. . AI: Source citations in ChatGPT! (12.13.23)

Hi friends, it's big company day today in AI. Not a single startup in today’s top stories, unless you count Anthropic. Some pretty interesting AI news nonetheless!

I aim for this to be the easiest, fastest way for you to find AI news you can use to fuel and inspire your innovation. As I have throughout my editorial career, I believe that watching a large community of specialists to see what they get most excited about is a great way to discover and differentiate important developments in a world of information overload.

Here’s what we found today, who we think it will impact first, and when,

First impacted: ChatGPT users
Time to impact: Short

Axel Springer and OpenAI have announced a global partnership with the goal of integrating journalism and AI technologies. The collaboration involves enhancing OpenAI's ChatGPT with content from Axel Springer's media brands (including Business Insider, Politico, BILD and WELT - with source citations and links!) and using this content to advance the training of OpenAI's LLMs, according to the companies. Once Insider posts start being cited in ChatGPT, I’m guessing a whole lot of other publishers will rush to do the same. [Partnership with Axel Springer to deepen beneficial use of AI in journalism] Share by email

First impacted: AI developers, Data scientists
Time to impact: Short

The Machine Learning Foundations team at Microsoft Research has launched Phi-2, a language model with just 2.7 billion parameters, which they say matches or outperforms models up to 25 times larger on complex benchmarks. Phi-2, now available in the Azure AI Studio model catalog (and coming soon to HuggingFace), is said to demonstrate exceptional reasoning and language understanding capabilities due to innovations in model scaling and training data curation. [Phi-2: The surprising power of small language models] Share by email

First impacted: App Developers, AI Researchers
Time to impact: Short

Google has launched Gemini Pro, a new AI model that the company says outperforms other models of similar size on research benchmarks and supports 38 languages across over 180 countries. Pro wasn't quite ready for launch with the other levels of Gemini, the company said last week, but now it is. Developers can access Gemini Pro for free through Google AI Studio, a tool that allows users to create prompts and obtain an API key for app development, offering a free quota of 60 requests per minute. [It’s time for developers and enterprises to build with Gemini Pro] Share by email

First impacted: Claude users
Time to impact: Short

Anthropic has launched a new tool, Claude for Sheets, integrating Claude with Google Sheets and allowing users to interact with Claude directly within cells. The tool, which requires an API key for activation, offers two main features for accessing Claude, an interactive guide on prompt engineering, and the ability to quickly test and tweak prompts with real-time results. [Claude for Google Sheets] Share by email

First impacted: New GPT-4 users
Time to impact: Short

Sam Altman has announced on X that subscriptions for ChatGPT Plus are once again available. "Thanks for your patience while we found more gpus," he said. Simple as that! [via @sama] Share by email

First impacted: Tesla customers, Robotics enthusiasts
Time to impact: Medium to long

Tesla has shared a video of its Optimus Gen-2, a humanoid robot with notable hardware enhancements, as per tweets from Tesla Optimus and team member Julian Ibarz. The robot's capabilities included in the video included a delicate task of handling an egg and dancing. [via @Tesla_Optimus] Share by email

First impacted: AI developers, Digital artists
Time to impact: Short

Google DeepMind has launched Imagen 2, a text-to-image diffusion technology, which is now accessible on Google Cloud's Vertex AI platform. According to DeepMind, this technology improves image-caption understanding, generates more realistic images, offers advanced image editing capabilities, and is integrated with SynthID, a toolkit for watermarking and identifying AI-generated content. [Imagen 2 - our most advanced text-to-image technology] Share by email

That’s it! More AI news tomorrow. Thanks for subscribing! Please tell a friend!