Total Assets$Micron Tech(MU.US)$Sandisk(SNDK.US) Does everyone still remember when GPT first came out and the context length was only 4k? Today's large models have increased it to 1M, and the difference in effectiveness is obvious to all. Even today's 1M still can't meet the demands of many AI scenarios. I really can't think of any chip manufacturer that would save a bit on memory costs and tell me to use less HBM. Large models will only get bigger, and their development speed far outpaces the speed of hardware updates.
The copyright of this article belongs to the original author/organization.
The views expressed herein are solely those of the author and do not reflect the stance of the platform. The content is intended for investment reference purposes only and shall not be considered as investment advice. Please contact us if you have any questions or suggestions regarding the content services provided by the platform.
