A chatbot made by Chinese artificial intelligence startup DeepSeek has rocketed to the top of Appleβs App Store charts in the US this week, dethroning OpenAIβs ChatGPT as the most downloaded free app. The eponymous AI assistant is powered by DeepSeekβs open-source models, which the company says can be trained at a fraction of the cost using far fewer chips than the worldβs leading models. The claim has riled financial markets, sending Nvidiaβs shares down over 12 percent in pre-market training,
Downloads for the app exploded shortly after DeepSeek released its new R1 reasoning model on January 20th, which is designed for solving complex problems and reportedly performs as well as OpenAIβs o1 on certain benchmarks. R1 was built on the V3 LLM DeepSeek released in December, which the company claims is on par with GPT-4o and Anthropicβs Claude 3.5 Sonnet, and cost less than $6 million to develop. By contrast, OpenAI CEO Sam Altman has said GPT-4 cost over $100 million to train.
DeepSeek also claims to have needed only about 2,000 specialized chips from Nvidia to train V3, compared to the 16,000 or more required to train leading models, according to the New York Times. These unverified claims are leading developers and investors to question the compute-intensive approach favored by the worldβs leading AI companies. And if true, it means that DeepSeek engineers had to get creative in the face of trade restrictions meant to ensure US domination of AI.
Nvidia, Microsoft, OpenAI, and Meta are investing billions into AI data centers β $500 billion alone for the Stargate Project, of which $100 billion is thought to be earmarked for Nvidia. Investors and analysts are now wondering if thatβs money well spent, with Nvidia, Microsoft, and other companies with substantial stakes in maintaining the AI status quo all trending downward in pre-market trading.
One of the big questions about the upcoming iPhone SE 4 is whether the display will have a notch or Dynamic Island. Two leakers have presented clashing views on this, with the latest dummy models appearing to support the notch.
The dummy models also give us our clearest view yet of what the new entry-level iPhone may look like, with high-quality photos, and a video β¦
moreβ¦If youβve been following the AI industry, Mistral should be a familiar name by now. The French AI startup with a $6 billion valuation is arguably the biggest AI company working on foundation models in Europe. Alan, on the other hand, isnβt as well known. The health insurance unicorn has been quietly growing to become [β¦]
Β© 2024 TechCrunch. All rights reserved. For personal use only.
If one were to observe that I have written critically about Blue Origin over the last half-decade, they would not be wrong.
The reality is that the space company founded by Jeff Bezos has underperformed. Its chief executive for most of this time, Bob Smith, was poorly regarded by his employees. He brought the worst of "old space" tendencies to Blue Origin from Honeywell. And under Smith's leadership, Blue was litigious, slow, and unproductive.
Frankly, it was a bad look for Bezos. He was pumping something on the order of $2 billion a year into Blue Origin for what, exactly? Lawsuits against NASA? Jokes about BE-4 rocket engine delays?
Β© Blue Origin
Chinese AI firm DeepSeek has emerged as a potential challenger to U.S. AI companies, demonstrating breakthrough models that claim to offer performance comparable to leading offerings at a fraction of the cost. The companyβs mobile app, released in early January, has lately topped the App Store charts across major markets including the U.S., UK, and [β¦]
Β© 2024 TechCrunch. All rights reserved. For personal use only.
