Alphabet (NASDAQ:GOOGL), the parent company of Google, surged 4.6% yesterday after Bloomberg reported that Apple (NASDAQ:AAPL) may integrate Google Gemini’s AI inside the iPhone. Indeed, a hypothetical Apple-Google Gemini agreement in artificial intelligence (AI) may be similar to their search deal, which has faced significant regulatory scrutiny in recent years.
And, while a Google-Apple AI alliance is sure to irritate antitrust regulators, I see it as a potentially significant win for both businesses as they attempt to gain (even more) ground in the AI race.
With Microsoft (NASDAQ:MSFT) leading the charge on generative AI through its OpenAI investment, Apple must find a means to make up for lost time in the AI space. Collaboration with Google offers an opportunity to catch up with industry heavyweights such as Microsoft. It could also be a simple method to get an advantage now that Apple has access to one of the most powerful large language models (LLMs), Google Gemini. Given this, I am still bullish on both AAPL and GOOGL.
While I’m sure Apple could build its own LLM into its devices, I believe Google is the superior option at this early stage of the AI growth. Why?
Even while LLMs have improved in recent years, they are remain susceptible to hallucinations.
Only time will tell whether we have a strong, hallucination-free LLM ready for market. Regardless, considering Apple’s history of selling polished goods that are actually ready for flight, I don’t see Apple launching a work-in-progress LLM product.
Even the most recent versions of Google Gemini and ChatGPT-4 Turbo, as competent and remarkable as they are, are still technically under development, in my opinion.
Neither LLM is ideal, especially at this stage in the AI lifecycle when hallucinations continue to occur on a daily basis. My estimate is that it will take some time until LLMs are polished enough to make mistakes extremely unlikely (or perhaps virtually impossible).
Perhaps we’re still a few quarters, if not years, away from the error-free AI that many of us see as the holy grail of AI. Given how quickly generative AI technologies are growing (at exponential rather than linear rates), hallucination-free LLMs may be far closer than many of us realize.
For the time being, I believe most people have accepted that the occasional hiccup is unavoidable while using even the best LLM available. Users of such LLMs can simply verify the sources to ensure they are not experiencing hallucinations. Indeed, fact-checking can be difficult when utilizing LLMs such as Gemini or GPT-4, but it’s a minor trade-off. Furthermore, it is a choice between that and utilizing a traditional search engine.
Read More: https://thesiliconleaders.com
