piercarlo.bsky.social
@piercarlo.bsky.social
@garymarcus.bsky.social's comment on the o3 vs. ARC-AGI issue: the fundamental point is that it is stated that o3 underwent extensive pre-training on the public ARC tasks. This is not a problem in itself, but it is to the extent that this aspect was not adequately communicated.
o3, AGI, and the art of the demo. Long read on what OpenAI didn’t tell you yesterday. garymarcus.substack.com/p/o3-agi-the...
o3, AGI, the art of the demo, and what you can expect in 2025
OpenAI’s new model was revealed yesterday; its most fervent believers think AGI has already arrived. Here’s what you should pay attention to in the coming year.
garymarcus.substack.com
December 22, 2024 at 7:00 PM
@garymarcus.bsky.social's comment on the o3 vs. ARC-AGI issue: the fundamental point is that it is stated that o3 underwent extensive pre-training on the public ARC tasks. This is not a problem in itself, but it is to the extent that this aspect was not adequately communicated.
arcprize.org/blog/oai-o3-...
o3 reaches 75.7% on ARC-AGI - however, @melaniemitchell.bsky.social notes that the method deviates so much from the test's spirit that a higher score doesn't necessarily mean progress towards AGI. Is another test needed? What do you think?
o3 reaches 75.7% on ARC-AGI - however, @melaniemitchell.bsky.social notes that the method deviates so much from the test's spirit that a higher score doesn't necessarily mean progress towards AGI. Is another test needed? What do you think?
December 21, 2024 at 8:48 AM
arcprize.org/blog/oai-o3-...
o3 reaches 75.7% on ARC-AGI - however, @melaniemitchell.bsky.social notes that the method deviates so much from the test's spirit that a higher score doesn't necessarily mean progress towards AGI. Is another test needed? What do you think?
o3 reaches 75.7% on ARC-AGI - however, @melaniemitchell.bsky.social notes that the method deviates so much from the test's spirit that a higher score doesn't necessarily mean progress towards AGI. Is another test needed? What do you think?
www.theverge.com/2024/12/13/2...
Ilya Sutskever highlights a crucial transition for AI: from the current approach based on pre-training and predictive responses to systems capable of developing more autonomous, creative, and innovation-oriented reasoning.
Ilya Sutskever highlights a crucial transition for AI: from the current approach based on pre-training and predictive responses to systems capable of developing more autonomous, creative, and innovation-oriented reasoning.
OpenAI cofounder Ilya Sutskever says the way AI is built is about to change
“We’ve achieved peak data and there’ll be no more.”
www.theverge.com
December 14, 2024 at 5:48 AM
www.theverge.com/2024/12/13/2...
Ilya Sutskever highlights a crucial transition for AI: from the current approach based on pre-training and predictive responses to systems capable of developing more autonomous, creative, and innovation-oriented reasoning.
Ilya Sutskever highlights a crucial transition for AI: from the current approach based on pre-training and predictive responses to systems capable of developing more autonomous, creative, and innovation-oriented reasoning.