4 ms·
One part really jumped out at me: Thus only a very small number of companies (currently TSMC, Samsung, and Intel) attempt to operate leading-edge nodes, an
by etaioinshrdlu 2y ago
One part really jumped out at me:
Thus only a very small number of companies (currently TSMC, Samsung, and Intel) attempt to operate leading-edge nodes, and the industry has shifted to a “fabless” model where companies like Apple and Nvidia design their chips but have them manufactured by “foundries” like TSMC. By pooling the orders of many different chip companies, the foundries can achieve the scale necessary to afford cutting edge fabs.
I wonder if AI training will end up being similar in the long-term (it's already partially true today).
- langsoul-com 2y agoIt is already true. The cost of training gpt jumped exponentially per model, in the realm of millions upon millions. There's a good reason why the latest gpt-4 competitor is meta or Google.
- everybodyknows 2y agoChatGPT-4 shows a certain awareness of its own training investment: If asked, it will say it "contains data up to December 2023".