According to a leaked internal Google memo written by a senior Google employee, OpenAI and Google’s AI technologies may not come out on top in the AI arms race. The document has been in circulation in Silicon Valley for several months and was recently made public by consulting firm SemiAnalysis.
In recent years, OpenAI and Google have been competing to come out on top in the AI industry. OpenAI became famous last year when its conversational AI chatbot ChatGPT became a global topic. Google, on the other hand, has been working in the AI field for more than 10 years and was seen as a leader in the industry. However, recently it has been reported that it is trying to catch up with OpenAI.
Despite the intense competition, Google engineer Luke Sernau believes that neither company will ultimately come out on top if they continue down this path. Sernau released an internal document in April to promote the idea.
According to Sernau, while Google and OpenAI are struggling with the impact of regulations, it is open-source AI that is leading the way. Large-scale language models running on smartphones and personal AI that can be fine-tuned overnight on laptops are examples of how the technology has evolved.
Sernau writes that AI models developed by private companies still have the upper hand, but they won’t last long. The open-source model is getting to the point where you can achieve the same results for $100 as companies spending billions of dollars can create new iterations in months; open-source technology has done the same in weeks.
Sernau attributes the huge models used by Google and OpenAI as a major reason for slowing their progress. The open-source community came across LLaMA brought from Meta and saw potential in it. Sernau emphasized the need for Google to shift to a smaller model and learn from the open-source community.
The leaked internal document raises questions about the future of the AI industry and whether open-source technology will ultimately lead the way. As better AI models become available for free, clients may not be willing to pay Google or OpenAI to use their inferior models. The race to dominate the AI industry may not be won by the biggest player but by the most innovative.