAR POSTS

POPULAR CATEGORY

ABOUT US

BigTechWire is a technology blog. We provide you with the latest breaking news from Big Tech companies: Amazon, Apple, Google, Meta and Microsoft.

Contact us: editor@bigtechwire.com

Logo created by @lorisobidesign