GPT-4 Will Be Released Next Week, Microsoft CTO Says

Is This The Dawn Of AGI Or Just Another Hype?

And then it happened.

There have been many rumors about GPT-4 and the fact that Microsoft Germany CTO Dr. Andreas Braun almost casually announced the release and hinted at some incredible new features is as confusing as it is exciting!

Multimodality: A New Chapter In The History Of Artificial Intelligence?

At a one-hour information event entitled “AI in Focus — Digital Kickoff” on 9 March 2023, Dr. Andreas Braun said:

“We will introduce GPT-4 next week, there we will have multimodal models that will offer completely different possibilities — for example videos”

Multimodal capabilities, that’s where the most exciting developments in AI are happening right now:

Until now, we have been using models such as DALL-E, Midjourney and Stable Diffusion for image generation separately from the AI text generators (GPT-3, ChatGPT, and the likes), the multimodality paradigm aims to combine several of these modalities into single models.

We are talking about AI models that can understand text, images, and sound at the same time and use them to get knowledge of the world around them as well as interact with it.

Kosmos-1

Earlier this month Microsoft already presented Kosmos-1, a “multimodal large language model” (MLLM), showcasing the sheer power of this new pathway of AI development.

Have a look:

Kosmos-1 uses text AND image information to answer questions, giving it unprecedented “cognitive capacity” and even the ability to solve visual puzzles:

Let that sink in.

For more info, here’s the link to original Kosmos-1 paper: https://arxiv.org/abs/2302.14045

What will GPT-4 be like?

Scientists even argue that multimodality can be a path to developing AGI, or “artificial general intelligence”, an AI that could theoretically replace humans in any intellectual task.

While Kosmos-1, the impressive showcase of multimodality, is a Microsoft project, AGI is the actual stated goal of OpenAI.

Now, this is what I call a fit.

However, besides the multimodality paradigm, we do not know for sure what GPT-4 will actually be like.

There are rumors, of course.

Some people talk about a 100-fold increase in parameters (GPT-3 currently already uses 175 billion parameters), but Dr. Braun did not say anything about it. I’m sure there are other people besides him who know a lot about what GPT-4 will be like because they’ve had the opportunity to test early versions of it.

For the rest of us: One more week!

See you on the other side. :)

➡️ For more information about AI & Creativity, follow me on Twitter or Medium (use my referral link to get full access to all my articles and those of thousands of other writers).

➡️ If you like my content, why not leave a “clap” at the end of this article, so more people can see it?