How open AI DeepSeek will likely disrupt the market

Jaspreet Bindra, Founder of AI&Beyond and Tech Whisperer Ltd, gives us a ringside view of the changes taking place as a result of DeepSeek entering the market. Excerpts from a freewheeling video interview.

- Sunil Rajguru

The difference between DeepSeek and chatGpt… A consumer may find a very similar experience. Both of them are AI chatbots. The scary thing is you’ll find a very similar experience because the level of DeepSeek is, depending on multiple parameters, approximately similar to ChatGPT or Claude or Gemini. Why it has created such a furor is because of the way it has been built. They are completely two different philosophies. There are three things. One, the total cost of building DeepSeek has been claimed to be about $5-6 million as against hundreds of millions of dollars for ChatGPT. Two, the GPUs used for DeepSeek were around 2000 chips which were of last or last to last generation from Nvidia. ChatGPT4.0/4.1 would use hundreds of thousands of latest generation chips. Three, while ChatGPT is a closed model (even though the company is called OpenAI), the real open AI is DeepSeek because they have published a very detailed technical paper which proves all these things. That has set the cat amongst the pigeons. The size of the model itself, in terms of the number of parameters, DeepSeek is 10x to 15x smaller than ChatGPT. My AI&Beyond co-founder has already downloaded it on his very normal laptop locally and is running it without an Internet connection. For Claude or ChatGPT you require stargate like data centers. DeepSeek will change the narrative of AI forever much the same way as ChatGPT changed it about 2 years back.

Losers and gainers in this disruption…