Former OpenAI VP Says Google Caught Up as ChatGPT Momentum Waned

Jerry Tworek, who recently stepped down as OpenAI’s Vice President of Research, has expressed concerns that the company failed to fully capitalize on its early lead after the launch of ChatGPT. Tworek said this misstep allowed Google to catch up quickly in the competitive field of Artificial Intelligence, closing what was once a significant capability gap between the two tech giants.

According to Tworek, OpenAI’s early dominance in consumer-facing AI was clear after ChatGPT transformed the market, but that momentum waned as other companies, particularly Google, ramped up their AI research programs. Google, he noted, “started treating it seriously at that moment, training large language models, and, like, through OpenAI fumbling its lead, they are very, very close now in capability and in terms of models trained.”

Tworek joined OpenAI in 2019 when the organization had only around 30 employees. He played a key role in designing several major innovations, including the sophisticated o-series reasoning models. While he did not specify a single misstep by OpenAI, he asserted that the company did not fully leverage the advantage it gained after launching ChatGPT.

He emphasized the importance of staying ahead in forward-thinking businesses. “If you are the owner of a forward-thinking business with all of OpenAI’s advantages, you should always be ahead. However, Tworek revealed that Google went on to accomplish a great deal of success,” he explained. “Very clearly, Google started treating it seriously, training large language models, and, through OpenAI fumbling its lead, they are very, very close now in capability and in terms of models trained.”

Tworek’s observations come at a time when OpenAI faces intense competition and scrutiny regarding user privacy and research priorities. He explained that the pressures of scaling user growth, covering computational costs, and competing for top AI models have constrained companies’ willingness to pursue riskier, long-term research that may not produce immediate results. Tworek believes that reaching artificial general intelligence will require radical, innovative approaches and predicts it could be achieved by 2029.

The ongoing AI race has been highlighted by other leading executives at Davos this week. Microsoft CEO Satya Nadella described the competition in AI as “pretty intense,” expressing both caution and gratitude for the progress made so far. Nadella also suggested that in five years, AI could be considered a factor proportional to global GDP, reflecting its growing economic significance.

Similarly, DeepMind CEO Demis Hassabis characterized the AI landscape as the “most intense environment” that veteran technology professionals have ever witnessed, emphasizing the rapid pace of innovation and the unprecedented level of competition across leading AI research organizations.

Tworek also shared his personal reflections on AI development, research Culture, and the responsibility that tech companies have in shaping the future of artificial intelligence. He highlighted the tension between short-term product milestones and the need for groundbreaking, long-term breakthroughs that can bring society closer to general AI capabilities.

His remarks underline the high-stakes environment of AI research today, where market dominance can be fleeting and small strategic missteps may allow competitors to catch up. The former OpenAI executive’s insights suggest that continuous innovation, careful planning, and bold research decisions are crucial for maintaining leadership in the rapidly evolving field of artificial intelligence.