Researchers got a better look at the thoughts of chatbots, amateurs learned exactly how complicated simple systems can be, and quantum computers passed an essential milestone.
The end of 2024 seems a particularly uncertain time in history, and theoretical computer science is no exception. Amid several breakthroughs and new findings, the field also confronted its own doubts and limitations.For example, artificial intelligence once again dominated the popular discourse this year. Researchers have begun to understand what might be going on within the “black boxes” of neural networks that power chatbots such as Bard and ChatGPT, and to show that these systems truly understand the data they’re manipulating and composing. But there’s also a growing sense that AI’s progress has already started to slow(opens a new tab).Other areas of computer science saw clearer successes. After decades of hiding, an elusive numerical critter known as the fifth busy beaver finally gave itself up. But even here the news is not all good: Initial searches for its successor — the sixth busy beaver — suggest that it may lie beyond impassable mathematical barriers.The field of error-correcting codes — mathematical constructions that can fix themselves if an error has occurred — also rose in prominence in 2024. Researchers showed, for the first time, that an error-correcting technique essential for quantum computers really does work. But another finding showed that certain classical error-correcting codes are fundamentally inefficient, dashing hopes for practical versions of this “magical phenomenon.”