Large Language Model Llama 2 No Longer Free for Big Cloud Providers – A Strategic Move Shaping the Future of AI in the Cloud Market
Meta to Charge Cloud Providers for AI Tech That It Said Was Free
Meta Platforms, formerly known as Facebook, has announced a change in its policy regarding the use of its artificial intelligence (AI) large language model (LLM), Llama 2, by major cloud-computing companies. The company had initially stated it would provide the technology for free, but it has now said that it will charge cloud providers like Microsoft, Amazon, and Google who plan to resell its services.
Meta’s CEO, Mark Zuckerberg, explained that while the company expects to gain only a small amount of revenue from this in the short term, it could become a significant income stream in the long run. The announcement represents a departure from a previous statement on July 18, where Meta had revealed a partnership with Microsoft and Amazon to offer its LLM for commercial use on their cloud services free of charge.
Despite the policy shift, Zuckerberg maintains that Llama 2 remains open and continues to highlight the benefits of this approach. He clarified that while Meta has made the technology open-source and freely available to a broad community of developers, large cloud companies would now be required to negotiate a business arrangement with Meta to use it.
Photoshop’s generative AI feature can now expand images seamlessly
Adobe has beefed up Photoshop by sticking in its Firefly AI models. First, this let users add, change, or chuck out parts of pictures, just by typing simple commands. Now, they’ve upped the ante with a new gizmo called ‘Generative Expand’ that makes changing the size of a picture a piece of cake.
Here’s the lowdown: you pull the Crop tool to stretch out the canvas. Then hit the ‘Generate’ button. Boom! The new empty space fills up with AI-created stuff that fits right in with the rest of your picture. This is perfect for fixing up pics where something’s half out of the frame, your aspect ratio is wonky, or something in the shot is out of whack.
The cool part is you can use Generative Expand with or without a command. No command? No problem. The AI will fill up your extra space, fitting it to the rest of the picture. Got a command? The new stuff will be just what you asked for. As Adobe puts it, “With Generative Expand, you can spend less time fussing and more time playing with your images.”
That ain’t all, either. Adobe’s letting its Firefly features in Photoshop understand commands in over 100 languages, a big step up from just English. This opens doors for creators worldwide.
Artifact is getting an AI-powered text-to-speech feature to read you the news
Artifact, the news app whipped up by Instagram’s original creators, Kevin Systrom and Mike Krieger, is jazzing up its game. It’s rolling out an AI-powered feature that’ll read the news to you, thanks to a deal with Speechify.
You can choose celebrity voices like Snoop Dogg and Gwyneth Paltrow to do the reading. Plus, there’s over 30 English voices up for grabs, says Artifact’s top dog, Robby Stein.
To get this show on the road, you tap a “play” button while you’re reading an article. You can tweak things like how fast it’s read out (up to 4.5 times regular speed!) and pick your favorite voice. Even better, the audio keeps rolling even if you navigate away from the app.
ServiceNow, NVIDIA, and Accenture Team to Accelerate Generative AI Adoption for Enterprises
ServiceNow, NVIDIA, and Accenture have partnered to launch the AI Lighthouse program, aimed at accelerating the development and adoption of enterprise generative AI capabilities. The program is designed to assist customers in designing, developing, and implementing new generative AI use cases.
The AI Lighthouse program will benefit enterprise customers by reducing manual work in customer service, promoting self-service options, generating content automatically, and boosting developer productivity with intelligent code recommendations.
The partnership aims to apply generative AI to various industries, recognizing the transformative potential that this technology has to reinvent workflows, strengthen services, differentiate companies, and elevate performance levels.
ServiceNow will serve as the front-end workflow automation and intelligence platform, NVIDIA will offer full-stack computing for model training and tuning, and Accenture will utilize its expertise to implement use cases for customers. The program intends to explore generative AI applications across IT service management, customer service management, and employee experience.
Microsoft’s and Alphabet’s results indicate the AI game is more of a long-term strategy
The recent quarterly results of Alphabet and Microsoft have illustrated that investments in AI are more of a long-term strategic play. Both companies have made significant investments in AI, and though revenues are beginning to materialize, the main financial impact is expected to be seen in future quarters.
Both Alphabet and Microsoft have invested heavily in the hardware necessary to handle AI computation workloads, including data centers. Alphabet mentioned during its earnings call its investment in “AI supercomputer options with Google TPUs and advanced Nvidia GPUs, and recently launched new A3 AI supercomputers powered by Nvidia’s H100.”
Meta posts strong revenue growth as it touts bet on AI tools
Meta’s stock is looking pretty spry after reporting its first double-digit revenue growth since 2021. CEO Mark Zuckerberg’s big bet on artificial intelligence (AI) seems to be working wonders already, even as he works to give the company a much-needed jumpstart.
Over the past few months, Zuck’s been making waves with a major restructuring, slimming down management, and letting go of around 20,000 employees – all part of his “year of efficiency.” And it seems his bets on AI – particularly generative AI that creates text, images, and code – are giving Meta a boost. He plans to build chatbots to help businesses and creators engage with users and use AI to power up advertising and productivity.
Meta’s going toe-to-toe with the big dogs like OpenAI, Google, and Microsoft with its own language model, Llama 2. But there are still big questions about how AI will contribute to Meta’s bottom line. Zuck is eyeing deals with giants like Microsoft, Amazon, and Google who might use Meta’s technology on a large scale.
In terms of figures, Meta’s looking at projected revenue between $32 billion to $34.5 billion for the current quarter, blowing past analysts’ predictions. Second-quarter revenue grew by 11% to hit $32 billion, driving up Meta shares by 6.8% in after-hours trading.
Samsung extends cut in memory chip production, will focus on high-end AI chips instead
Samsung Electronics, the world’s leading memory chip maker, is continuing to reduce its production of memory chips after suffering a significant operating loss in the second quarter of 2023. This follows on from April’s decision to cut back production in response to weak consumer device demand, which led to the company’s worst quarterly profits since 2009.
Despite these challenges, Samsung is looking to refocus its efforts on high-performance memory chips, such as high bandwidth memory (HBM), in response to increasing demand from AI applications. HBM chips are used in AI, 5G, the Internet of Things (IoT), and graphic processing applications, offering faster data processing and lower power consumption than traditional NAND chips.
SEC proposes AI crackdown for Wall Street firms
The U.S. Securities and Exchange Commission (SEC) is proposing new regulations on how investment firms use artificial intelligence (AI), seeking to prevent the technology from being used in ways that might harm their customers’ interests. The proposal comes in the wake of a review of the 2021 meme-stock frenzy, where investment platforms used gamified interfaces and behavioral prompts to encourage riskier trades by retail investors.
The proposed rules aim to extend existing requirements for investment advisors to act in their clients’ best interests, to include AI-powered features that use personal data to influence user behavior. If enacted, investment firms would be required to identify and eliminate any potential conflicts of interest arising from their use of AI, as well as establish written policies, procedures, and records to prevent violations.
Singapore looks for generative AI use cases with sandbox options
Singapore is pushing for the development and testing of generative AI applications by introducing two sandboxes for government agencies and businesses. These sandbox environments, powered by Google Cloud’s AI tools, will be free for three months for up to 100 selected use cases or organizations.
One sandbox is for government agencies and the other is for local businesses. Through a series of workshops, participants will receive training from Google Cloud engineers to identify use cases that can be developed using the sandbox’s tools.
These sandbox initiatives are part of a collaboration between Google and the Singapore government to promote AI adoption. Google Cloud’s Asia-Pacific VP, Karan Bajwa, stated that generative AI requires a different approach within an organization compared to general-purpose, public-facing generative AI applications. He emphasized the importance of data security, cost optimization, and industry-specific tuning for enterprise applications.
Stay Ai’s tech helps Shopify brands get customers to make that second purchase
Stay Ai, an artificial intelligence company founded by serial e-commerce entrepreneurs Gina Perrelli, Pierson Krass, and Taylor Cawiezell in 2021, helps Shopify brands build their subscription programs, reduce customer cancellations, and increase customer lifetime value. The platform offers a recurring commerce enablement system for subscriptions, memberships, and clubs with loyalty tools such as SMS chatbots, surveys, and rewards programs.
Stay Ai’s proprietary ExperienceEngine, an AI-powered system, enables merchants to perform marketing procedures, use A/B testing, and apply predictive analytics and machine learning models for more efficient marketing decision-making. For instance, one of the company’s clients, Olipop, managed to boost subscriber revenue by 35% within 90 days of using Stay Ai’s tools.
The Future Of Manufacturing: Generative AI And Beyond
The future of manufacturing is getting a major overhaul with new tech like generative AI, the metaverse, and web3. This is changing how companies work, making them more efficient and customer-focused.
In the metaverse, virtual reality (VR) and augmented reality (AR) are being used to improve employee training, quality control, and product design. Digital twin technology lets companies simulate products and factories, reducing costly mistakes. Creating virtual prototypes also speeds up product development and saves money.
Web3 tech, like blockchain and smart contracts, is enhancing logistics and supply chain processes. It ensures better data security, transparency, and faster customs clearance. Smart contracts automate tasks, reducing costs and speeding things up.
AI, especially generative AI, is another game-changer. It uses vast data to improve decision-making, optimize production, and reduce waste. Generative AI generates new designs, saving time and resources.
Wayfair’s new app uses generative AI to transform your space
Online furniture and home goods retailer Wayfair has launched Decorify, an app that uses generative AI to help customers redesign their rooms in different styles. Users can upload a photo of their room and select a design style, such as “Modern Farmhouse” or “Bohemian”. The AI then generates a “shoppable”, “photorealistic” image of the room furnished in the selected style.
The AI technology behind Decorify utilizes image-generating AI diffusion models like OpenAI’s DALL-E 2. The app is integrated into Wayfair’s web platform and does not require a separate download. Currently, it only supports living room redesigns, but Wayfair plans to expand this in the future.