DeepSeek live updates — OpenAI cries foul, Nvidia impact and latest news
There's a lot to unpack with this new AI player
What's going on at DeepSeek?
With users both registered and waitlisted eager to use the Chinese chatbot, it seems as though the site is down indefinitely. The homepage appears as normal, but once users attempt to log in they are blocked with a number of messages.
I have gotten "site underconstruction" and "unable to connect" and "major outage." When it will be back up is unclear. Luckily, there plenty of AI chatbots to consider no matter what your query. Until DeepSeek is back up, we will have to go back to life before we knew it existed. Hard to believe that was just 2 days ago.
Are you able to get in to DeepSeek? Send me an email at amanda.caswell@futurenet.com to share your experience.
DeepSeek is down
Having trouble logging in to DeepSeek? You aren't alone. I have tried from two computers and my phone and I'm getting a "site down for maintenance" notice. The site's popularity since Monday has made it a target for outages and malicious attacks, but perhaps it truly is down for updates.
What seems like overnight success has brought scrutinity as well as praise for the Chinese chatbot. As the site handles the mounting interest and users begin to join from the waitlist, keep it here as we dive into everything about this mysterious chatbot.
DeepSeek blocked in Italy
Reuters reported that the DeepSeek is inaccessible on Apple and Google app stores in Italy. This development occurred a day after Ireland's Data Protection Commission requested information from DeepSeek regarding its data processing practices.
The removal of DeepSeek from the app stores in Italy highlights the increasing scrutiny that DeepSeek and other AI applications face concerning data privacy and regulatory compliance. Investigations have revealed that the DeepSeek platform explicitly transmits user data — including chat messages and personal information — to servers located in China. The privacy policies found on DeepSeek’s site indicate comprehensive data collection, encompassing device information and user interactions.
Experts and critics warn that freely offering extensive data to the app could lead to exploitation by the Chinese government, potentially leading to surveillance and misuse of personal information.
As AI technologies continue to evolve, ensuring adherence to data protection standards remains a critical concern for developers and users alike.
What DeepSeek's 'cheap model' means for sustainable AI practices
DeepSeek is making waves not just for its performance, but also for its surprisingly low energy consumption. Traditional AI models like ChatGPT, Gemini, Claude, and Perplexity, take up a lot of energy. Training GPT-3, for example, consumed a whopping 552 metric tons of carbon dioxide. Yet, DeepSeek achieved similar results using significantly less computing power and energy. This has significant implications for the environmental impact of AI and the future of energy infrastructure, translating to a smaller carbon footprint and reduced reliance on energy-intensive cooling systems for data centers.
The success DeepSeek has already seen with less budget and less energy, underscores the importance of prioritizing energy efficiency in AI development. By optimizing algorithms and utilizing less power-hungry hardware, the AI industry can significantly reduce its environmental impact.
DeepSeek's energy-efficient model offers a promising path towards greener AI. While potential challenges like increased overall energy demand need to be addressed, this innovation marks a significant step towards a more sustainable future for the AI industry.
This shift towards sustainable AI practices is crucial as global demand for AI continues to skyrocket and DeepSeek's model challenges the assumption that AI development necessitates massive energy investments. This could pave the way for more sustainable energy solutions and a reassessment of future energy infrastructure needs.
DeepSeek's Janus Pro — what to know
More than a comprehensive chatbot, DeepSeek also has image generation capabilities through its model Janus Pro. The open-source nature and impressive performance benchmarks make it a noteworthy development within DeepSeek.
In benchmark tests, Janus Pro has demonstrated superior performance compared to other image generators. According to DeepSeek's technical report, the model outperformed OpenAI's DALL-E 3 and Stability AI's Stable Diffusion in text-to-image generation tasks. These results highlight Janus Pro's advanced capabilities in producing high-quality images from textual prompts.
Exceling in both understanding and generating images from textual descriptions, Janus Pro, introduces enhancements in training methodologies, data quality, and model architecture. These improvements from its predecessor, Janus, result in more stable and detailed image outputs, positioning Janus Pro as a formidable contender in the AI image generation landscape.
The model incorporates 72 million high-quality synthetic images, balanced with real-world data. This extensive dataset enables Janus Pro to generate more visually appealing and contextually accurate images.
With up to 7 billion parameters, Janus Pro's architecture enhances training speed and accuracy in text-to-image generation and task comprehension. This scalability allows the model to handle complex multimodal tasks effectively.
Janus Pro is accessed through platforms like Hugging Face and GitHub. The open-source availability of Janus Pro encourages experimentation and collaboration within the AI community, fostering further advancements in multimodal AI applications.
Understanding the differences and similarities between DeepSeek and ChatGPT
The AI world is abuzz with DeepSeek, the Chinese startup DeepSeek's namesake chatbot. I’ve compared the two with various prompts, but let’s take a look at their similarities and differences. While both are revolutionizing how we interact with machines, it is no secret that their differences are as striking as their similarities.
OpenAI keeps the inner workings of ChatGPT hidden from the public. DeepSeek, in contrast, embraces open source, allowing anyone to peek under the hood and contribute to its development. This fosters a community-driven approach but also raises concerns about potential misuse.
Both excel at tasks like coding and writing, with DeepSeek's R1 model rivaling ChatGPT's latest versions. However, DeepSeek's affordability is a game-changer. Developed at a fraction of the cost, it demonstrates that cutting-edge AI doesn't have to break the bank.
ChatGPT offers a free tier, but you'll need to pay a monthly subscription for premium features. DeepSeek, on the other hand, believes in democratizing access to AI. This has fueled its rapid rise, even surpassing ChatGPT in popularity on app stores. Giving everyone access to powerful AI has potential to lead to safety concerns including national security issues and overall user safety.
ChatGPT is a complex, dense model, while DeepSeek uses a more efficient "Mixture-of-Experts" architecture. This allows it to punch above its weight, delivering impressive performance with less computational muscle.
DeepSeek operates under the Chinese government, resulting in censored responses on sensitive topics. ChatGPT, while moderated, allows for a wider range of discussions. This raises ethical questions about freedom of information and the potential for AI bias.
DeepSeek's arrival has sent shockwaves through the tech world, forcing Western giants to rethink their AI strategies. However, its data storage practices in China have sparked concerns about privacy and national security, echoing debates around other Chinese tech companies.
ChatGPT and DeepSeek represent two distinct paths in the AI environment; one prioritizes openness and accessibility, while the other focuses on performance and control. Their contrasting approaches highlight the complex trade-offs involved in developing and deploying AI on a global scale.
Is DeepSeek safe?
Right now no one truly knows what DeepSeek’s long-term intentions are. DeepSeek appears to lack a business model that aligns with its ambitious goals. Unlike major US AI labs, which aim to develop top-tier services and monetize them, DeepSeek has positioned itself as a provider of free or nearly free tools—almost an altruistic giveaway. While this approach could change at any moment, essentially, DeepSeek has put a powerful AI model in the hands of anyone — a potential threat to national security and elsewhere.
DeepSeek’s rise demonstrates that keeping advanced AI out of the hands of potential adversaries is no longer feasible. As Anthropic co-founder Jack Clark noted, “DeepSeek means AI proliferation is guaranteed.”
The Chinese government will undoubtedly get more involved. It’s hard to imagine Beijing remaining silent about a Chinese company freely distributing powerful AI tools, even to adversarial nations.
In the meantime, DeepSeek’s broader ambitions remain unclear, which is concerning. In many ways, it feels like we don’t fully understand what we’re dealing with here. One of the less-discussed aspects of DeepSeek’s story is the foundation of its success. While its v3 and r1 models are undoubtedly impressive, they are built on top of innovations developed by US AI labs.
Today, DeepSeek shows that open-source labs have become far more efficient at reverse-engineering. Any lead that US AI labs achieve can now be erased in a matter of months. This is a serious challenge for companies whose business relies on selling models: developers face low switching costs, and DeepSeek’s optimizations offer significant savings.
While U.S. AI labs have faced criticism, they’ve at least attempted to establish safety guidelines. DeepSeek, on the other hand, has been silent on AI safety. If they have even one AI safety researcher, it’s not widely known. Accelerationists might see DeepSeek as a reason for US labs to abandon or reduce their safety efforts.
But it’s crucial to remember that the most pressing AI safety challenges remain unsolved. If a lab unexpectedly releases superhuman intelligence, there’s no guarantee it will align with human values or goals—and no clear plan for what to do next. This eventually leaves us vulnerable, in both AI development and global security.
Nvidia stock seeing recovery after largest single-day loss in US corporate history
The world of AI experienced a dramatic shakeup this week with the rise of DeepSeek. The Chinese AI startup sent shockwaves through the tech world and caused a near-$600 billion plunge in Nvidia's market value.
The unveiling of DeepSeek’s V3 AI model, developed at a fraction of the cost of its US counterparts, sparked fears that demand for Nvidia's high-end GPUs could dwindle. This concern triggered a massive sell-off in Nvidia stock on Monday, resulting in the largest single-day loss in US corporate history. The ripple effect also impacted other tech giants like Broadcom and Microsoft.
However, the panic proved short-lived. Nvidia's stock bounced back by almost 9% on Tuesday, signaling renewed confidence in the company's future. Experts point out that while DeepSeek's cost-effective model is impressive, it doesn't negate the crucial role Nvidia's hardware plays in AI development. In fact, the emergence of such efficient models could even expand the market and ultimately increase demand for Nvidia's advanced processors.
DeepSeek's breakthrough has seen mixed reactions. While Microsoft and OpenAI CEOs praised the innovation, others like Elon Musk expressed doubts about its long-term viability. Nvidia itself acknowledged DeepSeek's achievement, emphasizing that it aligns with US export controls and shows new approaches to AI model development. It’s clear that the crucial "inference" stage of AI deployment still heavily relies on its chips, reinforcing their continued importance in the AI ecosystem.
The past few days have served as a stark reminder of the volatile nature of the AI industry. Disruptive innovations like DeepSeek can cause significant market fluctuations, but they also demonstrate the rapid pace of progress and fierce competition driving the sector forward.
OpenAI probes if DeepSeek improperly acquired AI data
Tech giants Microsoft and OpenAI have launched an investigation into a potential data breach from the group associated with Chinese AI startup DeepSeek. The probe surrounds a look into the improperly acquired data from OpenAI's technology.
This investigation was sparked when Microsoft's security researchers observed what they believe to be DeepSeek-linked individuals extracting a substantial amount of data through OpenAI's application programming interface (API) in the fall of 2024.
The scale of data exfiltration raised red flags, prompting concerns about unauthorized access and potential misuse of OpenAI's proprietary AI models. Implications of this alleged data breach are far-reaching.
OpenAI, known for its groundbreaking AI models like GPT-4, has been at the forefront of AI innovation. Their technology, accessible through APIs, has become a cornerstone for numerous applications across various industries. These APIs allow software developers to integrate OpenAI's sophisticated AI models into their own applications, provided they have the appropriate license in the form of a Pro subscription of $200 per month.
The potential data breach raises serious questions about the security and integrity of AI data sharing practices. As AI technologies become increasingly powerful and pervasive, the protection of proprietary algorithms and training data becomes paramount.
Is DeepSeek safe to use?
Now that DeepSeek has risen to the top of the App Store, you might be wondering if this Chinese AI platform is dangerous to use. The platform has already been subject to a "large-scale " cyberattack, which has limited registrations.
In addition, major privacy concerns have been raised about DeepSeek. The service's policy explicitly states that DeepSeek's partners, including advertisers, share information about your actions outside of DeepSeek. These include activity on other websites as well as products or services purchased.
For more info see our Is DeepSeek safe to use? guide.
Could DeepSeek be banned in the U.S.?
DeepSeek's chatbot is designed to comply with Chinese government regulations, which mandate adherence to "socialist values." As a result, the chatbot avoids or censors discussions on topics deemed sensitive or politically controversial by Chinese authorities. For instance, when asked about events like the 1989 Tiananmen Square protests, the chatbot may decline to provide information or redirect the conversation.
On the heels of the TikTok ban in the U.S., DeepSeek is raising concerns and some countries are considering regulatory actions in response. There is still so much that we simply don’t know about DeepSeek. For instance, although the app is free now, it could start subscriptions at any time, potentially locking out users.
While DeepSeek's technological advancements are noteworthy, its data handling practices and content moderation policies have raised significant concerns internationally. The potential for censorship reflects a broader apprehension about differing approaches to user data management between China and other nations.
As a result, discussions about potential bans or restrictions are emerging, highlighting the need for users and policymakers to carefully consider the implications of adopting unknown platforms.
How does DeepSeek's AI technology differ from others?
The latest model, DeepSeek-R1, released in January 2025, focuses on logical inference, mathematical reasoning, and real-time problem-solving. It was trained using reinforcement learning without supervised fine-tuning, employing group relative policy optimization (GRPO) to enhance reasoning capabilities. This model achieves performance comparable to OpenAI's o1 across various tasks, including mathematics and coding with an accuracy rate of 97.3% on the MATH-500 test.
DeepSeek's AI models are distinguished by their cost-effectiveness and efficiency. For instance, the DeepSeek-V3 model was trained using approximately 2,000 Nvidia H800 chips over 55 days, costing around $5.58 million—substantially less than comparable models from other companies.
Beware of false info from fake accounts
To prevent any potential harm, we reiterate that @deepseek_ai is our sole official account on Twitter/X.Any accounts:- representing us- using identical avatars- using similar namesare impersonations.Please stay vigilant to avoid being misled!January 28, 2025
As reported by Tom's Guide yesterday, DeepSeek's popularity has resulted in outages, malicious attacks, and temporary limits on user accounts. In addition to technical challenges, DeepSeek has faced issues related to misinformation and impersonation. The company has warned users via Twitter about fake social media accounts impersonating its brand, underscoring the importance of verifying the authenticity of online sources.
Given these developments, users are advised to exercise caution. There is only one app, which can be downloaded from the Apple store and Google Play. New user accounts are temporarily limited to users with Chinese phone numbers, so any individual hoping to use DeepSeek should be vigilant about potential fake accounts and verify the authenticity of any DeepSeek-related profiles or communications.
While DeepSeek's rapid rise in the AI sector is noteworthy, it's crucial for users to remain informed and cautious. Be sure to engage with authentic sources and remain aware of impersonation attempts.
How does DeepSeek's censorship compare to other AI chatbots?
DeepSeek's popularity has been followed by debates over its censorship practices and data handling. While some users appreciate its advanced capabilities and cost-effectiveness, others are wary of the implications of its adherence to Chinese censorship laws and the potential risks to data privacy.
Unlike more familiar chatbots like ChatGPT, Gemini, and Perplexity, that may offer detailed responses on a wide range of topics, including politically sensitive ones, DeepSeek's chatbot aligns its responses with official Chinese narratives. Essentially, DeepSeek provides information that reflects the Chinese government's stance or chooses not to address certain subjects altogether.
DeepSeek's privacy policy indicates that user data, including chat interactions, is stored on servers located in the People's Republic of China. This has raised alarms about the potential for user data to be accessed by the Chinese government, especially given China's stringent data laws. There are also questions about how the Chinese government could use the user data and share it with the hedge fund for trading insights.
There is still so much unknown about this powerful AI agent.
Will DeepSeek be banned?
The big question is whether DeepSeek will survive in the US since a Chinese firm owns it. Will President Trump step in and ban or restrict the AI model and its chatbot?
It's too early to tell. While Trump called the rise of DeepSeek "a wake-up call" for US tech firms, he didn't comment on his long-term plans for the chatbot. He also called it a positive for the US AI space. "If you could do it cheaper, if you could do it [for] less [and] get to the same end result. I think that's a good thing for us," he told reporters on board Air Force One.
He rounded out his quick questioning session by saying he was not concerned and believed the US would remain dominant in the field.
An outage already?
Unsurprisingly, DeepSeek gained public attention and was immediately hit by a massive outage. The company blamed "large-scale malicious attacks" for the outage.
It didn't seem to hurt the AI tool's popularity any. It might have boosted it, as more publications covered the tool based on these attacks. If the attackers planned to slow down DeepSeek's momentum, it doesn't seem the plan worked.
Thankfully, everything is up and running today, so you should be able to try DeepSeek without any issues. What prompt will you try first?
Why are stocks falling?
The term "disrupts" is thrown around a lot in the tech space. But in the case of DeepSeek, it seems to be disrupting both the landscape in AI and the tech world.
Tech stocks are dropping in price as people speculate that chips won't be in nearly as high demand as first expected. If DeepSeek can make its AI model on a fraction of the power, what else can be done when the open-source model makes its way into the hands of more developers?
It's not just Nvidia stocks that dropped, either. ASML dropped by 6%, and Broadcom dipped by 17%.
"DeepSeek has taken the market by storm by doing more with less," said Giuseppe Sette, president at AI market research firm Reflexivity, in an email to CBSNews. "This shows that with AI the surprises will keep on coming in the next few years."
The idea of massive AI profits also inflates tech stocks. If a Chinese firm can make a model this powerful for cheap, what does that mean for all that AI money?
OpenAI's CEO weighs in
deepseek's r1 is an impressive model, particularly around what they're able to deliver for the price.we will obviously deliver much better models and also it's legit invigorating to have a new competitor! we will pull up some releases.January 28, 2025
With a new AI model making waves, it was only a matter of time before OpenAI's CEO Sam Altman offered his thoughts on the model. While he praised it as an "impressive model, particularly around what they're able to deliver for the price," he also said, "we will obviously deliver much better models."
It sounds like Altman sees the rise of DeepSeek as an opportunity for his company, citing that it's "invigorating to have a new competitor."
We'll have to wait and see if OpenAI is still excited based on how well DeepSeek catches on, but if the early hype is any indication, it could be a big deal in the AI game.
Images, too
As you might expect from a feature-packed AI chatbot, you can make images with DeepSeek's tools. The AI image maker is called Janus Pro, and it rivals many of the big names in the space, at least according to early testing. Midjourney and DALL-E need to take notice. Like its primary AI model, it's being trained on a fraction of the power, but it's still just as powerful.
Unfortunately, it's not available to the public yet, so if you're interested in trying it, you'll need to wait a bit longer. You can read about how it works on Hugging Face.
How to join DeepSeek
Joining DeepSeek and getting in on the fun is a relatively painless process. We have a complete guide breaking down each step individually, but if you've ever signed up for an online service, it should be mostly self-explanatory.
Here's the quick and dirty version of how to join DeepSeek:
1. Go to the website or download the mobile app
2. Choose a login method (Google or creating an account from scratch)
3. Enter all information if creating an account
4. Click the Google account you want to use if doing Google sign in
5. Confirm your sign up through email
It really is that easy. You don't need payment information or anything else. A valid email address is all you need to start crafting prompts and blowing your mind with the new AI trend.
DeepSeek vs ChatGPT
ChatGPT is the first name people think of when they mention AI chatbots. It's the one that rose to prominence early in the AI craze, and it's still one of the most well-rounded tools on the market. Considering that, we had to put DeepSeek against ChatGPT to see which tool could take our seven prompts and provide better results.
Read our ChatGPT vs DeepSeek piece for all the details regarding each of the seven prompts if you want all the details. I'm surprised that DeepSeek R1 beat ChatGPT in our first face-off. For a company that spent a fraction of the money on raw computing power to step in and win the leader in AI is genuinely astonishing.
What is DeepSeek?
At its core, DeepSeek is an AI model that you can access through a chatbot, similar to ChatGPT and the other major players in the AI space. However, what's making everyone notice is how much less powerful the systems that trained it are compared to those of other AI companies.
Despite the smaller investment (thanks to some clever training tricks), DeepSeek-V3 is as effective as anything already on the market, according to AI benchmark tests. With a fraction of the computing power, the company behind the AI managed to build a tool that rivals its competitors.
It is worth noting that this new model delivers almost the same performance as OpenAi's much-praised o1 model and is also open source. That means any AI researcher can apply what they have learned to the tool, which could lead to a massive breakthrough in the coming months and weeks.
If you want to learn more about it, look at our DeepSeek R1 deep dive that runs through everything in much greater detail.
