Weekly AI - 30 Sept 2023
Welcome to your weekly AI newsletter! We've got a jam-packed edition highlighting major advancements across conversational AI, generative creativity, and more.
Kicking things off is OpenAI rolling out new voice and image capabilities to ChatGPT, enabling more natural conversational interactions. This brings us one step closer to AI assistants that can truly understand and respond like humans.
In the world of generative AI, Adobe launched early web versions of Photoshop and Illustrator, making these powerful creative tools more accessible online. Meta also revealed new AI features for social media to transform photos and create custom animated stickers.
Other key developments include DeepMind's AI system that can predict the effects of genetic mutations to help diagnose diseases, and AWS launching services to accelerate generative AI innovation.
Stay tuned over the coming weeks as we bring you the latest on how AI is transforming the way we communicate, create, and solve problems. As always, let me know what you think of this week's selection. Enjoy the newsletter!
This newsletter is also available in Spanish and Catalan.
ChatGPT Gets New Voice and Image Capabilities
On September 25th, OpenAI announced the rollout of new voice and image capabilities in ChatGPT, allowing users to have more natural conversations by speaking to the AI assistant and showing it images. This more intuitive interface enables use cases like getting recipes by taking pictures of your fridge, or helping kids with math homework by circling problems in a photo. The new features are launching first to Plus and Enterprise users over the next two weeks. This update marks an important advancement in making AI systems easier to interact with in everyday life.
Spotify Launches AI-Powered Voice Translation for Podcasts
On September 25th, Spotify announced an exciting new feature that uses AI to translate podcast episodes into other languages while retaining the speaker's original voice. According to Spotify, this more personalized approach will help listeners around the world discover podcasts in an authentic way. The feature is being piloted with popular shows like Armchair Expert and Lex Fridman Podcast. As AI technology continues to advance, innovations like Voice Translation have the potential to break down language barriers and increase accessibility to content worldwide.
DeepMind's AlphaMissense Predicts Effects of 71 Million Genetic Mutations
DeepMind released AlphaMissense, an AI tool that classifies the effects of genetic "missense" mutations. Missense mutations are changes in a single DNA letter that can lead to diseases like cystic fibrosis if they disrupt protein function. AlphaMissense scored all 71 million possible missense mutations, finding 89% were likely benign or pathogenic. This catalogue of mutations will help researchers pinpoint disease-causing variants faster. DeepMind hopes AlphaMissense will accelerate studies into the root causes of diseases and development of new treatments.
Microsoft Brings New AI Features to SwiftKey Keyboard
Microsoft announced this week it is bringing new AI-powered features to its SwiftKey keyboard app for iOS and Android. The update, rolling out now, ties the app to Microsoft's AI systems to enable features like AI-powered grammar and spelling correction, AI stickers and avatars, and more. This integration allows Microsoft to leverage its investments in AI to improve one of the most widely used mobile keyboard apps. With over 500 million installations, enhancing SwiftKey with AI could significantly increase mainstream consumer exposure to the technology.
Microsoft Upgrades Bing with New AI Capabilities
Microsoft announced several new AI capabilities coming to its Bing search engine later this year. On September 22nd, the company said Bing will integrate OpenAI's latest DALL-E 3 model to generate images. The new version promises more realistic image rendering. Bing will also gain the ability to provide more personalized search results and answers in chat by taking into account users' prior interests and chat history. Additionally, Bing will add invisible digital watermarks to all AI-generated images to provide more transparency. These upgrades aim to improve Bing's helpfulness by leveraging the latest advances in AI.
Amazon Bets Big on AI with $4 Billion Investment in Anthropic
Amazon announced a massive $4 billion investment in AI startup Anthropic. This deal displays Amazon's commitment to competing with other tech giants like Microsoft and Google in the race to develop powerful artificial intelligence. The funding will allow Anthropic to further refine and scale up Claude, its conversational AI assistant, as well as build even more advanced AI systems in the future. Amazon's cloud computing capabilities will aid Anthropic in training complex models, while Anthropic's work could enhance Amazon's consumer products and cloud offerings.
Meta Unveils New AI Features for Messaging and Creativity
Meta AI Assistant - The Social Assistant: Meta has unveiled an AI assistant for Messenger, WhatsApp and Instagram that can have more natural conversations, search the web for real-time information, and generate images from prompts. Powered by Meta's LLaMA language model and integrated with Bing search, the assistant aims to provide helpful information in a conversational way as a more human-like chatbot experience.
Make Me a GIF - AI-Generated Stickers: New AI-powered stickers coming soon will let users instantly generate unique animated stickers to express themselves more creatively in chats. Using Meta's new image generation model Emu, users can type text prompts to create fun, personalized stickers in seconds to better convey emotions and reactions in conversations on Messenger, WhatsApp and Instagram.
Restyle My Photo - AI Image Editing: Meta is bringing new AI-powered image editing capabilities to Instagram. With the "Restyle" feature, users will be able to completely transform the visual styles of their photos by typing prompts like "watercolor", "collage", "anime" and more. The "Backdrop" feature will let users change the entire background scene. These new creative tools aim to empower self-expression and unlock new possibilities for customize photos.
Celebrity Chatbots - Talk to Your Favorite Stars: Life-like AI chatbots based on celebrities and influencers like Mr Beast and Kendall Jenner are coming to Messenger. Designed for casual conversation about topics like sports, fashion and pop culture, these bots could make chats more engaging, fun and personalized. Users may enjoy chatting about their interests with AI versions of their favorite celebrities.
AI Studio - Build Your Own Chatbot: Meta's new AI Studio platform will empower businesses and creators to easily build customized chatbots for Messenger, WhatsApp and Instagram. Meta envisions AI assistants that reflect brands' personalities to enhance customer service, marketing, ecommerce and more. The accessible toolkit aims to make the power of AI available to everyone for creating unique chatbots and virtual assistants.
Adobe Launches Test Versions of Photoshop and Illustrator Online
Adobe released early online versions of its popular creative apps Photoshop and Illustrator. This is the first time core editing tools from these complex programs are available on the web. By putting them online, Adobe wants to make it easier for people to work together and use them. Adobe also announced new AI features like auto-masking in Photoshop and better neural filters.
ChatGPT Regains Real-Time Web Search Capabilities
ChatGPT is once again able to search the web to provide up-to-date answers, after this feature was temporarily disabled earlier this year. OpenAI announced on September 27th that it has re-enabled "Browse with Bing" which allows ChatGPT to cite current authoritative sources, instead of just generating conjectural responses. This real-time web search was previously available but then pulled by OpenAI to prevent misuse of paywalled content. Now with added protections, OpenAI is reintroducing this key functionality for paid subscribers, moving ChatGPT closer to being an AI assistant that can provide reliable, factual information drawing from live web data.
AWS Launches New Generative AI Services
Amazon Web Services (AWS) announced the general availability of Amazon Bedrock, a new fully managed service that provides access to leading foundation models for building generative AI applications. AWS also revealed new capabilities in Amazon CodeWhisperer and Amazon QuickSight.
That wraps up this week's AI news! If you found this informative, don't forget to share it with friends and colleagues. And be sure to subscribe to get next week's news straight to your inbox. Thanks for reading!