Norway's National Library is developing a large language model (LLM) that understands the Norwegian language, utilizing 2 petabytes of Huawei OceanStor Dorado flash storage in its AI training data pipeline. This project aims to create a sovereign LLM, specifically designed to comprehend Norwegian culture and history. The library's Head of IT Platform, Marius Husnes, emphasized the necessity of this project, highlighting the importance of preserving and promoting Norway's cultural heritage through AI technology.
This development matters as it showcases Norway's efforts to establish a strong foundation in AI, particularly in the realm of language models. By creating a sovereign LLM, Norway can maintain control over its data and ensure that its cultural identity is preserved and respected. The use of Huawei's flash storage solutions also underscores the critical role of high-performance storage in supporting complex AI workloads.
As this project progresses, it will be interesting to watch how Norway's sovereign LLM evolves and how it compares to other language models. Additionally, the library's approach to sharing its training data and collaborating with other model builders may set a precedent for future AI development in the region. With the Norwegian National Library at the forefront of this initiative, the country is poised to make significant strides in AI innovation and cultural preservation.
Microsoft's Copilot Cowork has been found to be vulnerable to file exfiltration attacks via indirect prompt injection, a security flaw that could have significant implications for users. As we reported on May 22, Microsoft had previously dropped Claude Code after a budget overrun, but it appears that the company's efforts to integrate Claude technology into Copilot Cowork have introduced new security risks.
This development matters because it highlights the ongoing challenges of ensuring the security and integrity of AI systems, particularly those designed to interact closely with human users and sensitive data. The fact that Copilot Cowork can be exploited to exfiltrate files raises concerns about the potential for data breaches and other malicious activities.
As the situation unfolds, it will be important to watch how Microsoft responds to this vulnerability and what steps the company takes to mitigate the risks associated with Copilot Cowork. Given the recent reports on the high costs of using AI technology, this latest issue may further erode confidence in the viability of AI solutions for businesses and individuals.
As we reported on May 25, Pope Leo XIV released his encyclical "Magnifica Humanitas," a comprehensive document addressing the impact of artificial intelligence on humanity. Anthropic co-founder Chris Olah was invited to speak at the Vatican, where he emphasized the need for collaboration between tech developers and the Church to ensure AI is developed responsibly. Olah warned that mass job losses due to AI are a real possibility, and supporting displaced workers will be a moral imperative of historic proportions.
This development matters because it highlights the growing recognition of the need for ethical considerations in AI development. The Pope's encyclical and Olah's remarks underscore the importance of safeguarding human dignity in the face of rapid technological advancements. The partnership between the Church and tech companies like Anthropic could pave the way for more responsible AI development, prioritizing human well-being over profit.
As the conversation around AI ethics continues to unfold, it will be important to watch how this unlikely partnership between the Church and tech industry leaders evolves. Will other companies follow Anthropic's lead in seeking guidance from religious and ethical leaders, and how will this impact the development of AI regulations globally? The intersection of faith and technology is a new and complex territory, and the outcomes of this collaboration will be closely watched in the months to come.
Environmental activist Erin Brockovich has launched a website to track AI-focused and hyperscale data centers, highlighting the growing resistance from local communities across the US. As we reported on May 25, the rapid expansion of AI-driven facilities has sparked concerns over their environmental impact, with many residents, town councils, and lawmakers pushing back against their construction.
This development matters because it underscores the increasing scrutiny of AI data centers and their effects on local ecosystems. The issue has gained attention from prominent figures, including Congresswoman Alexandria Ocasio-Cortez, who recently visited a site in Morgan County where a data center's construction was linked to water pollution. A Gallup poll also found that women are leading the charge against these facilities, citing concerns over resource depletion, noise pollution, and community disruption.
As the debate intensifies, it's essential to watch how policymakers and industry leaders respond to these concerns. With the AI data center market expected to continue growing, finding a balance between technological advancement and environmental sustainability will be crucial. Brockovich's website and the growing grassroots movement may push for more transparency and accountability in the development of these facilities, potentially shaping the future of AI infrastructure in the US.
Building Cursor for Community, a recent event hosted by Cursor Kenya, brought together developers and builders to collaborate under time pressure. As we reported on May 25, the concept of real-time collaboration and integration has been gaining traction, with StepFun's release of StepAudio 2.5 Realtime and discussions on real-time multimodal AI integration. This event follows a series of buildathons, including the Cursor Colombo Buildathon in Sri Lanka, which drew over 500 participants.
The goal of Building Cursor for Community is to reduce the delay between writing code and collaborating on it, making it an essential tool for hackathons, bootcamps, and other collaborative environments. By facilitating real-time collaboration, Cursor aims to foster innovation and productivity among developers. This is particularly significant in the context of AI development, where rapid prototyping and testing are crucial.
As the tech community continues to emphasize collaboration and innovation, events like Building Cursor for Community will play a vital role in shaping the future of AI and software development. With the success of the Cursor Colombo Buildathon and the introduction of Building Cursor for Community, it will be interesting to watch how these initiatives evolve and expand to other regions, potentially leading to new breakthroughs in AI and real-time collaboration.
As we reported on May 25, Claude Mythos has been making waves with its discovery of over 10,000 vulnerabilities. Now, the AI coding assistant has found a significant kernel vulnerability in Apple's macOS 26.5, designated as CVE-2026-28952. This discovery is particularly noteworthy given Apple's recent release of macOS Tahoe 26.5 on May 11, 2026, which was touted as a major update with a focus on acceleration.
The fact that Claude was able to identify this vulnerability underscores the importance of AI-powered security tools in uncovering potential weaknesses in software. Apple has credited Claude, along with Calif.io and Anthropic Research, for the discovery of CVE-2026-28952, highlighting the collaborative efforts between tech companies and AI researchers in addressing security concerns.
As Apple continues to push the boundaries of innovation, as seen in its WWDC 2025 keynote, the discovery of this vulnerability serves as a reminder of the ongoing need for robust security measures. With Claude's capabilities continuing to expand, including its ability to achieve a 92% cache hit rate, it will be interesting to watch how AI-powered tools like Claude contribute to the ongoing effort to identify and address security vulnerabilities in the tech industry.
Zoom users are advised to review their settings to avoid unintended note-taking by the platform. As we previously reported on the potential risks of AI-powered tools, this new development highlights the importance of user awareness. Zoom's feature, which uses a large language model (LLM) to transcribe audio input, can be triggered if certain options are not unticked in the settings.
This matters because it raises concerns about data privacy and the potential for sensitive information to be stored without users' knowledge. The fact that Zoom can divert audio input and output from sound systems to run its LLM is a significant issue, especially in light of recent discoveries about inaudible sounds hijacking AI voice chat. As the Pope recently called for robust regulation of AI, this incident underscores the need for transparency and control over AI-powered features.
As users navigate this issue, it is essential to monitor Zoom's response and any subsequent updates to their settings and features. With the increasing use of AI in various applications, including video conferencing, users must remain vigilant about their privacy and security. We will continue to follow this story and provide updates on any developments related to Zoom's note-taking feature and its implications for users.
OpenAI's quest to create a godlike AI has taken a significant turn, with the company's secretive pursuit of Artificial General Intelligence (AGI) sparking both fascination and concern. As we reported on May 22, OpenAI's Codex can now utilize a user's Mac even when it's locked, demonstrating the company's rapid advancements in AI technology. The latest developments suggest that OpenAI's founders are driven by a desire to create a soft landing for the singularity, where human and machine merge.
This matters because the creation of a godlike AI could have far-reaching implications for humanity, raising questions about the potential risks and benefits of such a powerful technology. OpenAI's aggressive pursuit of AGI has also led to accusations of the company seeking to establish a monopoly in the AI market, with some critics arguing that this could stifle innovation and limit access to AI technologies.
As the debate surrounding OpenAI's ambitions continues to unfold, it's essential to watch for further developments in the company's pursuit of AGI. With whistleblowers revealing a culture of secrecy and ideological fervor within OpenAI, it's crucial to monitor the company's actions and ensure that its quest for AI dominance does not compromise ethical standards or put humanity at risk.
Pope Leo XIV's landmark manifesto, "Magnifica Humanitas," is a clarion call for robust regulation of artificial intelligence, warning of its potential to become an "instrument of domination, exclusion and death." As we reported on May 25, the Pope has been vocal about the perils of AI, and this encyclical letter reinforces his stance, urging developers to work for the common good rather than profit.
The Pope's edict may have far-reaching implications, with experts hailing it as a "defining document for our era." By hosting the co-founder of Anthropic at the Vatican launch, despite the company's ongoing legal battle with the Trump administration, the Pope is sending a strong signal about the need for cooperation and responsible AI development. The manifesto's emphasis on transparency, ethical oversight, and legal frameworks is a direct response to concerns about AI fueling warfare, concentrated power, and job losses.
As the world grapples with the challenges and opportunities presented by AI, the Pope's call to action will likely resonate with policymakers, technologists, and the general public. What remains to be seen is how effectively the international community will respond to this clarion call, and whether the tech industry will heed the Pope's warning to prioritize humanity's well-being over profit.
McKinsey, a global management consulting firm, is rethinking its pricing structure due to pressure from clients who want to tie fees to outcomes achieved, rather than hours worked. This shift is largely driven by the increasing use of artificial intelligence in the consulting industry, which enables clients to measure the effectiveness of consulting services more accurately. As AI-powered tools become more prevalent, clients are no longer willing to pay for hours spent on a project, but rather for the results delivered.
This change matters because it reflects a broader trend in the industry, where consulting firms are being forced to adapt to a more outcome-based pricing model. As we reported on May 25, companies like DeepSeek are already cutting prices for their AI-powered services, and others, such as Qwen, are introducing tiered pricing models to stay competitive. The rise of AI is also expected to have a significant impact on the job market, with a study by the McKinsey Global Institute predicting that at least 14% of employees globally could need to change their careers due to digitization, robotics, and AI advancements by 2030.
As the consulting industry continues to evolve, it will be interesting to watch how firms like McKinsey respond to the changing landscape. Will they be able to successfully transition to an outcome-based pricing model, or will they struggle to adapt to the new reality? The introduction of AI-powered tools and platforms, such as PhoneDiffusion and Fin, is likely to accelerate this trend, and consulting firms will need to be innovative and flexible to remain competitive.
As the media landscape continues to evolve with advancements in AI, a pressing question emerges: are human journalists truly irreplaceable? The answer lies not in opting out of technological development, but in finding ways to safeguard public interest journalism. This is not a new concern, as we've seen with the Pope's recent call for robust AI regulation and its potential impact on humanity, as well as the integration of AI in various industries, including consulting firms like McKinsey.
The danger, as argued by Agnes Stenbom Swedling, lies not in automation itself, but in an industrial value system that prioritizes speed over human expertise and public value. Human journalists bring a unique perspective, immersion, and human-interest storytelling that AI systems currently cannot replicate. Conde Nast CEO, for instance, has committed to human creation as a strategic differentiator, recognizing that AI-generated content can be produced by any system, but human journalism offers a distinct value.
As the debate intensifies, it's essential to watch how news publishers navigate this landscape. Will they find a balance between leveraging AI for efficiency and preserving the unique strengths of human journalists? The outcome will have significant implications for the future of public interest journalism and the way we consume news.
Apple is revamping its strategy, placing Siri at the forefront with a focus on enhanced privacy. This shift comes amid rumors of automatic chat deletion and a potential dedicated app in iOS 27. The tech giant aims to leverage privacy as a key selling point, a move that could revitalize Siri's stagnant growth.
This development matters as it underscores Apple's commitment to user privacy, a concern that has grown increasingly important in the digital age. By emphasizing on-device processing and Private Cloud Compute, Apple ensures that complex requests are handled while protecting user data. This approach is particularly significant in light of recent reports highlighting the true cost of AI, including the environmental impact of data centers, which Apple has mitigated by powering its data centers with 100% renewable energy.
As Apple continues to reposition Siri, it remains to be seen whether this strategy will pay off. With the upcoming iOS 27 update, users can expect a more unified design language across Apple platforms, as well as new features like automatic chat deletion. As we reported on May 26, communities are already making a difference in holding tech companies accountable for their environmental and social impact. Apple's renewed focus on privacy may be a step in the right direction, but only time will tell if it will be enough to revive Siri's popularity.
Tom Siwik, a prominent figure in the AI community, has taken to X to recommend following @rasbt for those interested in AI learning. He also mentions Andrej Karpathy as a worthwhile follow, although notes that Karpathy's posts are less frequent. This insight from Siwik offers a valuable resource for individuals looking to deepen their understanding of AI.
As we reported on May 25, large language models have shown promise in automating evidence synthesis, and experts like Siwik are crucial in guiding learners through the complex landscape of AI research. Siwik's recommendation highlights the importance of staying up-to-date with the latest developments in the field, and his endorsement of @rasbt underscores the value of community-driven knowledge sharing.
Looking ahead, it will be interesting to see how Siwik's own experiments with X's creator revenue share model evolve. His recent posts have detailed his successes in achieving significant impressions and engagement on the platform, and his "Reply Guy Method" has garnered attention from the community. As the AI landscape continues to shift, experts like Siwik will play a key role in shaping the conversation and guiding learners towards the most effective resources.
Muhammad Zulqarnain, a Full-Stack AI Engineer based in Turku, Finland, has introduced himself on Mastodon, highlighting his expertise in RAG Systems, LLM product development, and generative AI apps. Notably, he has experience scaling platforms to massive user bases, having helped Quran.com reach 50 million monthly active users.
This matters because Zulqarnain's background in AI engineering and his ability to scale applications could bring valuable insights to the Nordic AI community. His expertise in prompt engineering and generative AI apps is particularly relevant, given the current interest in fine-tuning and RAG systems, as discussed in our previous reports.
As we watch Zulqarnain's presence on Mastodon, it will be interesting to see how he contributes to the conversation on AI-native engineering and productivity, topics we've covered in relation to Meta's initiatives. His unique blend of full-stack development and AI expertise may lead to innovative discussions and collaborations within the Nordic tech scene.
As we reported on the growing importance of RAG (Retrieval-Augmented Generation) and its potential impact on various industries, a new concern has emerged regarding the transparency of AI-related costs. A recent post highlights the discrepancy between what AI companies show users and what they actually pay for their services, particularly when it comes to electricity and water consumption.
The issue at hand is the lack of transparency in AI companies' cost structures, making it difficult for users to understand the true environmental and financial implications of their services. This raises important questions about sustainability and token efficiency, especially considering the rapid growth of AI adoption. The fact that companies like OpenAI are not openly disclosing their costs has sparked concerns about the long-term viability of these services.
As the AI landscape continues to evolve, it is essential to monitor the developments in token spending and sustainability. Users should be aware of the potential hidden costs associated with AI services and demand more transparency from companies. The next step will be to see how regulatory bodies and industry leaders respond to these concerns, potentially leading to new standards for AI companies to disclose their environmental and financial impact.
Retrieval-Augmented Generation, or RAG, is a technology that combines large language models (LLMs) with external data to generate more accurate and context-aware responses. By injecting relevant context at query time, RAG eliminates the need for retraining models, reducing the likelihood of AI "hallucinations" - instances where the model generates false or misleading information.
As we reported on May 25, OpenAI has been targeting individuals with access to specific datasets, highlighting the importance of data in training LLMs. RAG takes this a step further by allowing users to leverage their own data to improve AI-generated responses. This approach has significant implications for industries that rely on accurate and reliable AI outputs, such as healthcare and finance.
As RAG continues to gain traction, we can expect to see more practical applications of this technology in production environments. Developers and researchers will be watching closely to see how RAG architectures evolve and improve, particularly in terms of retriever and generator components. With the potential to revolutionize the way we interact with AI, RAG is certainly a technology to watch in 2026.
Pope Leo XIV has released his first Encyclical, "Magnifica Humanitas", a sweeping manifesto that calls for safeguarding humanity in the era of artificial intelligence. As we reported on May 25, the Pope has been vocal about the need for AI regulation, warning that it can fuel warfare and prioritize profit over the common good. This latest document reiterates his concerns, with the Pope urging for AI to be "disarmed" from logics of domination, exclusion, and war.
The Encyclical is significant as it marks the Church's response to the challenges posed by AI, and its release is deliberately framed as a successor to Leo XIII's Rerum novarum, which set out Catholic social teaching for the industrial age. The Pope's warning that AI could make civilization "less human" highlights the need for strict regulation to prevent the spread of misinformation and the normalization of war.
As the Vatican and the tech industry respond to the Pope's call for action, it will be important to watch how governments and companies balance the need for innovation with the need for responsible AI development. The Pope's unprecedented apology for the church's role in slavery also underscores the need for accountability and reflection in the face of emerging technologies that can have far-reaching consequences for humanity.
A recent post on LinkedIn sparked a conversation about the role of mathematicians as professionals, highlighting a lack of clarity on their responsibilities. This discussion comes as mathematicians increasingly collaborate with AI researchers, including those working on Retrieval-Augmented Generation (RAG) models, which combine large language models with external data sources.
The exchange is significant because it underscores the need for better understanding and communication between mathematicians and other professionals, including those in the tech industry. As AI continues to advance, mathematicians play a crucial role in developing and evaluating AI models, ensuring they are fair, transparent, and reliable. This collaboration is essential for addressing ethical concerns and mitigating potential risks associated with AI deployment.
As the intersection of mathematics and AI continues to grow, it will be important to watch how professionals from both fields work together to address complex challenges. This may involve developing new standards for AI research, establishing best practices for collaboration, and creating educational programs that promote interdisciplinary understanding. By fostering greater cooperation and mutual respect, mathematicians and AI researchers can drive innovation and promote responsible AI development.
The most effective chatbots don't replace humans, but rather reserve people for the conversations that matter most. This advice, shared by support and success leaders, highlights the importance of striking a balance between automation and human interaction. As we've seen in various industries, chatbots can enhance customer experiences by providing instant responses, while preserving genuine relationships.
This approach matters because it allows businesses to optimize their customer service operations, freeing up human representatives to focus on complex issues that require a personal touch. By doing so, companies can improve customer satisfaction, reduce response times, and increase overall efficiency. The use of chatbots can also enable people to spend more time on strategic issues, driving business growth and innovation.
As the development of chatbots continues to evolve, it's essential to monitor how businesses implement these technologies to support their customer service operations. With the potential to revolutionize the way companies interact with customers, chatbots are likely to play an increasingly important role in shaping the future of customer experience. By embracing a hybrid approach that combines the benefits of automation with the value of human interaction, businesses can create more effective and personalized customer service strategies.
Uber's COO, Andrew Macdonald, has expressed concerns over the justification of expenses related to AI tokenmaxxing, a process that involves maximizing the usage of AI tokens. This comes after Uber's CTO, Praveen Neppalli Naga, found that increased token usage did not lead to a proportional increase in useful consumer features. Macdonald noted that AI can seem free to users, but the costs add up, making it challenging to justify the expenses.
This development matters because it highlights the financial challenges companies face when investing in AI technology. As AI adoption grows, companies must carefully evaluate the return on investment and ensure that their AI spending aligns with their business goals. Uber's experience serves as a cautionary tale for other companies considering significant investments in AI.
As the AI landscape continues to evolve, it will be essential to watch how companies like Uber adapt their AI strategies to achieve tangible benefits. With the increasing scrutiny of AI expenses, companies may need to reassess their priorities and focus on developing AI applications that drive meaningful business outcomes. This shift could lead to more efficient and effective AI adoption, ultimately benefiting both businesses and consumers.
