Boom! Have you seen what OpenAI's o3 model can do now? PhD level maths and beyond.
And take a look at some of these Veo2 text-to-video demos. We're at the point where they are often (but not always) indistinguishable from genuine video footage now.
Jerrod Lew
Creative project by Jerrod Lew focusing on using AI for storytelling, learning and content creation inspiration. This channel is a personal project.youtube.com
Hollywood as we know it, is over. The music industry, as we know it, is over.
Thanks, I thought you might like the following articles.
The scores on fairly standard pre grad and above average tests have sky rocketed in the latter parts on 2024. The cost is huge, so I'm guessing we won't see mass rollout for some time.
Beyond human: OpenAI’s o3 wake up call
A reasoning breakthrough arrives months - not years - after its predecessor. Here's why we need to talk
www.exponentialview.co
This is a good perspective from an established mathematician, and the comments provide insight too.
Can AI do maths yet? Thoughts from a mathematician.
So the big news this week is that o3, OpenAI’s new language model, got 25% on FrontierMath. Let’s start by explaining what this means.
xenaproject.wordpress.com
It is still in the realms of number crunching to ascertain exact answers, and apparently making large strides against the blind FrontierMath database, an advanced mathematics dataset used for machine tests, scoring an impressive 25%. Further questions remain on what level those 25% answered are... (there is a range in the data).
I agree with the xena article that true and eye popping advance will come when proofs accompany hard answers, and the next standard will be human usable proofs, consistently to advanced theorems.