There is no denying the huge impact and success that OpenAI has seen since the public release of its ChatGPT feature in December last year. While OpenAI remains the most dominant player in the market, a number of challengers are rapidly closing the gap and starting to shift market share.
Despite the $10b Microsoft invested into OpenAI, many of these competitors will simply be able to invest far more into this space to produce similar products. Google, not wanting to be left behind Microsoft in this space, released Bard, their ChatGPT competitor. Google’s ambition is to bring Bard into all of its products in the future, similar to how Microsoft is approaching OpenAI and ChatGPT.
Google is now so confident in the writing capability of its AI systems, that it recently announced a new solution that can write news articles. Codenamed ‘Genesis’, the system is designed to take in information from journalists, and then generate news copy. Google, which sees this tool as “responsible technology” is built with the ambition of being a journalist’s personal assistant, automating some tasks to free up the journalist to focus on other things. Google has reportedly already pitched this to The New York Times, The Washington Post, and News Corp, with feedback currently mixed due to ethical concerns.
Another innovator in the GPT space has been Anthropic, which released its rival to ChatGPT, Claude, in March this year. While initial reviews revealed nothing groundbreaking, combined with Claude currently only being available in the US and UK, Claude has not received a lot of coverage. However, this did not stop Google from investing $300m into Anthropic, seeing encouraging signs in the work Anthropic was doing. Now, Anthropic released Claude 2 earlier this month. While the same limitations are in place for access, overall the feedback on Claude 2 has been encouraging, looking to be close to truly threatening ChatGPT.
Surprisingly active in the AI space has been Meta, who have revealed a number of exciting innovations in this space. Unlike OpenAI and Google, Meta has looked to decentralise their AI learnings, making most of their AI solutions open source. This move allows people to download, train, and modify Meta’s code to iterate and evolve. One of these solutions is LLaMa, with Meta last week releasing LLaMa 2. While Meta has been the first to admit that LLaMa 2 is not yet ready to compete with GPT-4, early feedback has been encouraging that it’s getting close.
The real threat in this space, though, is only just starting to emerge. Apple, the titan of tech with its astounding $3 trillion valuation, has teased its entry into this AI realm. Little is known currently about what is reportedly referred to internally as ‘Apple GPT’, and codenamed AJAX. However, the thought of a company the size of Apple entering this space is making a lot of headlines.
Some have speculated a fusion between cornerstone Apple services, such as iMessage, Keynote, Pages, and Health, with Apple GPT. AI strength hinges on the data it feeds on, and Apple sits on an enormous amount of data on its highly loyal users. If Apple is able to connect all of its data signals together with its AI, while ensuring user privacy is consistently maintained, it could bring about a new evolution of AI. If successful, Apple could morph all of this into an unparalleled AI sidekick, adeptly grasping user rhythms, growing with them, and possibly evolving into a tailor-made life navigator, always guarding user well-being.
ChatGPT reportedly saw its first drop in user numbers engagement last month, down 9.7% in June compared to May. This could be driven by the increased competition in this space. OpenAI’s biggest investor, Microsoft, has productised their chat feature in Bing, powered by ChatGPT, while Google and Anthropic continue to iterate their AI offering. These continued improvements are rapidly closing the gap in competition, but two new threats are emerging, one with deep pockets, and one with very deep pockets.