Communications of the ACM
Supersized AIs: Are Truly Intelligent Machines Just a Matter of Scale?
Few experts thought machine intelligence would arrive as a mere exercise in engineering. Of course, many still doubt that it will. Time will tell.
Pablo Hurtado de Mendoza
While many believe that AIs will never truly achieve general intelligence, there are reasons to believe that AIs such as GPT-3 may soon develop human-level language abilities, reasoning, and other hallmarks of what we think of as intelligence. GPT-3's recent successes are being attributed to one thing: it is bigger than any AI of its type, meaning, roughly speaking, that it boasts many more artificial neurons.
No one expected that a shift in scale would make such a difference. But as AIs grow ever larger, they are not only proving themselves the match of humans at all manner of tasks, they are also demonstrating the ability to take on challenges they have never seen. Some in the field are beginning to think the inexorable drive to greater scales will lead to AIs with abilities comparable with those of humans. That would be huge if true.
From New Scientist
View Full Article (May Require Paid Registration)
No entries found