Ideation

Title and Thumbnail

The AI Mess, Explained Why Deepseek Matters
What Is Really Going on With Deepseek The AI Mess, Explained
DeepSeek Just Revealed The Future Of AI

Intro

So everyone just lost their minds because of

[deepseek, deepseek, deepseek, ai model, out of china]

But while all the news is about how insanely well it works, whether the US or China are going to steal your data or how it wiped XXB$ from the stock market

If you look closer what DeepSeek actually did is far more interesting, because it just revealed what is going to happen next with AI.

For example, OpenAI started this AI Wave with ChatGPT and with their models that have the most confusing naming scheme known to man. Seriously what the hell does this mean?

And now you might think they've been beaten by China at their own game, but by complete accident they have built something that could make them win in the end.

NVidia just lost 600B$ in value after the deepseek announcement. What will happen to them now? How have they become the most valuable company in teh world in the first place when just a few years ago they were making graphics cards to make this less pixelated

See there are 3 key things that DeepSeek just showed us about what's going to happen next.

So let's go through them one by one and in the process we are going to untangle the mess that is AI today.

  1. Models are not a defensible advantage

This is an v16 car engine from 1936. don't worry, this makes sense.

And this is the San Francisco office of a small non profit research lab called OpenAI founded in 2015. And in this cramped room OpenAI set out to build engines. In AI the model is the engine, this is your GPT 4, Claude 3.5 Llama 2.

And OpenAI was built with the idea that they were going to research and experiment with new models, new engines and make them available, open and free to use for everyone.

But then, they started to change.

In 2020 as they launched GPT3, they made the model closed source, not publicly available because in their view it could be dangerous.

And up until today, the strategy to win was clear: