Discover and read the best of Twitter Threads about #badalgorithm
Most recents (1)
More explicitly #cherrypicked #GPT3 output presented as somehow 'meaningful' text.
*sigh*
HINT: the #AI wasn't "advancing" any sort of argument, in any of the 8 runs @guardian editors chose to mix'n'match.
Read Searle & the #ChineseRoom!
theguardian.com/commentisfree/…
*sigh*
HINT: the #AI wasn't "advancing" any sort of argument, in any of the 8 runs @guardian editors chose to mix'n'match.
Read Searle & the #ChineseRoom!
theguardian.com/commentisfree/…
And then there's this:
Which you might hope a good (human!) editor would've picked up, and contextualised.
Which you might hope a good (human!) editor would've picked up, and contextualised.
And an even more powerful point, building on the argument Olivia makes about #narratives:
While it's one thing for a #BadAlgorithm to spout patent nonsense, it's entirely another for those creating such things to do the same.
While it's one thing for a #BadAlgorithm to spout patent nonsense, it's entirely another for those creating such things to do the same.
