OpenAI Could Lose Its AI Lead to Google For the First Time in 4 Years

The AI race is now at a tipping point

It was in 2019 that OpenAI released GPT-2, surpassing Google in the race to create better generative AI models.

Will Google retake the AI throne with Gemini before 2023 ends?

The Information scooped yesterday that in mid-2023 OpenAI had to stop working on a new model, codenamed Arrakis, that would presumably make ChatGPT run more efficiently.

This was the company’s main ongoing development after finishing GPT-4 in the summer of 2022.

Why did they stop? Because the model didn’t work as expected. In a space where a gap of a few months is the difference between being the leader or not, this was an important setback for OpenAI.

The young startup is more than fine — this isn’t a life-or-death situation. It’s making $1.3 billion in ARR, has been releasing new juicy models like GPT-4 vision, DALL-E 3, etc., and has more aces up the sleeve for the DevDay conference on November 6th.

But this “failure” (I guess we can say that if we compare it with the flawless 4-year run it’s had so far!) may allow Google to surpass OpenAI with Gemini, which is posited to beat GPT-4.