Excellent summary - I had spent the evening yesterday (I’m in Germany) tracking down all the quotes you have in your article, and they together with the o3 performance results made it very clear that OpenAI has just performed another “quantum leap” - a fitting conclusion to the 2024 leap year. It will be interesting to see all the franti…
Excellent summary - I had spent the evening yesterday (I’m in Germany) tracking down all the quotes you have in your article, and they together with the o3 performance results made it very clear that OpenAI has just performed another “quantum leap” - a fitting conclusion to the 2024 leap year. It will be interesting to see all the frantic attempts by the “stochastic parrot” crowd to move the goalposts, just so that the “crown of creation” can feel smug about themselves and go back to navel gazing, head bashing, environmental destruction, and illogical theorizing about made up phenomena. Oh, and in Europe, our “friends in Brussels” are now looking at the Brussels sprouts mess called the EU AI Act, which needs to be rewritten / renegotiated / re-whatevered prontissimo. While they are at it, they should also re-examine their confabulation that the Act will serve as a springboard for European innovation. — Just one little quibble about percentages. In discussing the SWE-Bench Verified results, you state: “We’ve never seen a direct 20% jump before …”. This is an understatement. We see a jump of 20 percentage points from 51% to 71%, which is a 20/51 or roughly 40% jump. — Anyway, Frohe Weihnachten, Feliz Navidad, Happy/Merry Christmas, and a smooth transition into 2025.
Excellent summary - I had spent the evening yesterday (I’m in Germany) tracking down all the quotes you have in your article, and they together with the o3 performance results made it very clear that OpenAI has just performed another “quantum leap” - a fitting conclusion to the 2024 leap year. It will be interesting to see all the frantic attempts by the “stochastic parrot” crowd to move the goalposts, just so that the “crown of creation” can feel smug about themselves and go back to navel gazing, head bashing, environmental destruction, and illogical theorizing about made up phenomena. Oh, and in Europe, our “friends in Brussels” are now looking at the Brussels sprouts mess called the EU AI Act, which needs to be rewritten / renegotiated / re-whatevered prontissimo. While they are at it, they should also re-examine their confabulation that the Act will serve as a springboard for European innovation. — Just one little quibble about percentages. In discussing the SWE-Bench Verified results, you state: “We’ve never seen a direct 20% jump before …”. This is an understatement. We see a jump of 20 percentage points from 51% to 71%, which is a 20/51 or roughly 40% jump. — Anyway, Frohe Weihnachten, Feliz Navidad, Happy/Merry Christmas, and a smooth transition into 2025.