Artificial intelligence

NVIDIA platform lifts leading MoE models

Frontier developers are adopting a mixture-of-experts architecture as the foundation for their most advanced open-source models. Designers now rely on specialised experts that activate only when needed instead of forcing every parameter to work on each token.

Major models, such as DeepSeek-R1, Kimi K2 Thinking, and Mistral Large 3, rise to the top of the Artificial Analysis leaderboard by utilising this pattern to combine greater capability with lower computational strain.

Scaling the architecture has always been the main obstacle. Expert parallelism requires high-speed memory access and near-instant communication between multiple GPUs, yet traditional systems often create bottlenecks that slow down training and inference.

NVIDIA has shifted toward extreme hardware and software codesign to remove those constraints.

The GB200 NVL72 rack-scale system links seventy-two Blackwell GPUs via fast shared memory and a dense NVLink fabric, enabling experts to exchange information rapidly, rather than relying on slower network layers.

Model developers report significant improvements once they deploy MoE designs on NVL72. Performance leaps of up to ten times have been recorded for frontier systems, improving latency, energy efficiency and the overall cost of running large-scale inference.

Cloud providers integrate the platform to support customers in building agentic workflows and multimodal systems that route tasks between specialised components, rather than duplicating full models for each purpose.

Industry adoption signals a shift toward a future where efficiency and intelligence evolve together. MoE has become the preferred architecture for state-of-the-art reasoning, and NVL72 offers a practical route for enterprises seeking predictable performance gains.

NVIDIA positions its roadmap, including the forthcoming Vera Rubin architecture, as the next step in expanding the scale and capability of frontier AI.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

AWS launches frontier agents to boost software development

AWS has launched frontier agents, autonomous AI tools that extend software development teams. The first three – Kiro, AWS Security Agent, and AWS DevOps Agent – enhance development, security, and operations while working independently for extended periods.

Kiro functions as a virtual developer, maintaining context, learning from feedback, and managing tasks across multiple repositories. AWS Security Agent automates code reviews, penetration testing, and enforces organisational security standards.

AWS DevOps Agent identifies root causes of incidents, reduces alerts, and provides proactive recommendations to improve system reliability.

These agents operate autonomously, scale across multiple tasks, and free teams from repetitive work, allowing focus on high-priority projects. Early users, including SmugMug and Commonwealth Bank of Australia, report quicker development, stronger security, and more efficient operations.

By integrating frontier agents into the software development lifecycle, AWS is shifting AI from task assistance to completing complex projects independently, marking a significant step forward in what AI can achieve for development teams.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot! 

Honolulu in the US pushes for transparency in government AI use

Growing pressure from Honolulu residents in the US is prompting city leaders to consider stricter safeguards surrounding the use of AI. Calls for greater transparency have intensified as AI has quietly become part of everyday government operations.

Several city departments already rely on automated systems for tasks such as building-plan screening, customer service support and internal administrative work. Advocates now want voters to decide whether the charter should require a public registry of AI tools, human appeal rights and routine audits.

Concerns have deepened after the police department began testing AI-assisted report-writing software without broad consultation. Supporters of reform argue that stronger oversight is crucial to maintain public trust, especially if AI starts influencing high-stakes decisions that impact residents’ lives.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Amazon rolls out Trainium3 AI chip to challenge Nvidia’s dominance

AWS has launched its in-house AI processor, Trainium3, marking a fresh push to compete with established players in the AI-hardware market. The chip and its associated UltraServer platform were unveiled at the launch event in Las Vegas.

According to Amazon, servers powered by Trainium3 deliver more than four times the performance of the previous generation while using around 40% less energy. Several AI firms, including startups working on large language models, are already utilising the new hardware to reduce their inference or training costs.

Looking ahead, AWS has signalled plans for a follow-up chip, Trainium4, which is expected to integrate with Nvidia’s NVLink Fusion interconnect technology. That would permit hybrid deployments combining Amazon’s ASICs with traditional GPUs, potentially appealing to AI workloads already built around Nvidia’s ecosystem.

The move highlights a broader trend: major tech firms are increasingly investing in their own AI infrastructure, aiming to reduce dependence on dominant vendors and lower costs. As AWS scales out its custom chips, the AI infrastructure market is poised to become more diverse with price-performance and energy efficiency as key differentiators, rather than raw hardware dominance alone.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot! 

UK ministers advance energy plans for AI expansion

The final AI Energy Council meeting of 2025 took place in London, led by AI Minister Kanishka Narayan alongside energy ministers Lord Vallance and Michael Shanks.

Regulators and industry representatives reviewed how the UK can expedite grid connections and support the necessary infrastructure for expanding AI activity nationwide.

Council members examined progress on government measures intended to accelerate connections for AI data centres. Plans include support for AI Growth Zones, with discounted electricity available for sites able to draw on excess capacity, which is expected to reduce pressure in the broader network.

Ministers underlined AI’s role in national economic ambitions, noting recent announcements of new AI Growth Zones in North East England and in North and South Wales.

They also discussed how forthcoming reforms are expected to help deliver AI-related infrastructure by easing access to grid capacity.

The meeting concluded with a focus on long-term energy needs for AI development. Participants explored ways to unlock additional capacity and considered innovative options for power generation, including self-build solutions.

The council will reconvene in early 2026 to continue work on sustainable approaches for future AI infrastructure.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

OpenAI faced questions after ChatGPT surfaced app prompts for paid users

ChatGPT users complained after the system surfaced an unexpected Peloton suggestion during an unrelated conversation. The prompt appeared for a Pro Plan subscriber and triggered questions about ad-like behaviour. Many asked why paid chats were showing promotional-style links.

OpenAI said the prompt was part of early app-discovery tests, not advertising. Staff acknowledged that the suggestion was irrelevant to the query. They said the system is still being adjusted to avoid confusing or misplaced prompts.

Users reported other recommendations, including music apps that contradicted their stated preferences. The lack of an option to turn off these suggestions fuelled irritation. Paid subscribers warned that such prompts undermine the service’s reliability.

OpenAI described the feature as a step toward integrating apps directly into conversations. The aim is to surface tools when genuinely helpful. Early trials, however, have demonstrated gaps between intended relevance and actual outcomes.

The tests remain limited to selected regions and are not active in parts of Europe. Critics argue intrusive prompts risk pushing users to competitors. OpenAI said refinements will continue to ensure suggestions feel helpful, not promotional.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Regulators question transparency after Mixpanel data leak

Mixpanel is facing criticism after disclosing a security incident with minimal detail, providing only a brief note before the US Thanksgiving weekend. Analysts say the timing and lack of clarity set a poor example for transparency in breach reporting.

OpenAI later confirmed its own exposure, stating that analytics data linked to developer activity had been obtained from Mixpanel’s systems. It stressed that ChatGPT users were not affected and that it had halted its use of the service following the incident.

OpenAI said the stolen information included names, email addresses, coarse location data and browser details, raising concerns about phishing risks. It noted that no advertising identifiers were involved, limiting broader cross-platform tracking.

Security experts say the breach highlights long-standing concerns about analytics companies that collect detailed behavioural and device data across thousands of apps. Mixpanel’s session-replay tools can be sensitive, as they can inadvertently capture private information.

Regulators argue the case shows why analytics providers have become prime targets for attackers. They say that more transparent disclosure from Mixpanel is needed to assess the scale of exposure and the potential impact on companies and end-users.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

OpenAI expands investment in mental health safety research

Yesterday, OpenAI launched a new grant programme to support external research on the connection between AI and mental health.

An initiative that aims to expand independent inquiry into how people express distress, how AI interprets complex emotional signals and how different cultures shape the language used to discuss sensitive experiences.

OpenAI also hopes that broader participation will strengthen collective understanding, rather than keeping progress confined to internal studies.

The programme encourages interdisciplinary work that brings together technical specialists, mental health professionals and people with lived experience. OpenAI is seeking proposals that can offer clear outputs, such as datasets, evaluation methods, or practical insights, that improve safety and guidance.

Researchers may focus on patterns of distress in specific communities, the influence of slang and vernacular, or the challenges that appear when mental health symptoms manifest in ways that current systems fail to recognise.

The grants also aim to expand knowledge of how providers use AI within care settings, including where tools are practical, where limitations appear and where risks emerge for users.

Additional areas of interest include how young people respond to different tones or styles, how grief is expressed in language and how visual cues linked to body image concerns can be interpreted responsibly.

OpenAI emphasises that better evaluation frameworks, ethical datasets and annotated examples can support safer development across the field.

Applications are open until 19 December, with decisions expected by mid-January. The programme forms part of OpenAI’s broader effort to invest in well-being and safety research, offering financial support to independent teams working across diverse cultural and linguistic contexts.

The company argues that expanding evidence and perspectives will contribute to a more secure and supportive environment for future AI systems.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

AI growth threatens millions of jobs across Asia

UN economists warned millions of jobs in Asia could be at risk as AI widens the gap between digitally advanced nations and those lacking basic access and skills. The report compared the AI revolution to 19th-century industrialisation, which created a wealthy few and left many behind.

Women and young adults face the most significant threat from AI in the workplace, while the benefits in health, education, and income are unevenly distributed.

Countries such as China, Singapore, and South Korea have invested heavily in AI and reaped significant benefits. Still, entry-level workers in many South Asian nations remain highly vulnerable to automation and technological advancements.

The UN Development Programme urged governments to consider ethical deployment and inclusivity when implementing AI. Countries such as Cambodia, Papua New Guinea, and Vietnam are focusing on developing simple digital tools to help health workers and farmers who lack reliable internet access.

AI could generate nearly $1 trillion in economic gains across Asia over the next decade, boosting regional GDP growth by about two percentage points. Income disparities mean AI benefits remain concentrated in wealthy countries, leaving poorer nations at a disadvantage.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot! 

Gemini Projects feature appears in Google app teardown

Google is preparing a Gemini feature called Projects, offering a small workspace for grouping chats by topic. Early clues show it works like a sandbox that keeps related conversations structured. It is still hidden and not active for anyone.

An Android Authority teardown of the Google app revealed the interface and onboarding prompts. These mention isolating chats, choosing a focus area and adding files for context. The feature remains dormant until Google enables it on the server.

When opening a project, users can name it and then view a simple dashboard. This includes options to set project goals that guide Gemini’s behaviour. The aim is to keep longer tasks organised in one place.

The teardown shows a limit of ten file uploads per project, with no clarity on whether paid tiers will receive more. This may affect complex tasks that require a larger context. Users will also be able to pin projects for quicker access.

Because all information comes from hidden code, Google has not confirmed any details. The design or limits may change before launch. Until the Gemini feature is announced, the findings should be treated as provisional.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!