"ChatGPT" developer talks about DeepSeek, the rapidly growing Chinese IoT model

The market was surprised by the achievements of the R1 model recently released by Chinese start-up DeepSeek. In addition, the Chinese start-up opened up the code of its models and revealed in an R1 research paper that it managed to train a highly advanced artificial intelligence system at a fraction of the cost of training OpenAI's o1 model, which led to a plunge in the shares of several major technology companies. It took time, but OpenAI CEO Sam Altman finally reacted to the DeepSeek uproar.

How did Mr Altman react?

S. Altman said he was happy to have a new competitor. 

"The DeepSeek R1 is an impressive model, especially considering what they can offer at this price. We will of course launch much better models, but it is really good to have a new competitor", said the OpenAI founder.

"OpenAI promises innovation

OpenAI is likely to release new versions, which may be related to the already introduced o3 model. This should be officially announced in the coming days or weeks.

S. Altman also stated that he and his team are very happy to be able to continue with their research plan.

"We believe that increased computing power is more important now than ever. The world will want to use a lot of artificial intelligence and will certainly be impressed by the next generation of models. We look forward to bringing you AGI and more," he said.

"Advantages of DeepSeek

However, while OpenAI has made several announcements in response to DeepSeek, there is a fundamental difference between them. 

The Chinese startup uses only 3-5% of the resources OpenAI needs to achieve similar progress with ChatGPT.

One of the problems with current artificial intelligence software is the cost of developing and operating the product.

Building advanced models like o1 can cost tens of millions. The process requires high-end graphics processing units (GPUs) that provide the necessary computing power and energy consumption.

That’s why end products like ChatGPT o1 can’t be free without restrictions. Companies like OpenAI need to cover their costs and make a profit.

Meanwhile, DeepSeek researchers took a different approach to R1 — they found ways to train an advanced reasoning model without having access to the same hardware.

DeepSeek also made it much cheaper to access R1 than ChatGPT, which is a significant change.

Add to that the open source nature of DeepSeek’s models, and it’s easy to see why developers are rushing to try out the Chinese company’s AI and why DeepSeek is so popular on the App Store.