image text translation
(1)OpenAI researchers are threatening humanity
(2)The model is out. Warning
(3)Input 20231124 AM 925
(4)Correction 20231124 AM 1044 Articles
(5)Reporter Kim Mi-jung, TALK
(6)1 a.m
(7)Dataset application solves computational problems by itself
(8)The cause of the dismissal of the OpenAI CEO over the past week is being identified little by little Artificial intelligence AI models powerful enough to threaten humanity have emerged, and several researchers have emailed the board of directors to warn them
(9)On the 22nd, the U.S. Deposition reported that OpenAI researchers sent a letter to the board of directors stating that they had developed a powerful general artificial intelligence AGI. The letter was sent shortly before the dismissal of Sam Altman
(10)According to reports, the AI model name is Q-star. The development speed of the existing model is several tens of times faster. Researchers say they were unable to hide the shock after testing the model’s mathematical computing ability
(11)Photo = OpenAI
(12)In general, Generative AIs such as ChatGPT generate characters by predicting the following words based on statistics The same goes for mathematical operations. They can only be calculated within the scope of data learning You can’t get the answer right to a new math question that you haven’t learned
(13)The cue star is different. You can do it yourself to understand computational problems without learning additional data Because you can apply existing data sets This is why new computational problems that were not in the dataset can be solved skillfully. In other words, new problems can be solved by applying existing knowledge like humans
The researchers wrote in a letter
A model cue star has destroyed humanity
powerful
Can be used as a system
I warned the board that
httpsnnewsnavercomarticle0920002312648