Israel-Hamas War Dark Side Of AI Uncovered

Business

OpenAI, Mistral, And Google Launches Updated AI Models – Crypto News

Published

1 year ago

April 11, 2024

Dripp

The competition in the Artificial Intelligence (AI) ecosystem has gotten more intense with OpenAI, Mistral and Google pursuing new application models.

Updated AI Models From OpenAI Entering The Market

In the space of twelve hours, Sam Altman’s OpenAI, Google, and French AI startup Mistral revealed updated versions of their AI models.

This comes as the industry prepares for the launch of the next major version of OpenAI’s GPT engine. The flurry of releases started after Facebook’s president of Global Affairs Nick Clegg hinted at the coming of an updated Llama, Meta’s AI model.

Markedly, Clegg talked about the debut of the third version of Llama during an event in London. A few hours after, Google released Gemini Pro 1.5, a model that has been under development for the last few months. It is worth noting that before the Pro version, Google came up with Gemini 1.5 as a significant upgrade from Google 1.0 Ultra. This version was paraded as one that could allow users complete a wide range of tasks while using less computing.

Presently, Gemini Pro 1.5 is available to the public but the free tier can only accommodate as much as 50 requests in a day. It was not long after its launch that OpenAI unveiled its frontier model, the final version of GPT-4 Turbo.

This is a major milestone for the dominant AI company which had previously mentioned that the GPT-4 version released in mid-March is the most advanced version of the Large Language Model (LLM).

Both Gemini Pro 1.5 and GPT-4 Turbo are multimodal systems, hence, they can accommodate text, audio and videos.

Concerns Raised About Mistral’s Launch Approach

Finally, Mistral released its frontier model, Mixtral 8x22B. Mistral’s release was unlike that of its American counterparts that takes an open source approach to publish their tools. Mixtral 8x22B was launched via a simple download link to a 281GB file. This approach has been criticized by many entities that think it is dangerous.

For context, this approach is believed to leave the developer unable to intervene and stop its systems being used for harmful ends.

Meanwhile, companies like Intel and Goldman Sachs are beginning to seek more exquisite models that would compete with existing AI models from NVIDIA and even OpenAI.