If 2023 was the year of AI, what might 2024 look like? Here are five AI predictions to track in the new year.
Apple has long had a penchant for showing up late to the party; iPods landed long after the first MP3 players emerged, the iPhone was years behind smart devices from Nokia and BlackBerry, and even the Apple Vision Pro - which is slated for an early 2024 release - comes more than a decade after the debut of the Oculus Rift DK1 headset.
But when it comes to AI, Apple is seemingly getting its skates on.
There have been reports that some within the company were caught off-guard by the quick rise of OpenAI, and so Apple is said to have a plan to catch up by launching a slew of AI features in 2024. These include an LLM-powered chatbot (RIP Siri?), AI-powered developer tools, and numerous generative AI features within iOS.
While this might all be speculative, I’d at least expect Apple to roll out an AI chatbot in beta - if only so the almost 50-year-old company can demonstrate it’s still down with the kids.
Right now, every person and their dog seems to be developing an LLM. That’s great for us as users, but with OpenAI’s GPT-5 projected to cost north of $2 billion, expect to see more consolidation than innovation in 2024.
It likely won’t be LLM-makers that look to consume their competition either; it’s easy to see a world where some text-to-image companies merge and AI music-makers blend into one.
Of course, not every AI startup can (or even should) survive, and so those that don’t get vacuumed could feel the chill of being on the outside. After all, investor money isn’t endless and AI models will only get more expensive. But those that can innovate in the space between the giants might one day find themselves calling the shots.
As many understand, cloud-based large language models (LLMs) are the brains behind the chatbots like Google’s Bard and OpenAI’s* *ChatGPT. But expect to see more focus on small language models (SLMs) that are geared towards mobile devices.
In some ways, the trend has already started, with Google pushing Gemini Nano to Pixel 8 smartphones, and Microsoft making its Phi-2 SLM available for mobile devices. In 2024, don’t be surprised if Meta, Amazon, and others get into the game with their own on-device AI models.
AI might have enjoyed its position as the Hot New Thing™ this year, but expect a lot of the good will to fade as 2024 settles in.
Thanks to countless think-pieces about machines taking jobs and kids outsourcing their homework to ChatGPT, many lawmakers are ready to tap into the anxiety. Indeed, the European Union has already fired the starting gun by agreeing a set of proposals designed to reign in the AI industry.
Nonetheless, some are prepping the fight back: investment heavyweight Andreessen Horowitz has said it will give financial support to US politicians who back “an optimistic technology-enabled future.”
So in short, expect the battle over AI regulation to get serious next year.
It feels like AI only really became a thing yesterday, but so many are expecting artificial general intelligence (AGI) to be just around the corner. And while expert opinions diverge on when true AGI will be here (some anticipate roughly five years, others say it will be much longer), when it does land, it could change everything.
Still, the first step to the future begins with the next one, and in that sense, expect to see the initial signs of AGI as soon as 2024.
OpenAI’s CEO Sam Altman has talked at length about his AGI ambitions (some have speculated it’s why he was temporarily ousted), and there has been a lot of buzz/hype about what the company is up to, including debunked rumorsof a GPT-4.5 release featuring “complex reasoning and cross-model understanding.”
Despite the off-base speculation, don’t be surprised if _someone _does make a splash in 2024 with a claim that their LLM really is showing signs of thinking for itself. After all, whoever gets to true AGI first will surely capture the minds (and wallets) of many.