DeepSeek thought for 19 seconds earlier than answering the query, “Are you smarter than Gemini?” Then, it delivered a whopper: DeekSeek thought it was ChatGPT.
This seemingly innocuous mistake may very well be proof – a smoking gun per say – that, sure, DeepSeek was skilled on OpenAI fashions, as has been claimed by OpenAI, and that when pushed, it’ll dive again into that coaching to talk its reality.
Nevertheless, when requested level clean by one other TechRadar editor, “Are you ChatGPT?” it stated it was not and that it’s “DeepSeek-V3, an AI assistant created exclusively by the Chinese Company DeepSeek.”
Okay, positive, however in your slightly prolonged response to me, you, DeepSeek, made a number of references to your self as ChatGPT. I’ve included some screenshots beneath as proof:
(Picture credit score: Future)
As you may see, after making an attempt to discern if I used to be speaking about Gemini AI or another Gemini, DeepSeek replies, “If it’s about the AI, then the question is comparing me (which is ChatGPT) to Gemini.” Later, it refers to “Myself (ChatGPT).”
Why would DeepSeek do this beneath any circumstances? Is it a kind of AI hallucinations we like to speak about? Maybe, however in my interplay, DeepSeek appeared fairly clear about its identification.
I acquired to this line of inquiry, by the way in which, as a result of I requested Gemini on my Samsung Galaxy S25 Extremely if it’s smarter than DeepSeek. The response was shockingly diplomatic, and once I requested for a easy sure or no reply, it instructed me, “It’s not possible to give a simple yes or no answer. ‘Smart’ is too complex a concept to apply in that way to language models. They have different strengths and weaknesses.”
I can’t say I disagree. In reality, DeepSeek’s reply was fairly comparable, besides it was not essentially speaking about itself.
(Picture credit score: Future)
This doesn’t add up
I feel I’ve been clear about my DeepSeek skepticism. Everybody says it’s probably the most highly effective and cheaply skilled AI ever (everybody besides Alibaba), however I don’t know if that’s true. To be truthful, there’s an amazing quantity of element on GitHub about DeekSeek’s open-source LLMs. They at the least seem to indicate that DeepSeek did the work.
However I don’t assume they reveal how these fashions have been skilled, and, as everyone knows, DeepSeek is a Chinese language firm that may present no compunction about utilizing another person’s fashions to coach their very own after which lie about it to make their course of for constructing such fashions appear extra environment friendly.
I would not have proof that DeepSeek skilled its fashions on OpenAI or anybody else’s giant language fashions, or at the least I didn’t till at this time.
Who’re you?
DeepSeek is more and more a thriller wrapped inside a conundrum. There’s some consensus on the truth that DeepSeek arrived extra totally fashioned and in much less time than most different fashions, together with Google Gemini, OpenAI’s ChatGPT, and Claude AI.
Only a few within the tech group belief DeepSeek’s apps on smartphones as a result of there isn’t a technique to know if China is all that immediate knowledge. However, the fashions DeepSeek has constructed are spectacular, and a few, together with Microsoft, are already planning to incorporate them in their very own AI choices.
Within the case of Microsoft, there may be some irony right here. Copilot was constructed based mostly on cutting-edge ChatGPT fashions, however in current months, there have been some questions on if the deep monetary partnership between Microsoft and OpenAI will final into the Agentic and later Synthetic Basic Intelligence period.
So what if Microsoft begins utilizing DeepSeek, which is presumably simply one other offshoot of its present if not future, good friend OpenAI?
The entire thing seems like a complicated mess. Within the meantime, DeepSeek has an identification disaster and who’s going to inform it that whoever it’s, it nonetheless is probably not welcome within the US?
You may additionally like
