Battery One Blog

One of the most exciting uses of artificial intelligence (AI) is in the realm of image generation, where the power of AI can turn any idea into a visual masterpiece. OpenAI pioneered this trend with its release of DALL-E 2, and more recently, the highly anticipated DALL-E 3. However, access to DALL-E 3 has been limited behind a paywall, requiring users to subscribe to a ChatGPT Plus plan for $20 per month. Now, Microsoft has provided a free workaround with its revamped Bing Image Generator.

Key Upgrades to Bing Image Generator

On Wednesday, Microsoft announced several exciting upgrades to its Bing Image Generator. The most significant of these updates is that the tool is now powered by OpenAI’s latest DALL-E 3 model, specifically the PR16 version. This enhancement promises users an experience that is both twice as fast and higher quality than before, providing a smoother and more impressive image generation process.

In addition to the model upgrade, Microsoft has revamped the Image Creator standalone site, making it more user-friendly. The redesigned interface is clean, minimalistic, and intuitive, offering a much better user experience compared to the previous design, which was darker and more overwhelming. This new look makes navigation easier and more enjoyable for users creating their images.

Easier Access and Improved Features

The Bing Image Generator is now more accessible than ever. Users can generate images directly from the Bing and Microsoft Edge search bars. All they need to do is type a prompt, such as “create an image of [insert your idea here],” and the generator will produce the requested image. This seamless integration makes it easier to create images on the fly without having to visit a separate website.

Furthermore, Microsoft has added enhanced sharing capabilities, allowing users to share their creations on popular social media platforms like Facebook, WhatsApp, and Instagram. These social media integrations make it much easier for users to showcase their AI-generated artwork and share it with friends or followers.

Commitment to Safety and Quality

To ensure a safe experience, Microsoft has implemented several guardrails. These include blocking potentially harmful prompts and including a watermark in the bottom corner of each generated image. Additionally, content credentials are now attached to each image to provide context and authenticity. These measures ensure that users can enjoy the tool without worrying about the creation or distribution of inappropriate content.

Free Access and Comparison with OpenAI

One of the standout features of the Bing Image Generator is its free access to DALL-E 3, which would otherwise require a $20/month subscription for unlimited access through OpenAI’s platform. While free users of OpenAI’s DALL-E 3 are limited to only two images per day, Microsoft’s version removes this restriction entirely, offering unlimited creations for all users without any additional fees.

However, while DALL-E 3 does produce impressive outputs, some users have noted that the results may not always be as realistic as those generated by other models, such as Google’s Imagen 3. Despite being more expensive, DALL-E 3 sometimes falls short in terms of realism, though it still offers unique and valuable results depending on the type of prompt.

Conclusion

With the launch of these upgrades, Microsoft has made DALL-E 3 accessible to a much broader audience. Users can now create high-quality images with ease, directly from their search bars, and share their creations across social platforms. These changes, coupled with the safety features and free access, make the Bing Image Generator a powerful and user-friendly tool for anyone looking to explore AI-generated imagery.

As Microsoft continues to roll out these updates, users can expect even more enhancements, including a refreshed homepage for mobile devices in the near future.

As we approach the end of the year, the Windows 11 24H2 update released in October continues to cause problems for many users. Microsoft has acknowledged two new issues, both of which are causing frustration for those affected. According to Neowin, one of the problems disrupts audio output, while the other impacts Auto HDR functionality.

Audio Output Issue

The audio issue affects devices with Dirac Audio and the cridspapo.dll file, which is responsible for enhancing audio clarity and precision. This bug prevents Windows 11 from outputting sound to a range of devices, including integrated speakers, Bluetooth speakers, headsets, and other audio peripherals. Microsoft has not disclosed which specific manufacturer is impacted, but the problem is significant enough that there is no immediate workaround for users who are affected.

Auto HDR Bug

The second issue involves Auto HDR, a feature in Windows 11 that automatically enhances standard dynamic range (SDR) content by converting it to high dynamic range (HDR). However, due to this bug, users are experiencing various glitches, including incorrect colors and even system crashes. The only known workaround at the moment is to completely disable the Auto HDR feature.

Microsoft’s Response

In response to these issues, Microsoft has temporarily blocked the Windows 11 24H2 update on affected systems. The company plans to lift the block once the bugs are resolved and a fix has been issued.

For now, users experiencing these issues will need to either wait for an official fix or, in the case of Auto HDR, disable the feature temporarily.

This isn’t the first time Windows 11 24H2 has caused problems, especially for gamers.For users who have been blocked from installing the update, it may be beneficial to wait until Microsoft releases a stable version. Given the recurring issues, waiting a few months before updating could prevent further frustration and ensure that any remaining kinks are worked out.

On December 17, Salesforce introduced Agentforce 2.0, a cutting-edge iteration of its AI-powered platform designed to integrate customizable, semi-autonomous AI agents into enterprise workflows. The new version enhances capabilities, empowering businesses to streamline processes with an AI-driven workforce, capable of handling complex tasks across various departments.

What’s New with Agentforce 2.0?

Agentforce 2.0 offers an array of new features aimed at improving business operations. The platform’s core innovation lies in its ability to allow businesses to choose from a library of pre-built AI agents or create custom agents using natural language prompts. These agents can follow multi-step plans, adhere to conditional workflows (e.g., “if/then” statements), and execute actions across systems and applications seamlessly.

Key Features of Agentforce 2.0:

• Pre-built Skills: Salesforce has introduced several pre-configured skills for specific tasks, including Sales Development, Sales Coaching, Marketing Campaigns, and Commerce Merchant. These skills can help automate and streamline routine processes, significantly reducing the workload on human employees.
• Tableau Integration: Agentforce agents now offer Skills for Analytics and Insight within Tableau. These capabilities help businesses track and analyze agent performance, providing valuable data visualizations for enhanced decision-making.
• Slack Integration: Slack is deeply integrated into the Agentforce experience. Through Slack Actions, users can automate routine updates and have agents send project updates via direct messages. In January, the Agentforce Hub will be available within Slack, making it easier for teams to interact with AI agents, access relevant data, and automate tasks within Slack’s communication channels.

These features allow businesses to integrate Agentforce 2.0 directly into their existing tools, including Salesforce CRM, Slack, Tableau, and Mulesoft, creating a powerful, AI-powered digital labor force.

Customizing Your AI Agents with Natural Language Prompts

For businesses that don’t find an exact match in the pre-built library, Salesforce has introduced Agent Builder. This tool enables users to create custom AI agents using simple, natural language commands. For example, users can instruct the AI to “Onboard New Product Managers,” allowing businesses to tailor the system to their unique needs without technical expertise. Additionally, Slack Actions will be incorporated into Agent Builder, making it even more adaptable.

Integration of the Atlas Reasoning Engine

A major leap forward in Agentforce 2.0 is its Atlas Reasoning Engine, which powers advanced AI reasoning and retrieval-augmented generation (RAG). RAG allows the AI to access unstructured data and pull insights from various parts of the Salesforce Platform to deliver more nuanced and informed responses.

For instance, Salesforce highlighted an example where advanced reasoning could be used to answer a complex question like, “What investment vehicle would be best for my child’s college fund based on my current income and risk preferences?” Through RAG, Agentforce can retrieve data from across Salesforce systems to ensure more accurate, contextually relevant recommendations.

Expanding the Scope of AI in Enterprises

Salesforce’s emphasis on AI is evident in its vision to make Agentforce 2.0 the cornerstone of enterprise operations, allowing employees to focus on more value-added tasks while automating routine functions. The power of Agentforce 2.0 lies in its adaptability, making it applicable to virtually any business function, from customer service to marketing and HR.

Melik Khoury, president and CEO of Unity Environmental University, shared how his institution is leveraging Agentforce to improve workflows. By automating common queries, such as financial aid details or class registration, the university is freeing up staff time to focus on more personalized guidance for students. This represents a significant shift in how businesses and institutions can integrate AI for improved efficiency.

Looking Ahead: A Future of AI-Powered Digital Labor

Salesforce’s aggressive investment in AI tools like Agentforce 2.0 is part of a broader strategy to integrate generative AI into its entire suite of products. The company aims to provide enterprises with AI-driven solutions that not only automate basic tasks but also enhance decision-making processes with advanced reasoning. Claire Cheng, VP of Machine Learning and Engineering for Salesforce AI, stated that Salesforce expects the reasoning engine to become one of the key factors businesses will consider when evaluating digital labor options.

Availability and Next Steps

While Agentforce 2.0 will be fully available in February 2025, businesses can begin deploying it within Slack as early as January 2025. The expansion of generative AI into business workflows represents a new chapter for Salesforce and the AI industry, offering enterprises a flexible, powerful toolset to transform their operations with minimal manual intervention.

On Wednesday, GitHub, owned by Microsoft, announced a major update to its Copilot AI-powered code completion and pair programming tool. For the first time, GitHub is offering a free version of Copilot, which will also be available by default with Microsoft’s Visual Studio Code (VS Code) editor. Previously, developers had to pay a monthly subscription fee, starting at $10 per month, with only verified students, teachers, and open-source maintainers getting free access.

This update marks a significant step for GitHub as it continues to expand its platform. The company also shared that it now has 150 million developers using its platform, a significant jump from 100 million just earlier this year.

In an exclusive interview ahead of the announcement, GitHub CEO Thomas Dohmke reflected on the company’s journey. “My first project at GitHub in 2018 was the introduction of free private repositories, which we launched in early 2019. Then we had a version two with free private organizations in 2020. We introduced free GitHub Actions entitlements, and at my first Universe conference as CEO, we announced free Codespaces. It only felt natural to extend this philosophy to Copilot, offering a free version, not just for students and open-source maintainers,” Dohmke shared.

What’s Included in the Free Copilot Plan?

While the free Copilot plan opens up the tool to a broader audience, it does come with certain limitations. The free version is designed for occasional users, not for developers working on large projects or enterprises.
Here are the key details of the free Copilot plan:

• 2,000 code completions per month: Each code suggestion from Copilot counts towards this limit, not just accepted ones.
• Foundation Model Access: Free plan users are limited to Anthropic’s Claude 3.5 Sonnet and OpenAI’s GPT-4o. Paid plans, on the other hand, include access to Google’s Gemini 1.5 Pro and additional OpenAI models like o1-preview and o-mini.
• Copilot Chat: Users on the free plan can send up to 50 chat messages per month.
• Extensions and Skills: Despite these limitations, users still get full access to all Copilot Extensions and skills.

Dohmke explained that the team analyzed years of Copilot usage data to determine the balance between occasional users and professional developers. The goal was to make it as easy as possible for developers to start using the tool and be productive without needing to jump through extra hoops.

The free version of Copilot will work across multiple platforms, including VS Code, Visual Studio, and JetBrains IDEs, as well as directly on GitHub.com. This broad compatibility ensures that developers can integrate Copilot into their existing workflow seamlessly, no matter which development environment they prefer.

Expanding Access and Competing

Since its launch, GitHub Copilot has become a prominent player in the AI coding tool market. However, competition has intensified with companies like Tabnine, Qodo (formerly Codium), and AWS offering similar tools, many with free plans. To remain competitive and expand its user base, GitHub’s decision to introduce a freemium model for Copilot is a logical step.

By offering a free version of Copilot, GitHub aims to make its tool more accessible, particularly in emerging markets where a $10 subscription fee may be prohibitively expensive. This approach aligns with GitHub’s broader mission of enabling one billion developers globally. In countries like Brazil, India, and Pakistan, where income disparities are significant, providing a free version of Copilot helps break down barriers to entry and empowers more people to pursue careers in software development.

In addition to supporting global developers, GitHub’s free plan also simplifies access for students. While the platform has offered free access to students before, they had to go through a verification process, which created friction. With the new freemium model, students can sign up and immediately start using Copilot, making the tool more accessible and user-friendly.

This move not only strengthens GitHub’s competitive position but also supports its long-term vision of democratizing software development, offering a powerful AI tool to developers regardless of their financial situation or geographic location.

The T-Mobile Starlink Direct-to-Cell beta program is set to change the landscape of mobile connectivity by providing coverage in areas that traditional cellular networks have long struggled to reach. As of December 2024, T-Mobile is accepting registrations for this exciting new service, which aims to eliminate the notorious “dead zones” that have plagued rural areas, remote locations, and even some airspaces. This initiative leverages Starlink’s growing constellation of low-Earth orbit (LEO) satellites to deliver high-speed connectivity where conventional towers and networks fall short.

The beta program, launching in early 2025, will initially support text messaging, allowing T-Mobile customers to send and receive messages even in the most isolated areas. The service uses a dedicated network of 300 Starlink satellites and promises to cover up to 500,000 square miles of land in the U.S. that are currently underserved by traditional cellular infrastructure​.

While the initial launch will focus on text messaging, T-Mobile and Starlink have ambitious plans to expand the service. By mid-2025, they aim to introduce voice and data capabilities. This will make it possible to not only send texts but also make calls and access the internet in areas where terrestrial networks cannot provide service​. 

The service will evolve to include more functionalities, with Internet of Things (IoT) connectivity also expected by the end of 2025​.

How Does T-Mobile’s Starlink Direct-to-Cell Work?

Unlike traditional cell towers that rely on terrestrial infrastructure, Starlink’s Direct-to-Cell technology utilizes satellites in orbit to connect directly to mobile devices. This innovation means that as long as your phone can “see the sky,” you’ll have access to text messaging, even when you’re out of range of conventional cellular towers​. T-Mobile’s integration of this technology ensures that users won’t have to manually search for signals or navigate cumbersome connection processes. Messages will be sent and received as easily as they would on a regular network​.

Who Can Benefit from T-Mobile Starlink Direct-to-Cell?

This service is designed with a broad spectrum of users in mind. Its primary audience includes people who often find themselves in areas outside the reach of traditional cellular networks. This includes:

Rural Communities: Areas that are underserved or completely ignored by traditional cellular providers will benefit significantly from this service.
First Responders: Emergency teams working in disaster-stricken areas where communications are critical will be able to rely on Starlink to stay connected​.

Outdoor Enthusiasts: Hikers, mountaineers, and campers can enjoy peace of mind, knowing they can stay in touch even when deep in remote areas.
Travelers and Pilots: T-Mobile Starlink is also looking to expand its reach to locations like airplanes, offering connectivity in-flight, which could revolutionize how people stay connected during long-haul journeys.

FAQs

1. How do I sign up for the T-Mobile Starlink beta program? 

You can sign up directly through T-Mobile’s website. The beta program is open to select T-Mobile postpaid voice customers with compatible devices​.

2. Will T-Mobile Starlink Direct-to-Cell work with my current phone? 

The service is designed to work with most modern mobile phones. However, the initial beta program will support only specific “optimized” smartphones​.

3. When will T-Mobile Starlink support voice and data? 

The voice and data capabilities are expected to be available by mid-2025​.

4. Is T-Mobile Starlink free?

The beta program is free for T-Mobile postpaid customers, but once the service expands, there might be a monthly fee, depending on your mobile plan​.

5. What areas will T-Mobile Starlink cover?

 Initially, the service will cover up to 500,000 square miles of land in the U.S. that are not served by traditional cell towers​.

6. Can I use T-Mobile Starlink Direct-to-Cell while traveling abroad?

 The service is currently limited to U.S. territory, but T-Mobile has partnerships with several international carriers, so expanded global coverage may be available in the future​.