There has been important breakthroughs in understanding Large Language Models (the most consequential seems to be the Simulator theory by Janus (September 2022)).
Also, the deployment of practically useful systems reached entirely new levels in the last few months. Rapid progress in text-to-image, breakthroughs associated with ChatGPT and related systems, etc.
We are approaching the third anniversary of the GPT-3 revolution, and it seems that a new rather unprecedented "phase transition" is starting to happen.
Comments 3
There has been important breakthroughs in understanding Large Language Models (the most consequential seems to be the Simulator theory by Janus (September 2022)).
Also, the deployment of practically useful systems reached entirely new levels in the last few months. Rapid progress in text-to-image, breakthroughs associated with ChatGPT and related systems, etc.
We are approaching the third anniversary of the GPT-3 revolution, and it seems that a new rather unprecedented "phase transition" is starting to happen.
Reply
GPT-4: The Third Deep Learning revolution
AlexNet; GPT-3; GPT-4
Reply
(the intervals between those revolutions are getting shorter: 7.5 years, 3 years, 18 months)
Reply
Leave a comment