AI's battle (rap) begins

Rush for more data to define who wins the AI wars

Who could have predicted back when ChatGPT first went public in November 2022 that it would transform the entire tech industry so quickly?

The chatbot, made by tech startup OpenAI, used the GPT 3.5 language model to talk to correspondents in an eerily human-like way. It could answer questions, tell jokes, code with a frightening level of competency – and even write battle raps, much to this writer’s delight. And just like the humans it was mimicking, if it didn’t know the answer, it would bluff its way through and make up something plausible-sounding instead.

But it wasn’t all fun and games. It was also an earthquake through Silicon Valley as it became clear that this is the future. By January, Microsoft had invested $10bn into the company, and quickly built GPT functionality into its search engine, Bing, with the promise of direct integration into Office yet to come.

By March, Google had responded in kind with the launch of its own AI-powered chatbot named Bard.

At the same time, other companies launched their own “generative” AI tools, powered by the same sorts of large-language-models that made ChatGPT so powerful. Most notably, this included a small startup called Midjourney, which began generating eerily realistic AI images, and Adobe attempted to do the same with the launch of what it calls Firefly.

Little wonder that it has also been a boom time for Nvidia, which makes many of the chips that AI uses to crunch data – its stock price has more than doubled since January.

In short then, the AI war is on. But why is it taking off now? And just how much better can it get?