🗿 DeepSeek is Great for AI

In partnership with

GM. Happy Tuesday. How does January seem to get longer and longer every year? And how does everyone on the internet have the same experience? If you reply to today’s newsletter and try to tell me that May has the same number of days as January you will be getting blocked. It’s a glitch in the simulation, it has to be. 

Anyways… Deepseek dropped and crashed the market yesterday. We’re here to present a counter-argument to the Wall St. sentiment.  

Without Further Ado. ☕ *knuckle cracks* ☕ Let’s get into it.

There’s a reason 400,000 professionals read this daily.

Join The AI Report, trusted by 400,000+ professionals at Google, Microsoft, and OpenAI. Get daily insights, tools, and strategies to master practical AI skills that drive results.

Sign up now for free and work smarter, not harder.

Why DeepSeek’s Release is a Good Thing

Monday saw Nvidia shed almost $600 billion in market cap, the biggest one-day loss in U.S. history. So the message was clear – the sky was falling. But, we’re here to tell you that the sky is fine. 

We’re going to skip any technical analysis, or any specific stock predictions. You can get that elsewhere I’m sure. 

Why the DeepSeek Model is So Transformative

First, it’s unfathomably cheaper. Training other models like OpenAI costs $100m+ just for computing power. DeepSeek is doing it for $5m. Quite the savings. 

It’s cheaper because it uses significantly less memory. Around 75% less. Crazy. 

Next, DeepSeek is processing words over 2x as fast, which is pretty significant when you need to process billions of them. 

The real transformative bit is the expert system they have built. This basically only activates specialized experts for tasks when needed, instead of having the entire system running at once. 

“Traditional models? All 1.8 trillion parameters active ALL THE TIME. DeepSeek? 671B total but only 37B active at once. It's like having a huge team but only calling in the experts you actually need for each task.”

This is how they made such a shock. If it’s cheaper, and faster we won’t need Nvidia and the impossibly big numbers of GPUs anymore, right? 

Not exactly. 

We still need GPUs

Everyone freaked out yesterday like DeepSeek meant the end of compute. But we think that’s just shortsighted. 

The advancements DeepSeek made (open source btw), are great for AI in general. Is it bad for OpenAI’s crazy valuations? Maybe for now. But they aren’t going to stop developing. If anything, DeepSeek will push OpenAI forward faster. 

And let’s be very clear – everyone still needs GPUs to do any of this. So everyone can stop freaking out about Nvidia. The Scale CEO even said they are hiding 50k H100s over at DeepSeek. 

THIS IS GREAT FOR CONSUMER PRODUCTS

The cheaper process also opens the door to hundreds of others who don’t have to build entire cities of computing power to compete. 

This is what really moves the needle for us. Everyone building in AI may not have to wait for giants like OpenAI to advance the space, and then build on top. It may now be financially possible to create more, custom reasoning models tailored to very specific products. 

Of course the big guys are still going to win. But a young up-and-coming AI company now has a fighting chance – it’s theoretically possible to build something real with an achievable amount of money.  

We Just Don’t Know Where AI Will Go, Yet

Last week at this time no one knew DeepSeek was a serious player, and now they might be one of, if not the most important. 

These are world changing ideas with implications for the future of all of human society, and you think Wall St. had it perfectly figured out in the first few hours?

AI App of the Week 💻 

AI gets a lot of hype. But how does that help me?

This section is dedicated to No Fomo’s favorite AI tools.

We’ll try to find things that make your life easier. Or we’ll give honest, informative reviews of apps capturing Internet attention.

This Week’s App De Jour: DeepSeek

If you read today’s edition and haven’t checked it out yet, we’ve made it easy for you. 

Reach Over 100 Million Humans

This is NOFOMO, the newsletter keeping you up-to-date on all things emerging tech.

No Fomo is part of Supernova’s media conglomerate - a media network reaching 100 million humans and over 10 billion impressions/month across socials, gaming, newsletters, and podcasts.

Request a media kit.

DISCLAIMER: None of this is financial advice. This newsletter is strictly educational and is not investment advice or a solicitation to buy or sell assets or make financial decisions. Please be careful and do your own research.