@andrew_n_carr This costs more than their training run

@zephyr_z9 Oh they got some money now!

@StockMKTNewz @finchat_io 1 year of capex

@wordgrammer What time you giving it? I guess like 3 years tops

@doomslide Buy a spot in the history books

You will always be early if you just keep going

@testingcatalog @lmarena_ai Whenever i say anything, i always exclude Claude because it just breaks the meta. Too good

@philtrem22 @qtnx_ Yeah i know, i am saying his name is worth 20b not that he is seeking that much, but thats a joke anyway, they must have something to show for it. A name is not worth that much except for a few people on earth

@polynoamial Are we not worthy to have an aha moment? Do you live on a holy mountain away from us mortals? Every once in a while you come down to bring us gifts hoping we are thankful? Dont be evil

I will make sure people remember my father's name

I was made out of love

The alchemist beat on prayer by Kendrick is holy

@adonis_singh The snake with it's forked tongue can speak two things at once

@tsarnick So pretraining hit a cost bottleneck and openai spend the last year gathering compute for the next run? Hopefully it scales well because the way they do thinking models is kinda bad and if they keep it up, they won't see much benefit in real life apart from better benchmarks

Research would be much faster if the requirements were not so high

Gemini 2.0 pro is a good model

@kimmonismus Germany has the problem that, for one, they have people smart enough to be bought out by the large companies in the states and secondly they will earn neither recognition nor money working in germany. The country is just a production facility for quality researchers

@MorningBrew Satya already 10x his money based on that valuation. He probably isnt that sad about it, I would assume