@NaeemShn@PTAofficialpk It's not PTA's. You need to check it with the office administration.
Our speed is around 300 MBs, netflix mobie buffering bht karti hai, apne request ma mere request bh add kardy
DeepSeek R1 is:
27x cheaper than OpenAl's models
Open-source and permissively licensed
A model that fully discloses reasoning steps
This isn't just competition. As Marc Andreesen said It's Sputnik for Al.
The game is changing and fast.
hashtag#deepseek hashtag#Ai hashtag#OpenAi
China's DeepSeek just built a reasoning Al model that rivals OpenAl's best-and did it for 1/1000th of the cost.
OpenAl's GPT-4: $600M+ training cost
DeepSeek: $6M total
This isn't just a cost difference-it's a paradigm shift.
Silicon Valley's Al advantage? It's evaporating.
Tech investors in panic mode
Global attention shifts to China's Al dominance. What if we dont need that many GPUs or even electricity to build better Models!