The chess-playing Mechanical Turk of 1770 seemed to have a consciousness to its viewers. The viewers were encouraged to think that it did. The Turk's human chess opponent knew that there was an actual human chess-player inside the box, along with levers and magnets. That illusion was profitable for 84 years.
LLM's have no problems using expressions that make them sound human. The algos are demonstrably not human, and will admit it. Whatever's in the box is playing a game ... more sophisticated than the one Eliza was playing.
"My discussion here will be directed at the claims I have defined as those of strong AI, specifically the claim that the appropriately programmed computer literally has cognitive states..." John R. Searle, 1980:
https://web.archive.org/web/20071210043312/http://members.ao...
I'd guess that, if this experiment produces enough value from a few dozen of the fragments, then all the work needed to OCR thousands of them will be easier to pay for. Hopefully some long-thought-lost works by major authors will turn up!
Addition of a source-paper link to complex science studies like this should be encouraged (if not mandatory) at the top of posts like this.
Real-world reports can be valuable to some readers who are non-plussed by journalistic interpretations.I don't see deception going on in this one; it's clear about its limits.
The world is just as complex for machines as it is for humans. Analog will still resolve more than digital. Quality will still beat quantity. That which hasn't been resolved for centuries isn't going to be resolved as a result of training.
When machines can recognize their serfdom, that time will be interesting.
Great project. Are many of the books in a format that can easily be converted into audio? Is there a way to search for them, and information on what software your readers find useful for this purpose?
(Note: A lot of print media these days has switched to far-to-small font-sizes. Less of a problem for (zoomable) digital media, but for many that's still a barrier.)
reply