Dipcic AI
Chinese AI model Deepsek AI had launched an earthquake in the American Silicon Valley. Many of the legendary tech companies, including Google, OpenAI, had swelled hands and feet but the condition of the Chinese AI model has now become bad. The lack of chip has opened the pole-patti of this Chinese AI model. At one time it seemed that this Chinese AI model would come out of AI model like Chatgpt, Google Gemini. However, now the whole picture has changed.
What is the matter?
China claimed that this AI model is made entirely on indigenous technology and is not dependent on America technology. Dipcik used the Ascend chip of the Chinese company Huawe to train its R2 model. The company claimed that this chip would work in further development of an AI. However, now the Chinese AI model needs an American chip. Engineers of the company have admitted that Huawei’s chip is not reliable to train the model on such a large scale.
Engineers say that Huawei chips are not suitable for pursuing AI on such a large scale. Its performance is not stable and lack of low speed network connections and software tools. Dipcic will have to depend on the chips of American company NVIDIA to make its R2 model advanced. Huawei chips will only be used for interfaces.
Where is the screw stuck?
Donald Trump has banned the transporting advanced-level chips to China due to trade-war between the US and China. Dipcik’s project is stuck due to hardware shortage. The CUDA software ecosystem is not working in this AI model, due to which its performance is being affected.
According to the rules, NVIDIA cannot sell its most advanced A100 and H100 chips to Chinese company. Dipcik has now collected these chips through Shell companies of Singapore and Middle East to run its AI model. Nevertheless, more NVidia chips will be required to continue the work of this AI model smoothly.
Also read –
New scam in the market, captcha code will be game! Learn how to avoid
