I hadn't thought about the DeepSeek angle. But yes, I would say that in general, any data coming out of China ought to be viewed with the assumption that it has been massaged, especially since China itself probably does not know the truth!

For those who can't read WSJ, Wikipedia has a summary <https://en.wikipedia.org/wiki/Economy_of_China#Disputes_over...>:

>"We do not know the true number of China’s real growth figure", Gao Shanwen, chief economist at SDIC Securities and adviser to the government, said at a Peterson Institute event on 12 December 2024. He thought that the economy had grown by about 2% annually, compared to the official 5% figure. Gao expected that the economy would grow by 3–4% annually over the next several years "but we know the official number will always be around 5%". His words reportedly caused the government to ban Gao from speaking publicly. While he did not lose his job, the Securities Association of China ordered brokerages and investment firms to make sure that their economists "play a positive role"; those that do not obey may be fired. In January 2025, Xi aide Cao Qi ordered propaganda officials to block negative economic news.

With respect to DeepSeek, I assume you mean the relatively-low cost at which it was allegedly trained? Because the model itself can obviously be downloaded and tested directly.