Most interesting... so I wonder how long it will be before AI learns that there is a connection between text and images portraying text? And language is so subtle and nuanced that it might be a steep learning curve for AI to make the connection. Any text (such as the faux 'headlines' depicted in the above image) would have to be pithy, very pointed, and have memetic value.
If I were to insert text in the image above, ti would read "DEMS LOSE BIG!" but for AI to synthesize all it knows about language into those 3 words and then incorporate them into a graphical representation, seems like it would take a fairly large intuitive leap of understanding.
AI might learn that the 'headline' means something, but to come up, unprompted, with a pithy phrase that is both true AND funny seems like a very large cognitive leap.
Most interesting... so I wonder how long it will be before AI learns that there is a connection between text and images portraying text? And language is so subtle and nuanced that it might be a steep learning curve for AI to make the connection. Any text (such as the faux 'headlines' depicted in the above image) would have to be pithy, very pointed, and have memetic value.
If I were to insert text in the image above, ti would read "DEMS LOSE BIG!" but for AI to synthesize all it knows about language into those 3 words and then incorporate them into a graphical representation, seems like it would take a fairly large intuitive leap of understanding.
AI might learn that the 'headline' means something, but to come up, unprompted, with a pithy phrase that is both true AND funny seems like a very large cognitive leap.