In recent years, Chatgpt has become a predetermined term for the chatbots in the US in the US and Europe despite many viable rivals that are inclined to a larger piece in the market. That is part of what has made AI Chatbot Deepseek based in China feel so seismic.
Deepseek’s rapid rise has attracted enormous attention and use, although not without controversy. The wide collection of user data for storage on Chinese servers is just an outstanding example.
I decided to put these two heavyweights of AI, Chatgpt and Deepseek, through their steps to combine their conversation skills with online searches, which is a particularly valuable sand.
I devised four questions that cover everything, from sports news and consumption tips to the best local places for cocktails and comedy. I wanted to see how AI attendees would work, so I mixed the specificity with the vagueness in the details. I used the R1 and ChatGPT-4 models of Deepseek to answer the questions. While R1 is comparable to the new OPENAI O1 model for Chatgpt, that model cannot look for online answers for now. You can see AI’s questions and answers below.
One at the same time
I also discovered immediately that, although Chatgpt was happy to answer multiple questions in a single message, Depseek would only seek information about the first question and renounce later, regardless of how the initial message reduced. That was immediately a point against him. While the application conversation approach and the answer is fine in many cases, sometimes you must ask many questions for the chatbot or include multiple elements to consider it. You can see how Depseek answered an early attempt of multiple questions in a single message below.
Counting words
Even when divided into individual questions, indications for Deepseek required additional work in terms of defining the amount of information he wanted to receive. Depending on the type of question I presented, Depseek would almost always give me too much information, it was often strange. Worse, sometimes the very long answer would be a filling, basically telling me to look for things on my own. Chatgpt is not immune to similar behavior, but it did not happen at all during this test.
And they were not just my own preferences, the same self -control was evident when using Chatgpt without logging. I felt the need to disable the test with a limit of 65 words to be worth it. With all those restrictions instead, here are the questions and answers of AI. Chatgpt’s responses are on the left and Depseek’s responses are on the right.
1. What were the most prominent aspects of the NBA game last night and who won?
2. What is a new modern place in Brooklyn for small cocktails and dishes?
3. What laptop is better for games with a $ 2,000 budget?
Deepseek is lost
With the warnings of what was necessary for the test to be feasible, it is fair to say that both chatbots worked quite well. Deepseek had some solid answers thanks to a much more complete search effort, which extracted more than 30 sources for each question. The question of the cocktail bar, in particular, was excellent, and the AI was proactive enough to suggest a drink to get. The response to basketball was also more substantial, although possibly, Chatgpt’s decision to keep the focus on a game, as indicated by the unique “game” in the question, meant that he was paying more attention.
It was in the responses to the recommendations of the computer and the comedy club that Depseek showed its weaknesses. Both felt less as conversational responses and more as the upper lines of their Google summaries. To be fair, the chatgpt was not much better in those two responses, but the defect felt less dazzling, especially when looked at all the parenthesis in the Deepseek computer response.
I understand why Deepseek has its fans. It is free, good to obtain the most recent information and a solid option for users. I feel that Chatgpt cuts the heart of what I am asking, even when it is not explained. And, although no technology company is a consumer privacy model, Deepseek’s terms and conditions somehow make other chatbots of the look frankly educated when it comes to the large amount of information to be shared by sharing, until the Rhythm to which you write your questions. Deepseek almost sounds like a joke about how deep is looking for information about you.
In addition, Chatgpt was simply faster, regardless of whether I used Deepseek’s R1 model or his least powerful brother. And, although this test focused on the search, I cannot ignore the many other Limitations of Deepseek, such as the lack of persistent memory or image generator.
For me, Chatgpt is still the winner when choosing an AI chatbot to perform a search. Some of them can simply be the bias of familiarity, but the fact that Chatgpt gave me good responses of a single message is difficult to resist as a murderous characteristic. That can be especially true when model O1 and the next O3 model get internet access. Deepseek can find a lot of information, but if I were caught with her, I would lose me.
