The AudioCraft family of models are capable of producing high-quality audio with long-term consistency, and they’re easy to use. The tech brands all reported dips in advertising demand, which affected their revenue outlooks. Google’s Search and YouTube businesses reported year-over-year ad sales declines of 2% and 8%, respectively. Meta’s ad business was down 4% year over year, and Snap’s revenue was flat year over year. Meta’s AI Sandbox highlights the Facebook parent company’s generative AI strategy and how generative AI is changing advertising.
AI Sandbox is available to select advertisers at the moment with expanded access in July. Within the next month, brands will be able to switch over their manual campaigns to Advantage+ at the click of a button, a bid to spur even more brands to try out the AI tools. Automated reports that compare the outcomes of Advantage+ campaign to manual sales campaigns are in the process of going live as well. This allows researchers to test new approaches to limiting or eliminating the risk of bias, hallucinations, and toxic responses that users have been exposed to when interacting with an AI. In addition to new monetization opportunities, he notes that with the capabilities of generative AI, creators will be able to level up their storytelling and production value. Instead of simply transforming their face into an animal or adding glowing layers, users can see themselves as they were as a teenager or get a preview of what they would look like with a glamorous makeover.
The company’s process uses a technique called supervised fine-tuning to train text-based transformer models using a dataset of licensed images and captions from Shutterstock, allowing for the parsing of complex text and objects to better follow the inputs of a user. We’re open-sourcing these models, giving researchers and practitioners access so they can train their own models with their own datasets for the first time, and help advance the field of AI-generated audio and music. As Meta refines the best ways to develop generative AI systems, this Community Forum will give us insights into how people would like models to behave for nuanced topics, and therefore inform future product and policy considerations.
Meanwhile, Google recently debuted Bard and is now planning a more capable large language model dubbed Gemini. Although Meta plans to first launch its AI agents on Messenger and WhatsApp, Zuckerberg envisions a broader future for these technologies, potentially expanding across its family of applications and eventually into smart glasses. In this scenario, users could wear smart glasses, similar to Meta’s Yakov Livshits Ray-Ban Stories, to interact with AI agents via voice. During an investor call, Zuckerberg first hinted at Meta’s plans to investigate chat experiences in WhatsApp and Messenger, as well as visual creation tools for social media posts on Facebook and Instagram. In collaboration with Ahmad Al-Dahle, VP of Engineering, Generative AI at Meta, Zuckerberg shared more about these projects during the meeting.
The brand also uses many of Meta’s “Advantage” AI- and machine learning-powered automated advertising tools, and Plofker is a big proponent of the potential for generative AI to further streamline marketing processes. Meta has been actively releasing new AI tools alongside many other tech giants, including Google and Microsoft, in a race to develop and deploy the most powerful models. In May, the company announced that it was accepting “early testers” of the products via its AI Test Kitchen platform.
Meta Releases ‘Code Llama’ Generative AI Model to Assist in Code ….
Posted: Thu, 24 Aug 2023 07:00:00 GMT [source]
Meta is leveraging its formidable data science team and computing infrastructure to advance state-of-the-art models like CM3leon. While diffusion-based AI like MidJourney’s has grabbed headlines, Meta is betting on autoregressive transformer architectures (the same tech used by ChatGPT). The company claims CM3leon needs 5x less training compute than other comparable methods. Meta has put AI front and center in recent months, viewing it as a central pillar to growing its TikTok clone Reels and its longer-term metaverse ambitions. Meta’s revenue grew again in the opening stretch of the year after several consecutive quarters of declines, a bump leadership attributed to the new efficiency enabled by AI.
Meta and companies like ours input “values” which guide the AI model and can help protect against bias and unintended consequences by giving it a way to evaluate its own outputs. We think it’s important these values are reflective of different viewpoints from throughout society, and the Community Forum will build on the feedback we’ve received from experts as part of our policy development processes. For most people, it’s impossible to know what to think when confronted by new technologies that inspire such hyperbolic optimism and pessimism. Microsoft has been rushing to get ChatGPT into Bing, and Google has now accelerated its plans. It’s possible Google was being more cautious about unleashing a nascent AI brain for consumers to poke and prod. Now, Google is getting ready to uncage LaMDA, a language model that is similar to ChatGPT.
Yakov Livshits
Founder of the DevEducation project
A prolific businessman and investor, and the founder of several large companies in Israel, the USA and the UAE, Yakov’s corporation comprises over 2,000 employees all over the world. He graduated from the University of Oxford in the UK and Technion in Israel, before moving on to study complex systems science at NECSI in the USA. Yakov has a Masters in Software Development.
Mark Zuckerberg’s Remarks at AI Forum.
Posted: Wed, 13 Sep 2023 17:20:02 GMT [source]
Images generated by CM3leon, using text prompts such as “a small cactus wearing a straw hat and neon sunglasses in the Sahara desert” (left) and “a stop sign in a Fantasy style with the text ‘1991’” (right). CM3leon recognizes both text and visuals, allowing both image-to-text and text-to-image Yakov Livshits generation. Eventually, generative AI could give anyone the power to create in augmented reality. “Imagine playing around with your kids wearing AR glasses and pointing, ‘Oh my gosh, there’s a pirate ship and a big monster,’ and we can bring those to life using generative AI,” Spiegel said.
Meta said it wasn’t publically releasing the model, however, citing the potential risks of the technology. There are several actions that could trigger this block including submitting a certain word or phrase, a SQL command or malformed data. Meta’s release of its generative AI music tools comes shortly after Google launched similar tools that turn text into music, called MusicLM. In a recent interview with Cointelegraph, the CEO of the Recording Academy, Harvey Mason Jr., also likened the emergence of AI-generated music to the early days of synthesizers coming onto the music scene. VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Zuckerberg seems determined to put his foray into the metaverse behind him and is spending billions to catch up with AI.
Meta, Google and Snap are preparing for generative AI, as the technology has caught on with consumers through products like ChatGPT, and they see it as a powerful tool in advertising and creator economies. Today, advertisers on Facebook and Instagram often manually generate several variants of titles, description text, and images for each ad. Those variants are launched simultaneously, and Facebook’s ad algorithms evaluate ad performance and automatically shift impression volume to the best performing ads.
That gives us the opportunity to now take that technology, push it forward, and build it into every single one of our products. He further emphasized Meta’s unique role in the industry, as the company is poised to bring these capabilities to billions of people in unparalleled ways. Axios initially reported the news of the consumer-facing AI agents and the photo-editing tools.
Meta has unveiled its plans for new generative AI tools intended for consumer products, including Instagram, WhatsApp, and Messenger, as well as some that will be used internally within the company. In a recent all-hands meeting, Meta CEO Mark Zuckerberg announced several AI technologies in different stages of development. These include AI chatbots for Messenger and WhatsApp, AI stickers, along with tools that will enable photo editing in Instagram Stories. Additionally, internal-only products like an AI productivity assistant and an experimental interface for interacting with AI agents powered by Meta’s large language model, LLaMA, were announced. Meta this week showcased its work on new generative AI technologies for its consumer products, including Instagram, WhatsApp and Messenger, as well as those being used internally at the company. This included an experimental internal-only interface to an “agents playground” powered by its large language model LLaMA where Meta employees could have conversations with AI agents and then provide feedback to help the company improve its systems.
That’s the promise of AudioCraft — our latest AI tool that generates high-quality, realistic audio and music from text. Omneky, which presented at TechCrunch Disrupt last year, was using OpenAI’s DALLE-2 and GPT-3 to create campaigns. Movio, which is backed by IDG, Sequoia Capital China and Baidu Ventures, is using generative AI to create marketing videos.
The advantages of this are that it requires less compute power and resources to retrain in order to test new approaches and use cases. Models such as this could conceivably run on far smaller devices than the cloud servers that are needed for ChatGPT or Bard – potentially opening the way for self-contained instances to run on personal computers or even smartphones. This could have important implications for businesses that want to use generative language models while keeping their data private. Not to be outdone, Spiegel also sees the creative possibilities of generative AI, especially around Snap’s augmented reality projects. Snap’s Lens Studio has 300,000 creators building augmented reality filters, which are animated 3D digital creations.
