Google has quickly narrowed the gap with OpenAI in the AI artificial intelligence race, partly because OpenAI no longer maintains its leading advantage after launching ChatGPT.
This assessment was made by Jerry Tworek, former Vice President in charge of research at OpenAI, who just left the company in early January.
According to Mr. Jerry Tworek, OpenAI once held a clear dominant position after ChatGPT created a major turning point for AI towards consumers. However, that advantage has not been fully exploited.
If you lead and have all the advantages OpenAI once had, then you need to maintain that position," said Mr. Tworek.
Tworek joined OpenAI in 2019, when the company had only about 30 employees. He is considered one of the important architects behind major breakthroughs in advanced reasoning models.
Although not specifying any specific mistakes, he believes that OpenAI wasted the outstanding gap they created in the early stages of ChatGPT.
Meanwhile, Google was once assessed as reacting slowly to the AI-generating wave but has strongly changed its strategy.
According to Tworek, Google began to consider AI as a top priority, investing in training a series of large-scale language models. As a result, by last year, Google not only caught up with OpenAI in terms of the number of models, but according to some assessments, was also superior in capabilities in some aspects.
Tworek's statements were made in the context that OpenAI is facing increasing competitive pressure, while facing stricter supervision of privacy and user data usage.
According to him, the current competitive environment has fundamentally changed the way companies conduct AI research.
The pressure to both increase user growth and pay for huge computing infrastructure, while racing to create the best model, makes companies increasingly cautious.
This reduces the level of willingness to take risks for bold research directions, which may not yield immediate results.
Tworek believes that general artificial intelligence (AGI) will require revolutionary methods, instead of just expanding existing models.
He predicts AGI may appear around 2029, if the industry is willing to accept new approaches.
This assessment is consistent with the overall picture of the industry. At the World Economic Forum in Davos (Switzerland), Microsoft CEO Satya Nadella described competition in AI as "quite fierce" and said that its impact will soon be clearly reflected in global GDP growth.
Meanwhile, DeepMind CEO Demis Hassabis called this the "fiercest competitive environment" the technology world has ever witnessed.
The AI race is entering a new phase, where there is no room for complacency, even for those who used to lead.
