Who will benefit most from DeepSeek, the Chinese AI startup that is laying waste to US stocks? Who will be hurt the most?

So is everybody. This is why on Monday after the weekend rout, DeepSeek got to be the #1 download app, and Chatgpt is #2. We’re still looking under the hood.

OpenAI with its access to capital has as wide of AI fields to pioneer. It's moving on from LLMs. And this is to extend the reasoning capacities of AI and moving forward in the direction of AGI.

Another avenue that I personally see that DeepSeek has opened up is in the training. Today, AI companies are paying literally hundreds of thousands of lowly paid people to “teach” AI. But now with “chain of thoughts” exposed by DeepSeek to allow users to “intervene” the inference process, this is now a more effective way of doing “reinformed learning” of the AI model. Incorporating this within the process ecosystem means a more efficient and cost efficient way to upgrading AI.

It is reported that President-elect Donald Trump meets with the TikTok CEO Shou Chew at Mar-a-Lago. Will TikTok take a turn for the better?

Silicon Valley is still franctically figuring this out.

Meanwhile, there are the new opportunities. Smaller AI startups now have the open source option to develop for much greater deliverables with much less capital. This is expanding the horizons for smaller teams with programming skills to widen AI capacities to discover new fields and user cases to enrich our daily lives.

What is known so far is that DeepSeek was able to distill large trained frontier models to derive synthetic dataset of larger model to train to get its smaller models. The integrity for frontier technology is still intact but OpenAI & co certainly now realize that the methodologies of the open source approach can deliver faster at less cost AI applications to market. They must now recalibrate pricing to justify their budget.

Pharmacists warn drug shortage affecting cancer patients - BBC

DeepSeek is democrartizing the field of AI development. Processing firepower is just as critical but there is always room to squeeze out efficiency. And this is how Huawei and Xiaomei are managing without EUV and yet get their chips to perform at the 2nm level.