DeepSeek - new AI model from China

• Silicon Valley Is Raving About a Made-in-China AI Model: As of Saturday, DeepSeek models R1 and V3 were ranked in the top 10 on Chatbot Arena, a platform hosted by University of California, Berkeley, researchers that rates chatbot performance. A Google Gemini model was in the top spot, while DeepSeek bested Anthropic’s Claude and Grok from Elon Musk’s xAI.
  
  DeepSeek said training one of its latest models cost $5.6 million, compared with the $100 million to $1 billion range cited last year by Anthropic.



• How China’s new AI model DeepSeek is threatening U.S. dominance: “To see the DeepSeek new model, it’s super impressive in terms of both how they have really effectively done an open-source model that does this inference-time compute, and is super-compute efficient,” Microsoft CEO Satya Nadella said at the World Economic Forum in Davos. “We should take the developments out of China very, very seriously.”
  
  DeepSeek also had to navigate the strict semiconductor restrictions that the U.S. govt has imposed on China, cutting the country off from access to the most powerful chips, like Nvidia’s H100s. The latest advancements suggest DeepSeek either found a way to work around the rules, or that the export controls were not the chokehold Washington intended.
  
  “They can take a really good, big model and use a process called distillation,” said Benchmark General Partner Chetan Puttagunta. “Basically you use a very large model to help your small model get smart at the thing you want it to get smart at. That’s actually very cost-efficient.”



• DeepSeek R1 Explained to your grandma:




• How small Chinese AI start-up DeepSeek shocked Silicon Valley: DeepSeek’s R1 release sparked a frenzied debate in Silicon Valley about whether better resourced US AI companies, including Meta and Anthropic, can defend their technical edge.
  
  Industry insiders say DeepSeek’s singular focus on research makes it a dangerous competitor because it is willing to share its breakthroughs rather than protect them for commercial gains. DeepSeek has not raised money from outside funds or made significant moves to monetise its models.
  
  DeepSeek claimed it used just 2,048 Nvidia H800s and $5.6mn to train a model with 671bn parameters, a fraction of what OpenAI and Google spent to train comparably sized models.
  
  Ritwik Gupta, AI policy researcher at the University of California, Berkeley, said DeepSeek’s recent model releases demonstrate that “there is no moat when it comes to AI capabilities”. “The first person to train models has to expend lots of resources to get there,” he said. “But the second mover can get there cheaper and more quickly.”
  
  Gupta added that China had a much larger talent pool of systems engineers than the US who understand how to get the best use of computing resources to train and run models more cheaply.



• Twist in the tale: Nvidia Stock May Fall As DeepSeek’s ‘Amazing’ AI Model Disrupts OpenAI.. America’s policy of restricting Chinese access to Nvidia’s most advanced AI chips has unintentionally helped a Chinese AI developer leapfrog U.S. rivals who have full access to the company’s latest chips. This proves a basic reason why startups are often more successful than large companies: Scarcity spawns innovation.



• The Empire Strikes Back: China Prepares One Trillion Yuan AI Plan to Rival $500 Billion US Stargate Project.

Driverless Tata Nano

Just so you don't feel that Google and Tesla are having all the fun with autonomous self-driving AI vehicles, a robotics researcher in Bangalore has modified a Tata Nano to turn it into a driverless vehicle:



Perhaps Ratan Tata should take a closer look at what this fellow has done. Maybe sponsor it for a cross-country self-driving road tour. It would at least be impressive for the trade shows. But "world's cheapest driverless car" also has a nice ring to it. Made in India!

Consider that if a large number of such cheap systems were to be put into operation, they could each be transmitting back information on various situations the vehicle encounters - just like your web browser sends information back each time it crashes. That accumulating pool of information could be continually analyzed to make the vehicle's AI smarter and more robust. The cheapest system would become the most widely deployed system would become the smartest and most robust system.

Given that companies like Uber are already clamouring for driverless cars to improve their business model, one of India's homegrown Uber-rivals might benefit from looking at this robo-Nano invention. The technology could also be adapted for defense applications and even space systems, after being stress-tested in the everyday world.

Maybe if any of Modi's "smart cities" take off, they'd be the ideal place to showcase such smart self-driving vehicles, and such cities could even be designed around facilitating them.