OpenAI Unveils O3 Model, But Access Remains Restricted
Welcome to Week in Review. Today, we are discussing OpenAI’s final and perhaps greatest announcement during the 12 Days of OpenAI’, Apple’s possible dabble into the foldable market; and why Databricks is opting to wait to go public. Let’s dig in.
OpenAI CEO Sam Altman introduced the successors to its o1 reasoning model family: o3 and o3-mini. The models are not currently publicly available, but safety researchers can apply for early access. This revelation culminates the company’s “12 Days of OpenAI” event, which had included the introductions of real-time vision capabilities, ChatGPT Search, and even a Santa voice for ChatGPT. You can catch up on everything you have missed right here.
Big upgrade this week or Ray-Ban and Meta smart glasses. The newest feature of busines darlings early access program is now allow members to download firmware-v11, which incorporates live AI. It reports that users can continuously converse with Meta AI, thereby making references to things mentioned earlier in the conversation-works with real-time video.
It turns out that UnitedHealth’s Optum left an AI chatbot exposed to the Internet – for anyone to access with a web browser-that employees would use to ask questions about claims.
Although the chatbot did not seem to purport or contain sensitive personal or protected health information, its exposure comes at the most unfortunate of times when parent company UnitedHealth is already under fire for using AI tools and algorithms to bypass doctors’ medical decisions and deny patient claims.
Join Our Social Media Channels:
WhatsApp: NaijaEyes
Facebook: NaijaEyes
Twitter: NaijaEyes
Instagram: NaijaEyes
TikTok: NaijaEyes