Yesterday, OpenAI introduced a new AI model, the GPT-4o (“o” as in “omni”), which the company claims will be more user-friendly than the previous GPT-4, released publicly 14 months ago.
GPT-4o will communicate verbally in real time without delays, allowing users to interrupt mid-conversation. Both characteristics have been challenging to achieve, as the company aims to resemble “realistic” conversations as much as possible. During the model presentation, GPT-4o verbally assisted a researcher in solving an equation on paper and provided real-time verbal translations.
As stated by OpenAI, the model’s new capabilities will include improved and faster responses in languages other than English and will retain the memory of its conversations. GPT-4o will be released for free in the coming weeks and will be ad-free, with a subscription option offering additional features.
Rumors suggested that OpenAI would unveil a new AI search engine to rival Google, but the company has postponed the announcement.
This represents OpenAI’s latest strategic move, backed by Microsoft, to maintain its edge in the competitive AI landscape. GPT-4o’s debut comes just ahead of the expected unveiling of Google and Apple’s AI model updates in a few days.
While OpenAI advances its models, it’s concurrently embroiled in legal battles with media outlets like the New York Times. These outlets accuse the company of training its AI models on their content without consent, seeking billions of dollars in damages.
What is fyi.news?
