Open AI Building New AI Device, Amazon Invests in Anthropic, ChatGPT Can Hear, See, & Speak, New Open Source Model Released, and Meta's Chatbot Invasion
OpenAI is working with Jony Ive, Amazon answers Microsoft's OpenAI partnership, ChatGPT has new senses, a new open source model is out, and Meta released celebrity chatbots
OpenAI Building New AI Device with Apple’s Former Chief Design Officer
OpenAI is reportedly in talks with former Apple designer Jony Ive & SoftBank’s Masayoshi Son to kickstart a venture aimed at crafting the “iPhone of artificial intelligence” with SoftBank pledging over $1 billion.
Their goal is creating a more natural user interface using AI, aiming for a new paradigm shift, as the iPhone’s touchscreen was in 2007.
While the conversations are early, SoftBank’s financial commitment, coupled with Arm's technological prowess, shows that this is more than just brainstorming.
What could this device look like? Possibly more like the movie Her, and less like an iPhone 📱.
Amazon Invests $4 billion in Anthropic & Announces Strategic Partnership
Amazon is investing up to $4 billion in AI startup Anthropic. They’re initially investing $1.25 billion, with an option to increase the investment to $4 billion.
Anthropic is known for their AI-powered chatbot Claude, as well as for developing their “frontier model” (tentatively called “Claude-Next”), which would be 10x more capable than the most powerful AI currently available.
Amazon is essentially responding to Microsoft’s partnership with OpenAI, by providing Anthropic with funding and compute resources through AWS. Similar to how Microsoft is helping OpenAI with compute resources through Azure.
Given that, we’ll likely see Anthropic services through AWS, similar to what Azure is offering with OpenAI. Amazon already announced that “Anthropic will provide AWS customers with early access to unique features for model customization and fine-tuning capabilities”.
OpenAI’s ChatGPT Can Now See, Hear, & Speak
OpenAI is rolling out voice and image capabilities in ChatGPT, offering a more intuitive user experience.
Their demo video shows what this is capable of, but I believe the reality shown by Mckay Wrigley is even more impressive.
Peter Yang’s photo also shows just how powerful this new version of ChatGPT is.
And the crazy thing is, this update still hasn’t been fully rolled out to ChatGPT’s users. Given that, we’ll probably see even crazier use cases over the next several weeks.
Mistral AI Launches Their Open Source Generative AI Model
French AI startup Mistral, who raised a massive $113 million seed round in June, released their inaugural language model for free public use under the highly permissive Apache 2.0 license.
Mistral 7B, which can be downloaded here, offers comparable capabilities to existing models at a lower computational cost.
Their performance is quite impressive, considering how much more restrictive Llama 2 is. Given how restrictive Llama 2 is, and that OpenAI has not been “open”, Mistral’s release is a win for the open source community.
Meta's Chatbots Leap into WhatsApp, Messenger, and Instagram
Meta is diving headfirst into AI chatbots, launching their own assistant along with celebrities like MrBeast and Dwayne Wade, across WhatsApp, Instagram, and Messenger.
Their chatbot can generate high resolution images on command, and can execute searches thanks to a partnership with Microsoft’s Bing.
The core of Meta’s innovation is a custom-designed large language model, drawing inspiration from Llama 2. With billions of daily users across their platforms, Meta's massive distribution network might just catapult their AI assistant to become a commonplace chat companion, bringing more competition against OpenAI's ChatGPT.