kristintynski

17 ms·

151.
by kristintynski 7y ago
It could be easily trained with additional fine-tuning to have whatever bias you want him to have that is different from his current biases. What makes it so scary is the volume of realistic text that can be generated. It could be used to m
152.
by kristintynski 7y ago
I trained this by fine-tuning the Salesforce CTRL transformer model, then surveyed 1,000 to see if people could tell the difference between RoboTrump and real Trump. They can't.
153.
RoboTrump – A fully convincing Trump AI text generator(lawsuit.org)
33 points by kristintynski 7y ago|12 comments
154.
by kristintynski 7y ago
Written with Grover model (transformer), from AllenAI (IMO, does way better with this type of content than anything I've seen from GPT-2)
155.
660 Semi-convincing marketing articles written by AI in 45min(thismarketingblogdoesnotexist.com)
6 points by kristintynski 7y ago|1 comments
156.
by kristintynski 7y ago
hahah, my brother and I also sold gamed StumbleUpon front page, and digg, and early reddit, and delicious back in the day and sold it as a service. It was insane how easy it used to be...
157.
by kristintynski 7y ago
It is curated, but not heavily. I usually took the first result of the prompt. Occasionally i'd skip a result that was totally off, but all of them were the best out of the first 2-3 for any given prompt.
158.
by kristintynski 7y ago
working on it... sorry!
159.
This Marketing Blog Does Not Exist(thismarketingblogdoesnotexist.com)
107 points by kristintynski 7y ago|50 comments
160.
Exploring the reasons behind 193 failed startups(research.frac.tl)
13 points by kristintynski 10y ago|0 comments