New Delhi, Feb. 8 -- Dear reader,

AI has been dominating the news for the past week, ever since DeepSeek, a Chinese AI startup, shook up the AI ecosystem-and the US stock markets-by producing an advanced and highly competent AI assistant model seemingly out of nowhere, and effectively starting an AI arms race.

Training an AI LLM (large language model) is an extremely cost intensive process. This is the stage at which the model is "taught" using vast amounts of data scraped from the internet. The next stage-called "inference"-is where the AI generates responses to queries. DeepSeek's R1 model's biggest success is the fact that its performance is at least as good as that of Open AI's latest 01 model for ChatGPT, or of Meta's Llama 3.1 mod...