@tenobrus completely expected and fair tbh
they were all counting on index buying liquidity to keep these inflated IPO valuations going
now they'll have to earn it
Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946.
For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids.
An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better.
This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.
@giffmana these are 9B diffusion models. most individual users will use a hosted service to run them
granted it's marginally better to have published weights but the reason why they do it is they know they're not losing on much
@giffmana Do you think those people will be able to run it on their own hardware?
If not they’ll use a model provider who will have to pay a license fee for each generation so it’s basically the same as a closed model
@svlevine there sadly wasn't enough space in the room for everyone who wanted to attend and dozens of people were lining up outside
was it recorded by any chance?
@datagenproc@distributionat i mean there is most definitely a correlation between job postings and hiring but for a n=1 observation you can't infer much from the presence of job listings
@datagenproc@distributionat not really either
it happens often that companies forget to remove job listings they're no longer hiring for. someone has to go in and clean up so it's lagging
@paschalidoud_1@nvidia was at the talk moments ago and this looks so cool!
thanks for sharing! Any idea when it will be released? will the release also include the multi embedding spaces composability techniques you hinted at during the talk?
"a small but growing number of recent work from various robotics labs have demonstrated that—at least conceptually—a pretrained video renderer can be used as the backbone for joint world-and-action prediction, suggesting a bridge between the renderer and the planner by letting one model imagine what will happen and what to do."