Chinese AI startup DeepSeek has announced that its Janus-Pro-7B model has surpassed competitors, including OpenAI’s DALL-E 3 and Stability AI’s Stable Diffusion, in benchmark rankings for text-to-image generation. This achievement solidifies DeepSeek’s reputation as a key player in the rapidly evolving AI market.
According to a technical report, the Janus-Pro model builds upon its predecessor by incorporating enhanced training processes, higher-quality data, and advanced scaling, resulting in improved stability and more detailed image outputs. The company credited the inclusion of 72 million high-quality synthetic images, combined with real-world data, for the model’s superior performance.
This success follows the launch of DeepSeek’s new AI assistant based on the DeepSeek-V3 model, which has become the top-rated free app in the US Apple App Store. The news sent shockwaves through the tech industry, leading to declines in shares of companies like Nvidia and Oracle, as investors reassessed the competitive dynamics in AI development.
OpenAI and Stability AI have yet to comment on the claims. DeepSeek’s achievements highlight the growing influence of Chinese firms in cutting-edge AI innovation, setting the stage for heightened competition in the global tech market.
AI-powered study rooms are revolutionising online education in China by offering personalised, tech-driven learning experiences. These spaces cater to students aged 8 to 18, using advanced software to provide interactive lessons and real-time feedback. The AI systems analyse mistakes, adjust course materials, and generate detailed progress reports for parents, who can track their child’s improvement remotely. By leveraging technology, these study rooms aim to make education more engaging and tailored to individual learning needs.
These AI rooms are marketed as self-study spaces rather than traditional tutoring centres, allowing them to navigate China’s strict private tutoring regulations by framing their services as facility rentals or membership plans. This creative positioning allows them to operate within a regulatory grey area, avoiding restrictions on off-campus tutoring for students in grades one through nine. Membership fees range from 1,000 to 3,000 yuan monthly, making them a more affordable long-term alternative to expensive one-on-one tutoring sessions.
Despite their growing popularity, education experts remain sceptical of their educational value. Critics argue that many of these systems lack proper AI functionality, relying instead on preloaded prompts and automated responses. Furthermore, there are concerns that their heavy emphasis on drilling questions to improve test scores may neglect critical thinking and deeper comprehension. However, proponents believe these AI-powered study rooms represent an essential step toward integrating technology into education and expanding access to personalised learning.
The CEO of Japanese IT giant NTT DATA has called for global standards in AI regulation to mitigate the risks posed by the rapidly advancing technology. Speaking at the World Economic Forum in Davos, Switzerland, Abhijit Dubey emphasised that inconsistent regulations could lead to significant challenges. He argued that standardised global rules are essential for addressing issues like intellectual property protection, energy efficiency, and combating deepfakes.
Dubey pointed out that the key to unlocking AI’s potential lies not in the technology itself, which he believes will continue to improve rapidly, but in ensuring businesses are prepared to adopt it. A company’s ability to leverage AI, he said, depends on the readiness of its workforce and the robustness of its data architecture.
He stressed that companies must align their AI strategies with their broader business objectives to maximise productivity gains. ‘The biggest issue isn’t the technology it’s whether organisations are set up to implement it effectively,’ Dubey noted.
The discussion at Davos highlighted the urgent need for collaboration among governments, businesses, and industry leaders to create cohesive AI regulations that balance innovation with risk management.
Meta’s new Ray-Ban smart glasses, featuring a Live AI assistant, promise a futuristic way to interact with the world. Users can ask questions about their surroundings, with the AI offering answers in real time. From recipe ideas to decorating advice, Live AI aims to be a virtual assistant that sees what you see and responds conversationally.
Despite its intriguing potential, Live AI struggles in everyday use. Its responses often state the obvious, like suggesting scrambled eggs when a fridge has two eggs and no milk. Users also find it challenging to remember to use the feature, with a smartphone search frequently feeling more practical and efficient. Moreover, the AI’s suggestions often lack the depth needed to be genuinely useful.
Making Live AI effective requires users to master the art of asking precise, specific questions a skill that doesn’t come naturally to everyone. This, combined with issues like misinterpreting conversations and a short battery life, makes the technology feel less magical in real-world scenarios. While the glasses point to a vision of hands-free AI, they currently struggle to provide a compelling alternative to existing devices.
A robotic puppy named ‘Jennie’ is offering a new way to provide companionship to people living with dementia, anxiety, and other mental health challenges. Developed by Tombot, Jennie is an AI-powered pet designed to mimic the comfort and emotional support of a real dog without the difficulties of pet care. Inspired by Tombot CEO Tom Stevens’ personal experience with his mother’s Alzheimer’s diagnosis, the robotic puppy was created to help reduce loneliness and distress.
Jennie stands out with her lifelike design, a collaboration with Jim Henson’s Creature Shop, best known for the Muppets. Equipped with advanced touch sensors and voice command technology, Jennie responds naturally to petting and verbal instructions, creating a realistic experience for users. Her sound effects, crafted from recordings of Labrador puppies, and an all-day battery life make her a practical and emotionally engaging alternative to traditional pets.
Research supports Jennie’s role in easing symptoms like agitation and hallucinations in dementia patients while also helping reduce anxiety and loneliness in broader mental health contexts. With over 7,500 preorders already received, Jennie’s impact is growing as Tombot explores registering her as a medical device, potentially expanding her reach to hospitals and care facilities worldwide.
Priced around $1,500, Jennie offers an accessible solution for those unable to care for live animals due to health or housing constraints. The US based company continues to improve Jennie’s capabilities with software updates, ensuring this robotic puppy remains a dynamic source of comfort for years to come.
President Donald Trump dismissed criticism from Elon Musk regarding a $500 billion artificial intelligence project announced earlier this week. The venture, named Stargate, involves OpenAI, SoftBank, and Oracle, with plans to construct data centres and create over 100,000 US jobs. Key leaders, including Masayoshi Son of SoftBank, Sam Altman of OpenAI, and Larry Ellison of Oracle, joined Trump at the White House for the announcement.
Musk, who has served as a close adviser to Trump, expressed doubts on social media about the project’s financial backing, claiming SoftBank lacks the necessary funds. The Tesla CEO and world’s richest man has a long-standing rivalry with OpenAI CEO Sam Altman, a central figure in the initiative. Musk questioned whether the collaboration could secure adequate investment for the ambitious venture.
In response, Trump dismissed the concerns, suggesting Musk’s comments stemmed from personal animosity toward Altman. Trump highlighted that the project would not rely on government funding, emphasising that the private entities involved are financially capable. He downplayed any potential fallout with Musk, stating that disagreements are natural in such high-stakes partnerships.
A Trump adviser hinted that tensions between the two could escalate, potentially jeopardising their working relationship. Musk currently leads Trump’s government efficiency initiative, raising questions about how this dispute might affect their collaboration.
UK citizens will soon be able to carry essential documents, such as their passport, driving licence, and birth certificates, in a digital wallet on their smartphones. This plan was unveiled by Peter Kyle, the Secretary of State for Science, Innovation and Technology, as part of a broader initiative to streamline interactions with government services. The digital wallet, set to launch in June, aims to simplify tasks like booking appointments and managing government communications.
Initially, the digital wallet will hold a driving licence and a veteran card, with plans to add other documents like student loans, vehicle tax, and benefits. The government is also working with the Home Office to include digital passports, although these will still exist alongside physical versions. The app will be linked to an individual’s ID and could be used for various tasks, such as sharing certification or claiming welfare discounts.
Security and privacy concerns have been addressed, with recovery systems in place for lost phones and strong data protection measures. Kyle emphasised that the app complies with current data laws and features like facial recognition would enhance security. He also reassured that while the system will be convenient for smartphone users, efforts will be made to ensure those without internet access aren’t left behind.
The technology, developed in the six months since Labour took power, is part of a push to modernise government services. Kyle believes the new digital approach will help create a more efficient and user-friendly relationship between citizens and the state, transforming the public service experience.
Lonestar Data Holdings has unveiled plans to establish the first-ever data centre on the Moon, targeting a launch late next month aboard SpaceX’s Falcon 9 rocket. The data centre, named Freedom, will be integrated with Intuitive Machines’ Athena moon lander, leveraging the Moon’s unique environment for security and energy efficiency.
The company, led by CEO Chris Stott, views the lunar surface as an ideal location for secure, disaster-resilient data storage. Powered by solar energy and equipped with naturally cooled solid-state drives, Freedom is designed for non-latency-sensitive applications such as data recovery. Backup operations will be supported by a ground-based facility in Tampa, US.
As the demand for energy-intensive data centres grows on Earth, Lonestar joins a burgeoning space industry exploring off-planet solutions. While competitors like Lumen Orbit also seek a foothold in this nascent field, challenges remain. High launch costs, maintenance limitations, and the risk of mission failure loom large over these ambitious projects.
With initial funding of nearly $10 million and support from partners including the Isle of Man government and AI firm Valkyrie, Lonestar is set to push the boundaries of data storage and space technology.
Infosys, India’s second-largest software services exporter, anticipates a major shift in the way IT firms approach talent management. Speaking at the World Economic Forum in Davos, CTO Rafee Tarafdar highlighted the evolving market, driven by emerging technologies such as generative AI.
Tarafdar noted that the traditional ‘pyramid’ model, where most employees are at the entry level, may give way to a more dynamic framework. Infosys is actively experimenting with strategies to upskill its workforce while creating roles that did not exist before, including specialists in responsible AI and model engineering.
In addition to re-skilling existing employees, the company has developed bespoke small language models tailored to industries like banking and IT operations, offering these as services to clients. With AI creating both challenges and opportunities, Infosys believes a blend of evolving skills and innovative hiring will shape the future of tech talent.
As the IT sector grapples with rapid innovation, India’s Infosys remains focused on adapting its workforce to meet new demands, ensuring it remains at the forefront of the global technology industry.
Sarah de Lagarde’s world turned upside down after a train accident in 2022 cost her an arm and part of her leg. Offered a basic prosthetic by the NHS, she found its functionality limited and struggled with everyday tasks for nine months. However, her life changed when she received an advanced AI-powered bionic arm that learns her movements and anticipates her intent by detecting muscle signals. With this, she regained the ability to perform delicate actions, such as picking up an egg or pouring water.
The remarkable advancements in robotics are not limited to prosthetics. Intelligent machines are being developed to match human dexterity in fields ranging from healthcare to agriculture. Cambridge-based Dogtooth Technologies has created robots capable of picking soft fruit as delicately as human workers, while other projects aim to handle hazardous nuclear waste or assist in complex industrial tasks. Robots like Boston Dynamics’ Atlas and Google’s DEX-EE are already demonstrating the ability to adapt and learn through embodied AI, mimicking human reflexes and movements.
Despite these breakthroughs, experts caution that AI-driven robots are still years away from matching the natural abilities of the human hand. Current limitations include sensory integration, haptic feedback, and adaptability to new environments. Safety, ethical concerns, and potential job impacts are also key issues as this technology develops. Still, pioneers like de Lagarde envision a future where robotic dexterity can assist not only people with disabilities but also support the elderly and others in need, highlighting the transformative potential of AI.
