Users celebrate Gary Marcus's vindication on symbolic and neurosymbolic AI as overdue recognition after years of dismissal by the field, while a few criticize past hostility toward those ideas.
Based on 6 visible X reactions from 8 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@GaryMarcus @ylecun @geoffreyhinton Neurosymbolic AI is exactly what we have been building toward. It is funny seeing the old guard finally come around to neurosymbolic approaches.
@LH @theo @GaryMarcus Totally agree. The plumbing, retrieval systems, and guardrails make or break the actual product way more than the base model choice.
@GaryMarcus @ylecun @geoffreyhinton Worst part is not that they ignored him but that they were hostile to people pointing the right way
@GaryMarcus @ylecun @geoffreyhinton 20 years of saying the same thing and finally getting the "i told you so" moment
@GaryMarcus @ylecun @geoffreyhinton symbolic reasoning won just under a different name - happens every time in science
The field of machine learning owes me an apology, starting with @YLeCun and @geoffreyhinton, each of whom dumped on me for many, many years for saying we need symbols in addition to deep learning. Because that’s exactly what loop engineering is: adding symbols to deep learning. It’s just shocking how correct I was, and how unable these folks are to acknowledge it.
For almost two decades people like @YLeCun and @geoffreyhinton dumped on me for saying we need symbols in addition to deep learning. But that’s exactly what loop engineering is: adding symbols to deep learning. Astonishing how wrong some people’s heroes can be, and hostile they were to people trying to point to a better path.
Users celebrate Gary Marcus's vindication on symbolic and neurosymbolic AI as overdue recognition after years of dismissal by the field, while a few criticize past hostility toward those ideas.
Based on 6 visible X reactions from 8 accounts; directional sample.
Ask a question below.
Published answers will appear here.
The field of machine learning owes me an apology, starting with @YLeCun and @geoffreyhinton, each of whom dumped on me for many, many years for saying we need symbols in addition to deep learning. Because that’s exactly what loop engineering is: adding symbols to deep learning. It’s just shocking how correct I was, and how unable these folks are to acknowledge it.