n the rapidly evolving world of artificial intelligence, a new contender has emerged from the East, poised to disrupt the established order and redefine the landscape of AI chatbots. DeepSeek, a Chinese startup, has unveiled its groundbreaking R1 chatbot, a free-to-use AI marvel that has taken the tech world by storm. This impressive creation has even managed to temporarily dethrone OpenAI’s ChatGPT from the top spot on Apple’s US App Store, signaling a potential shift in the balance of power within the AI industry.
But what exactly makes DeepSeek’s R1 so remarkable? It’s not just the fact that it’s free; it’s the sheer power and capability of the model. DeepSeek’s R1 is comparable to OpenAI’s impressive o1 “reasoning” model, but without the hefty $20-a-month paywall. This accessibility alone poses a significant challenge to OpenAI’s monetization strategy for ChatGPT, forcing the industry to reconsider the economics of AI.
DeepSeek’s innovation doesn’t stop at offering a free alternative to a paid service. The R1 model was trained using less powerful AI chips, showcasing a remarkable feat of engineering and efficiency. This breakthrough could redefine the way AI models are developed and deployed in the future, potentially leading to more cost-effective and accessible AI solutions for a wider audience.

First Impressions: A Glimpse into the Future of AI
I’ve had the opportunity to explore numerous generative AI tools over the past few years. Naturally, I was eager to put DeepSeek’s R1 to the test and see how it stacks up against the competition, particularly ChatGPT, the current reigning champion of AI chatbots.
After spending several hours with DeepSeek’s R1, my initial impressions are that it has the potential to be a major disruptor in the AI industry. Its power and accessibility are undeniable, and it’s clear that DeepSeek has pushed the boundaries of what’s possible with AI. The R1 model exhibits impressive language processing capabilities, generating human-like text that is often indistinguishable from that written by a human. It can answer questions, summarize factual topics, write stories, and translate languages, all with remarkable accuracy and fluency.
However, like other generative AI tools, DeepSeek’s R1 is not without its flaws. It still suffers from common issues such as hallucinations, invasive moderation, and questionable sourcing of information. These challenges, while not unique to DeepSeek’s R1, highlight the ongoing need for improvement and refinement in the field of AI.
Accessing DeepSeek’s R1: A Multi-Platform Experience
DeepSeek has made its R1 model readily accessible through various platforms, ensuring that users can experience its power and convenience regardless of their preferred device. The R1 chatbot is available through smartphone apps for both Android and Apple devices, as well as a desktop website. This multi-platform approach allows users to seamlessly integrate DeepSeek’s R1 into their daily lives, whether they’re on the go or at their desks.
In addition to its own platforms, DeepSeek has also made its R1 model available through third-party services like Perplexity Pro, further expanding its reach and accessibility. This collaborative approach fosters a more inclusive AI ecosystem, allowing developers and users alike to benefit from DeepSeek’s innovation.

To access the most advanced features and personalize their experience, users will need to create an account, which will track their chats and preferences. DeepSeek’s privacy policy states that the information collected is stored on secure servers located in China. It’s important to be mindful of this and avoid sharing sensitive personal information during interactions with the chatbot.
DeepSeek vs. ChatGPT: A Comparative Analysis of Features
While DeepSeek’s R1 offers a compelling free alternative to ChatGPT, it’s important to understand the key differences between the two. Both chatbots excel at generating human-like text and performing various language-based tasks, but they also have distinct strengths and weaknesses.
One notable feature of DeepSeek’s R1 is its ability to search the web and incorporate information from external sources into its responses. This allows the chatbot to provide more comprehensive and up-to-date answers, drawing on a vast pool of online knowledge. However, unlike OpenAI, DeepSeek does not have formal agreements with publishers to use their content. This raises questions about copyright and intellectual property, which DeepSeek will need to address as it continues to develop its R1 model.
DeepSeek’s R1 also lacks some of the features that ChatGPT users may be accustomed to, such as the memory feature that allows the chatbot to recall details from past conversations. This can be a significant drawback for users who rely on the chatbot to maintain context and continuity across multiple interactions. Additionally, DeepSeek’s R1 does not yet offer voice interaction capabilities, although the company is reportedly working on multimodal features that would allow users to interact with the chatbot through voice and other modalities.
In terms of search i’ve tested both with a simple “What the top news headlines for today?”, and below are the results compared:




The Hallucination Problem: A Persistent Challenge in AI
Despite its impressive capabilities, DeepSeek’s R1 still struggles with the issue of hallucinations, which is a common problem among generative AI models. Hallucinations occur when the chatbot generates inaccurate or fabricated information in its responses, often with a high degree of confidence. This can be misleading and even harmful, especially in situations where users rely on the chatbot for accurate information.
The hallucination problem is a complex challenge that researchers and developers are actively working to address. It stems from the way AI models are trained, which involves feeding them massive amounts of data and allowing them to learn patterns and relationships within that data. However, this process can also lead to biases and inaccuracies in the model’s understanding of the world, which can manifest as hallucinations in its responses.
DeepSeek, like other AI companies, is aware of the hallucination problem and is actively working to mitigate its effects. However, it’s important for users to be aware of this limitation and to critically evaluate the information provided by the chatbot, especially when making important decisions or forming opinions.
The Data Conundrum: Unveiling the Mystery of Training Data
DeepSeek has been praised for making its R1 model available locally with open weights, allowing users to customize the model’s capabilities and adapt it to their specific needs. This openness and transparency are commendable and contribute to a more democratic and accessible AI ecosystem.
However, the company remains tight-lipped about the specific data used to train the model. This lack of transparency raises concerns about potential biases and limitations in the model’s understanding of the world. It’s essential for AI companies to be open about the data they use to train their models, as this data can significantly influence the model’s behavior and output.

It’s clear that a massive amount of data was required to achieve the level of sophistication exhibited by DeepSeek’s R1. However, without transparency regarding the training data, it’s difficult to assess the potential biases and limitations of the model. This lack of information can erode trust and hinder the development of responsible AI systems.
DeepSeek, like other AI companies, should strive for greater transparency regarding its training data. This would not only foster trust and accountability but also enable researchers and developers to better understand and address potential biases and limitations in the model.
The Future of AI: DeepSeek’s Impact and Beyond
DeepSeek’s R1 has undoubtedly shaken up the AI landscape, challenging the dominance of established players like OpenAI and Google. Its innovative approach to model training and its commitment to accessibility have set a new standard for the industry, forcing competitors to re-evaluate their strategies and offerings.
While it’s too early to declare DeepSeek’s R1 as a death blow to American AI leadership, its success raises important questions about the future of AI development. It’s likely that companies like OpenAI will learn from DeepSeek’s achievements and incorporate its innovations into their own models. This cross-pollination of ideas and technologies can only benefit the AI field as a whole, leading to more powerful and accessible AI solutions for everyone.
DeepSeek’s R1 is not just a powerful AI tool; it’s a symbol of the changing dynamics in the global AI landscape. As China continues to invest heavily in AI research and development, we can expect to see more groundbreaking innovations emerge from the East, pushing the boundaries of what’s possible and shaping the future of AI. This increased competition and collaboration on a global scale will accelerate the pace of AI development, leading to new breakthroughs and applications that were once thought impossible.
Verdict
DeepSeek’s R1 is a remarkable achievement in AI, offering a powerful and accessible alternative to ChatGPT. While it still faces challenges such as hallucinations and censorship, its innovative approach to model training and its commitment to open weights make it a force to be reckoned with in the AI industry. As DeepSeek continues to refine and develop its R1 model, it has the potential to become a leading player in the global AI landscape, shaping the future of AI and its impact on our lives.