Did China's DeepSeek Steal from OpenAI?

In partnership with

Hi everyone,

This is Lore, the newsletter that helps you thrive in the age of AI.  

Thank you for being a part of 24k+ techno-optimists and growing. 🙏

P.S. Watch the latest episode of The Next Wave, where we discussed OpenAI’s Stargate Project!

It Costs $0 to Look This Good

Optics matter, and Agree helps you look like the pro you are. Whether you're a founder raising funds, sending Y Combinator SAFEs, or an entrepreneur billing clients, Agree combines (free) e-signature and payments into one seamless platform. Streamline agreements, speed up payments, and impress clients with sleek, branded processes that show you’re running a tight ship. Your business deserves tools that make you look as sharp as you are.

Upgrade for Free

1) Things Are Getting Better

Google’s AI keeps getting better. If you haven’t tried it in a while, I suggest you try their O1-like thinking model. It’s close to O1 quality but with a massive context window. 

2) Did China's DeepSeek Steal from OpenAI?

Chinese AI company DeepSeek is facing serious accusations of stealing data from OpenAI to train its newly announced R1 model. The Financial Times reports that OpenAI and Microsoft claim they have evidence that DeepSeek used the OpenAI API to train their model in a way that was against the terms of service. David Sacks, a famed member of the PayPal Mafia and now head of AI & Crypto for the Trump Administration, says there's "substantial evidence" of the theft.

Meanwhile, users on X have shared screenshots (whose validity I haven’t been able to confirm) purporting to show V3, the underlying LLM powering R1, responding to queries about its model identity with “ChatGPT, developed by OpenAI.”

Last week's R1 announcement briefly rattled markets—causing Nvidia shares to plummet 17% as investors panicked over the prospect of ultra-cheap AI training. DeepSeek's bold claim that they spent only $5.6 million without premium GPUs raised immediate eyebrows—count me skeptical. But even if true, there's no reason that alone should tank Nvidia. After all, bigger models still require more compute power, and there's no upper bound on how smart we want AI to become.

The reality? Most experts remain unconvinced that R1 represents any real leap forward. Anthropic’s CEO Dario commented in a recent blog post that while there might be minor efficiency gains, R1 performs roughly on par with top U.S. models from nine months ago.

My own testing confirms this: while R1 is indeed cheaper to run and admirably open source, its performance trails GPT-4 and Claude. It’s also refreshingly transparent about its thinking process—except, tellingly, when asked about Taiwan or Tiananmen Square.

Despite the hype around R1 itself, China's AI progress is undeniable, and this incident will only accelerate global competition. Sam Altman is already hinting at new OpenAI models that could overshadow DeepSeek's announcement entirely. Anyone hoping for slower, more measured AI development, should prepare themselves. Things are going to accelerate even faster now that China is in the game with open-source AI that anyone can use.

For techno-optimists, it’s both thrilling and sobering. If the U.S. and its allies are vying with China for the first true superintelligence, the leader might stay ahead for decades—if not longer. Free speech and human rights hinge on who sets global norms for advanced AI. So while R1 is intriguing, the broader implications are enormous, with the war for AI dominance in full swing.

AD

Learn how to make AI work for you

AI won’t take your job, but a person using AI might. That’s why 1,000,000+ professionals read The Rundown AI – the free newsletter that keeps you updated on the latest AI news and teaches you how to use it in just 5 minutes a day.

Sign up to start learning.

3) OpenAI’s Stargate Project is INSANE!

In this week’s episode of The Next Wave, Matt Wolfe and I discussed everything you need to know about OpenAI’s Stargate Project. The biggest technology infrastructure project in human history. Check it out! 

Also, please subscribe to The Next Wave on Apple, Spotify, and YouTube. 

4) Things I’m Learning From

That’s all for today. Please consider sharing the newsletter with your friends if you think they’d enjoy it. 🙏

Please give me feedback on Twitter. I am still figuring out the best format for this newsletter. Which part did you like? Dislike? Do you have other suggestions? Please let me know—just tweet @NathanLands.

-Nathan