$NVIDIA Corp(NVDA)$  $Taiwan Semiconductor Manufacturing(TSM)$ 

I am an active AI user and I have tried out Deepseek extensively in the past few days and this is my take on Deepseek R1. 

So basically, DeepSeek R1 just made headlines for achieving state-of-the-art AI performance, matching OpenAI's o1 at just 3% of the cost.

1. It used NVIDIA chips—still the king of AI hardware.

2. Cost efficiency is open-sourced, so big tech wins.

3. It’s compared to an older model (gpt-o1), which still holds up.

4. AGI is far off—cost efficiency isn’t the full story.

5. U.S. chip export controls will keep China in check.

Overall, while Deepseek is performing great at this current level, it is extremely optimised. by extremely, I mean the software has eliminated all the uncessary processes. 

In the longer term, AI will continue to progress, my view is that part of the reason for the flabby systems used by the present big tech is because they are not at the stage of trimming processes, but are in a race towards AGI. An analogy for those familiar with bodybuilding, one have to bulk up first before doing all the cardio and diet to trim the fats. 

Happy Lunar New Year, may all of us Ssssssslay in the coming year ahead. 

@TigerStars  @CaptainTiger  @Tiger_Newspress