AI is a con
AI is a grift
AI is a scam
"The people who have temporarily gotten rich or famous on AI have done so by pretending that this outlier problem simply doesn’t exist, or that a remedy for it is imminent. When the bubble deflation that I have been predicting comes, as now seems imminent, it will come because so many people have begun to recognize that GenAI can’t live up to expectations."
https://garymarcus.substack.com/p/this-one-important-fact-about-current
@gerrymcgovern A question:
How is this different from how human beings approach solving outlier problems? Are we not subject to the same failure modes?
In other words - is this an issue inherent in the systems themselves or in our *expectations* and assumptions of the systems? What does a system that does not have these failure modes even look like, given the infinitely broad problem space we expect them to operate in.
@dogzilla We have a lived experience, although, true, some of us don't have much common sense. However, if we see a huge thing sprawled across the motorway, most of us will slow down. AI will keep going and crash into it because it's not in its training set. Not all AI will do that all the time, but AI has very little common sense.
@gerrymcgovern @dogzilla Basically, the machine models need to “understand” their data. The only way for that to work is to establish a symbol composition scheme for nodal datasets and then a correlation regime for the symbols. This might be enough to manage the edge/corner case situations and allow the models to extrapolate and adapt.
