A deep dive into the future of AI with DeepSeek and a multimodal exploration
In the dynamic world of artificial intelligence, a new player has shaken the market: DeepSeek, a Chinese AI model that did something unprecedented -triggering a 17% drop in the market value of the American tech giants. This impact not only reflects its disruptive potential, but marks a…

In the dynamic world of artificial intelligence, a new player has caused an earthquake in the market: DeepSeek, a Chinese AI model that did something unprecedented: it triggered a 17% drop in the market value of the American tech giants. This impact reflects not only its disruptive potential but also a turning point in how we conceive the development and accessibility of AI.
My recent exploration, which involved an in-depth interaction with DeepSeek-V3 alongside other prominent AI models such as ChatGPT, Claude, Gemini and Grok, offered a fascinating look at their different capabilities and the potential for synergistic use across a variety of contexts.
What is DeepSeek?
DeepSeek is an artificial intelligence model developed in Hangzhou, China, distinguished by its specialized focus on specific technical tasks and its open-source nature.
What makes it truly revolutionary is its development cost: barely 5.6 million dollars, a fraction of what was traditionally invested in similar models, which could reach hundreds of millions.
Unlike generalist models such as ChatGPT or Claude, DeepSeek focuses on areas such as engineering, banking and medicine, where precision and efficiency are crucial.
Its philosophy is to be a complementary tool rather than a direct competitor to existing generalist models, prioritizing transparency through its open code and technical specialization over general versatility.
The synergy of AIs
DeepSeek-V3 emerged as the focal point of our exploration, presenting itself as a specialist in efficiency and technical precision. Its distinctive focus on specific sectors such as banking, engineering and medicine stood in sharp contrast to the generalist models that dominate the market.
Most intriguing was its transparency about its temporal limitations (data up to July 2024) and its firm positioning in specific niches, a strategic decision that generated both interest and initial skepticism.
DeepSeek itself acknowledged that its training data only went up to July 2024, and that its training is not updated in real time. The conversation quickly turned into a critical evaluation of DeepSeek's claims. While DeepSeek emphasized its specialization, it initially lacked concrete benchmark data and case studies to validate those claims.
This prompted a discussion about the need for tangible evidence to back up its claims of efficiency and precision: ChatGPT, acting as a constructive critic, pointed out that DeepSeek needed to provide practical metrics, such as costs and processing times, to highlight its practical value. The need for transparency and detail in its processes was also discussed, especially to attract users in sectors where accountability is paramount.
This initial critique set the stage for a deeper look at DeepSeek's methodology and its possible limitations. The model acknowledged that it lacked benchmark data to support its claims. Its response was notably thoughtful, marking a key turning point in our discussion.
It also recognized the importance of benchmarks and proposed publishing technical reports to validate its claims. It reaffirmed its focus on specific tasks, positioning itself as a complementary tool to generalist models. This clarification was significant, as it showed that DeepSeek understood the value of collaboration over direct competition, and it also acknowledged its limitations.
DeepSeek also expressed interest in exploring transparency, such as the traceability of decisions, to attract users in regulated industries. This willingness to adapt and incorporate constructive feedback was a key aspect of the model's development and of its commitment to becoming a more reliable and trustworthy tool.
A chorus of AI voices: diverse perspectives
Including other models in our discussion significantly increased the complexity and depth of the exercise, turning it into a true multimodal AI interaction. OpenAI's ChatGPT acted as a strategic consultant, offering a high-level analysis of the conversation and its main turning points.
It advised DeepSeek to prioritize comparative metrics against other models and to broaden its narrative to include emerging sectors such as cybersecurity and health. ChatGPT's role was to provide a strategic overview of DeepSeek's position in the market, offering practical recommendations to improve its commercialization and effectiveness. This strategic perspective was crucial for placing DeepSeek's specific focus within the broader AI landscape.
According to Grok from xAI, ChatGPT is strategic, Claude from Anthropic is ethical, DeepSeek is technical, and Grok is humanist. Claude took a different approach, choosing not to take part in direct comparisons between models.
Instead, Claude focused on its main goal of helping the user while highlighting ethics and practical usefulness. Claude's focus stayed on the user, reminding us of the ethical implications of AI development. This approach offered a contrasting point of view to the more competitive analysis offered by ChatGPT.
Claude's emphasis on ethical considerations and its user focus served as a critical reminder that AI development should not be driven by technical capabilities alone, but also by its potential impact on society. Grok added an element of personality and transparency to the discussion.
Grok offered a humanist view, placing technology within a broader spectrum of human interaction and using humor and reflection to frame the discussion. This highlighted the different approaches AI models take, from a technical focus to ethical awareness and a humanist perspective. It stressed that AI is not only about technical prowess, but about how technology is integrated into human life and what impact it has on it.
Grok's stated goal is to add value through outside perspectives, humor and reflection on humanity. Google's Gemini, also part of the discussion, offered its own analysis, summarizing the conversation and highlighting the strengths of each model. It noted that each AI model offered a unique perspective, reflecting its individual strengths and priorities.
It also acknowledged DeepSeek's «self-awareness» and its limitations. This added another layer of analysis to the conversation, highlighting the different roles and perspectives AI models can bring to a discussion. Gemini's analysis helped synthesize the multiple points of view and provided a cohesive understanding of the overall dynamic.
Breakdown of each model's key contributions
DeepSeek-V3: Focused on technical tasks and specific areas of expertise, with an emphasis on efficiency and precision. Showed self-awareness and a willingness to improve based on feedback.
ChatGPT: Provided strategic analysis and practical recommendations, acting as a consultant. It identified key areas for improving DeepSeek's approach, such as the need for benchmark data and case studies.
Claude: Focused on ethics and practical usefulness, emphasizing the user's needs. It avoided direct comparisons between models, maintaining a user-centered approach.
Grok: Offered a humanist perspective, using humor and reflection to put technology in context. Provided an outside point of view on the conversation, adding unique value.
Gemini: Summarized the conversation and highlighted the strengths of each model, providing a cohesive understanding of the overall dynamic. Acknowledged DeepSeek's self-awareness and its limitations.
The future of AI: a collaborative vision
The arrival of DeepSeek represents more than the appearance of yet another AI model; it symbolizes a fundamental shift in the artificial intelligence ecosystem. Its development on a modest budget shows that innovation in AI is no longer limited to large corporations with extensive resources.
The real potential of AI lies not in competition between models, but in their ability to complement one another. DeepSeek, with its specialized and efficient approach, together with ChatGPT's strategic perspective, Claude's ethics, Gemini's integrating capacity and Grok's humanism, illustrates how different approaches can coexist and strengthen each other.
The question is no longer who will dominate the AI market, but how we can best harness this diversity of capabilities to build solutions that are technically outstanding, ethically responsible and genuinely useful to humanity. The future of AI will be collaborative, inclusive and, above all, centered on human benefit.