SEB to adopt new custody platform from S&P Global
0
The new platform will enhance SEB's ability to provide state-of-the-art services to several customer segments.
'The new engagement with S&P Global gives us access to a tested platform that allows for the scale and flexibility we need, and enables quicker product development,' says Ann Magnusson, head of Investor Services at SEB.
Operated within S&P Global's Market Intelligence division, the new platform is API-enabled and designed to support the full lifecycle of post-trade and asset servicing functions.
The solution enables higher levels of automation, faster time to market for new services, and improved operational resilience through modern architecture and exception-based processing.
'Collaborating with SEB on this transformational journey represents a key moment in the evolution of our Securities Processing Platform,' said Ankush Zutshi, Head of Securities Processing & Asset Servicing at S&P Global Market Intelligence.
SEB will gradually move functionality and data to S&P Global's platform over several years, which means customers will also gradually experience improvements.
The bank will use the new platform in all Nordic countries for retail, private banking, family office, and sub-custody customers.
Modern technology and cloud-based storage provide higher operational reliability and easier updates. Scalable solutions such as S&P Global's make it easier to administer and develop new services, benefiting both SEB and customers.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
BBC News
2 hours ago
- BBC News
Arsenal sign Gyokeres from Sporting for £63m
Arsenal have completed the signing of Sweden striker Viktor Gyokeres from Portuguese club Sporting for 73m euros (£63m).The 27-year-old has signed a five-year deal with the Gunners after the move was held up because of negotiations between the clubs over add-on final sum is made up of a 63m euro transfer fee plus 10m euros in move takes Arsenal's summer spending up to about £204m, following the arrivals of Martin Zubimendi, Kepa Arrizabalaga, Christian Norgaard, Noni Madueke and Cristhian Sport revealed at the start of the month how Mikel Arteta's side had turned their attentions to Gyokeres after hitting an impasse in their bid to sign Benjamin Sesko from RB sporting director Andrea Berta flew to Lisbon to lead negotiations with the Portuguese deal concludes the long-running saga over Gyokeres' future, after he was linked with Manchester United earlier this failed to report back at Sporting for pre-season earlier this month, and the club's president Frederico Varandas said the player would face disciplinary action. More to follow.
Geeky Gadgets
8 hours ago
- Geeky Gadgets
New QWEN 3 Coder : Did the Benchmark's Lie?
What if the future of coding wasn't just about writing better code, but about rethinking how code is created altogether? The QWEN 3 Coder, a new open-weight AI model, promises to do just that. With its staggering 480 billion parameters and a token context window that scales up to 1 million, this model is designed to tackle coding challenges at a scale and precision that few others can match. Yet, as with any innovation, it doesn't come without its complexities. While the QWEN 3 Coder excels in practical applications like UI design and automation, it grapples with reasoning-heavy tasks, raising questions about the balance between capability and specialization in AI-driven coding tools. This guide by Prompt Engineering provides more insights into the capabilities, challenges, and future potential of the QWEN 3 Coder, offering an in-depth exploration of what makes it both a powerful asset and a work in progress. From its advanced training methodologies to its open source accessibility, readers will uncover how this model is reshaping the coding landscape while also confronting key limitations like benchmark reproducibility and reasoning inefficiencies. Whether you're a developer seeking to streamline workflows or a researcher curious about the next frontier in AI, the QWEN 3 Coder offers a fascinating glimpse into the evolving relationship between humans and intelligent coding agents. QWEN 3 Coder Overview Model Specifications The QWEN 3 Coder is engineered with scalability and high performance in mind, incorporating advanced specifications that set it apart in the field of AI-driven coding. Key features include: Parameter Count: The model features 480 billion parameters, with 35 billion actively used during runtime, making sure efficient processing of complex tasks. The model features 480 billion parameters, with 35 billion actively used during runtime, making sure efficient processing of complex tasks. Token Context Window: Starting at 256 tokens and scaling up to an impressive 1 million tokens, it can handle extensive coding tasks with ease. Starting at 256 tokens and scaling up to an impressive 1 million tokens, it can handle extensive coding tasks with ease. Training Data: Trained on 7.5 trillion tokens, 70% of which are code-based, providing a strong foundation for coding-related applications. Trained on 7.5 trillion tokens, 70% of which are code-based, providing a strong foundation for coding-related applications. Optimization: Specifically designed for agentic coding, browser integration, and external tool usage, enhancing its versatility. These specifications enable the QWEN 3 Coder to manage intricate, multi-turn tasks and create user interfaces (UIs) with precision. Its scalability and adaptability make it a powerful tool for a wide range of coding applications, from front-end development to automation. Performance Highlights The QWEN 3 Coder demonstrates competitive performance on benchmarks such as SweepBench Verified, showcasing capabilities comparable to those of Claude Sonnet 4. Its strengths are particularly evident in practical coding tasks, including: Designing and implementing front-end interfaces. Creating dynamic animations for web and app development. Generating single-chart visualizations for data representation. For instance, developers automating repetitive coding tasks or designing UIs can rely on the model for efficient and accurate results. However, its performance diminishes in tasks requiring intricate reasoning, such as solving abstract problems or navigating complex mazes without external tools. This limitation underscores the model's focus on practical applications rather than abstract problem-solving. New QWEN 3 Coder : Did the Benchmark Lie? Watch this video on YouTube. Here are more detailed guides and articles that you may find helpful on AI coding. Training and Development The advanced training methodologies employed in the development of the QWEN 3 Coder are central to its capabilities. These include: Pre-Training: The model relies heavily on synthetic data to establish a strong initial learning base, enhancing its ability to handle diverse coding tasks. The model relies heavily on synthetic data to establish a strong initial learning base, enhancing its ability to handle diverse coding tasks. Post-Training: Reinforcement learning techniques are used to refine its capabilities, making sure improved performance over time. Reinforcement learning techniques are used to refine its capabilities, making sure improved performance over time. Infrastructure: Training is conducted on 20,000 parallel environments hosted on Alibaba Cloud, providing scalability and efficiency. Training is conducted on 20,000 parallel environments hosted on Alibaba Cloud, providing scalability and efficiency. Framework: Built on the Gemini CLI framework, the model integrates seamlessly into Quinn Code and Cloud Code ecosystems, enhancing its usability. These training and development strategies ensure that the QWEN 3 Coder is both adaptable and efficient, catering to the diverse needs of developers and researchers. Its ability to integrate with existing ecosystems further enhances its appeal as a versatile coding tool. Community and Open source Contributions As an open source model, the QWEN 3 Coder is accessible on platforms such as Hugging Face and Open Router, fostering collaboration and innovation within the AI community. Its open availability encourages developers and researchers to contribute to its growth and refinement. Notable features include: Support for seamless integration with other coding agents and tools, expanding its functionality. Customizable features that allow users to tailor the model to their specific requirements. Ongoing community efforts to verify benchmark performance claims, making sure transparency and reliability. This collaborative approach not only strengthens the model's utility but also promotes its adoption across various coding environments. By encouraging open source contributions, the QWEN 3 Coder benefits from continuous improvement and innovation. Observations and Trends The QWEN 3 Coder excels in short-duration reasoning tasks, such as generating concise code snippets or resolving straightforward queries. In these scenarios, it often exceeds expectations, delivering results with speed and accuracy. However, its performance declines during prolonged reasoning tasks, particularly those requiring abstract problem-solving or extended logical deductions. This focus on practical coding applications over abstract reasoning reflects broader trends in AI development, where utility and real-world applicability are prioritized. As developers increasingly seek tools that can address immediate, tangible challenges, models like the QWEN 3 Coder are well-positioned to meet these demands. Limitations and Challenges Despite its many strengths, the QWEN 3 Coder is not without its limitations. Key challenges include: Benchmark Reproducibility: Discrepancies in RKGI scores have raised concerns about the model's consistency, particularly in standardized evaluations. Discrepancies in RKGI scores have raised concerns about the model's consistency, particularly in standardized evaluations. Reasoning Challenges: The model struggles with complex reasoning tasks, especially those requiring abstract problem-solving or extended logical analysis. These limitations highlight the need for further optimization and refinement. While the QWEN 3 Coder is a powerful tool for specific applications, it is not yet a comprehensive solution for all coding-related tasks. Addressing these challenges will be crucial for its continued development and adoption. Future Potential and Applications The QWEN 3 Coder stands as a robust and versatile coding model, offering significant potential for practical applications such as code generation, UI creation, and agentic tasks. Its advanced training techniques and open source availability make it a valuable resource for developers and researchers. As the AI community continues to refine and explore this model, it is poised to play a pivotal role in shaping the future of coding and artificial intelligence. By addressing its current limitations and building on its strengths, the QWEN 3 Coder has the potential to become an indispensable tool in the evolving landscape of AI-driven development. Media Credit: Prompt Engineering Filed Under: AI, Technology News, Top News Latest Geeky Gadgets Deals Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.
The Guardian
10 hours ago
- The Guardian
‘Excellent size': UK blueberry crop up nearly a quarter after warm spring
British blueberries are the latest fruit to benefit from the warmest spring on record, with the harvest up by almost a quarter so far this year. Growers say the weather has produced an early crop with more and larger berries, while new varieties can bring higher yields and better resilience. About 5,133 tonnes are expected by the end of August, up significantly from almost 4,187 tonnes by the same point last year. Sales of the spherical fruit are already 9% ahead year on year as growing awareness of the health properties and a healthy crop bolster demand. Daniel Martin, the group commercial director at S&A Produce, says: 'This blueberry season has seen a really positive early start, about two weeks ahead of schedule, thanks to strong light levels and an absence of rain. 'As a result, we're seeing excellent fruit size and nice, even ripening across the crop. In summary, we are optimistic about the season and expect to see excellent flavour and shelf life from this year's crop.' The harvest is also being aided by new technology as growers invest in automated picking and other methods to reduce labour costs and improve efficiency. Jim Floor, the managing director of Hall Hunter, one of the UK's leading blueberry growers, said the company now used machines to harvest 10% of its crop, and this was expected to rise to 30% next year – with the rest picked by hand. Blueberries are leading the technology push, with the more delicate strawberries, raspberries and blackberries now all picked by hand at Hall Hunter. 'When it comes to robotic fruit picking, we're seeing steady and encouraging progress from robotics and automation companies. However, the technology isn't quite ready for widespread commercial adoption just yet,' Floor said. The blockbuster blueberry summer comes hot on the heels of a perfect spring for strawberries. The sunny, very dry spring, with the warmest start to May on record, led to a glut of early strawberries, aubergines and tomatoes at the start of May. Sign up to Business Today Get set for the working day – we'll point you to all the business news and analysis you need every morning after newsletter promotion Growers said they were producing 'giant' 50g strawberries you 'cannot fit in your mouth'. However, the average was a more modest 30g. Nick Marston, the chair of the industry body British Berry Growers, said that all UK berry crops, including strawberries, raspberries and blackberries, were up by about 25% so far thanks to 'cracking weather' in June that had helped ripen crops that had developed well during the cooler start to the spring. 'The weather has been very beneficial for all berry crops,' he said. However, Marston said it was not clear if the overall berry harvest for the year would be up by that amount as some fruit would have just have been harvested earlier than usual and picking might tail off depending on conditions.
