Bobby Kotick's AI comments perfectly sum up the thoughtlessness behind the large-language model gold rush
One of the big problems with the tech industry's current obsession over large language models(opens in new tab) (commonly referred to under the umbrella term «AI»), is separating out the genuine potential applications and implications of this technology, from the marketing hype, misunderstanding, and outright lies proffered by industry leaders and blue-tick grifters pivoting from NFTs. A good example of how the two are easily conflated was recently provided by Activision CEO Bobby Kotick, who shared his thoughts on AI's potential applications in gaming.
As reported by Kotaku, Kotick was asked his opinion of AI tech in a company-wide meeting last week. «I don't know how much people realize that a lot of modern-day AI including ChatGPT started with the idea of beating a game, whether it was Warcraft or Dota or Go or Chess. But what is now these large language learning model AI technologies, all started from this idea of beating a game.»
Presumably, Kotick is referring to specialised supercomputers like IBM's Deep Blue(opens in new tab), as solving games and game-like puzzles were an early target for developing artificial intelligence. It's a perfectly reasonable comment, but it isn't especially relevant to what lies ahead for modern LLMs.
Kotick then expresses his belief that LLMs will be as influential as the «first MacIntosh» in terms of «how meaningful the impact of AI would be on society both positive and negative.» He then adds, «For what we do, I think it will have a profound positive impact on the things we'll be able to do in game development for a long time. It will enable us to do things that we haven't been able to do for a long time.»
Fair enough. But what things, exactly? Kotick uses Guitar Hero as an example.
Read more on pcgamer.com
