NYU Researcher Questions Fears of Releasing Mythos-Level AI Models
AI researcher at NYU questions release fears while a builder flags training risks.
Ravid Shwartz-Ziv, an assistant professor at NYU’s Center for Data Science, posted that Mythos-level models have existed for several months and asked if earlier fears about releasing them were justified. Will Brown, research lead at Prime Intellect, replied that evidence so far indicates training such models is likely a bigger risk than releasing them. The exchange consists of these two public posts on the topic.
Combined views
19.9K
2 posts, first seen 3d ago
NYU Researcher Questions Fears of Releasing Mythos-Level AI Models
AI researcher at NYU questions release fears while a builder flags training risks.
Ravid Shwartz-Ziv, an assistant professor at NYU’s Center for Data Science, posted that Mythos-level models have existed for several months and asked if earlier fears about releasing them were justified. Will Brown, research lead at Prime Intellect, replied that evidence so far indicates training such models is likely a bigger risk than releasing them. The exchange consists of these two public posts on the topic.
Reactions from ranked influencers
2 posts
Now that we have Mythos-level models several months already, do you think the fear that was of releasing it was right?
@ziv_ravid the evidence thus far suggests that training them in the first place is likely a bigger risk than releasing them