Alphabet and Google CEO Sundar Pichai in Warsaw, Poland, March 29, 2022. — NurPhoto/Getty Images
Warren Buffett’s flock heads to the Midwest this weekend to attend Berkshire Hathaway’s 59th annual shareholder meeting. In the recent turmoil of the banking industry, investors will look to Oracle of Omaha and his witty right-hand man Charlie Munger for reassurance. Even today, the impact of his under-90 pairs on investor confidence and the market for ideas in economics is evident.
In an internal document released Thursday by the consulting firm Semi-Analysis, senior Google engineer Luke Sernau referred to one of Buffett’s most famous theories, the “economic moat,” and explained it. clarified. And neither does OpenAI. In a document published inside his Google in early April, Sernau said the company is losing artificial intelligence dominance not to his flashy Microsoft-backed OpenAI, but to ChatGPT. It claimed to have been a huge hit since its release last November. A large-scale language model published in February, he has been working on open source platforms such as his LLaMa on Meta.
“We have looked into a lot about OpenAI…but the uncomfortable truth is that we are not in a position to win this arms race and neither is OpenAI. While we were there, a third faction quietly ate our lunch,” he wrote. “Obviously, I’m talking about open source. Quite frankly, they’re cornering us.”
selnau did not respond luckrequest for comment, as well as Alphabet, Google’s parent company.
The engineer’s argument ties in with Buffett’s idea that, like a castle in the Middle Ages, a successful business must guard its profitable cattle with a moat that prevents competitors from making credible challenges. These moats can come in the form of size advantages that help reduce costs, patents that protect key innovations, high entry costs for competitors, and even brand recognition.Billionaire detailed his economic trench theory in 1999 luck article:
“The key to investing is not in assessing how much an industry will impact society or grow, but in assessing the competitive advantage of a particular company and, above all, the sustainability of that advantage. It’s a matter of judgement: products and services that have a wide and sustainable moat around them will reward investors.”
For decades, Google’s search engine dominance, brand recognition, and technological superiority have served as a buffet moat, allowing the company to stay ahead of its competitors. But Sernau argued in his internal letter that the trench is drying up as the artificial intelligence arms race heats up.
He explained that within a month of Meta’s LLaMa leak, they had already developed a sophisticated model that allowed independent researchers to compose emails and post them on social media.
“Barriers to entry for training and experimentation have been lowered from the gross output of major research organizations to one person, one night, and a rugged laptop,” he writes. “There is no secret sauce.”
Google introduced ChatGPT rival Bard in February, but the chatbot has been plagued by errors that AI researchers have dubbed “hallucinations” since its release. The company said last month that it plans to add Bard to its search engine to compete with Microsoft’s Bing, which currently uses ChatGPT, but so far nothing has happened.
However, in his letter, Sernau argued that Google shouldn’t focus so much on competing with ChatGPT. Instead, we need to open up the AI platform to the public and allow anyone to create their own applications with it to prevent the competition from gaining an edge.
“The more tightly controlled the model, the more attractive and open options we can create,” he wrote. “Our best hope is to learn from and collaborate with what others are doing outside of Google.”
Finally, the engineer argued that consumers would only pay for closed models such as ChatGPT and Google’s Bard for free, as open-source alternatives are now “equivalent in quality.”
“We need to consider where the added value really lies,” he wrote.
