Sorry to break it to you, but I think we have become a little too obsessed with keeping humans in the loop. Every discussion about autonomous AI eventually reaches the reassuring conclusion that humans will always supervise the important decisions. It sounds responsible, but I suspect we use human supervision to compensate for systems not properly designed for autonomy. My research in AI suggests that humans should increasingly sit at governance, defining objectives, permissions, risks, and operating limits. They should intervene when decisions genuinely exceed the authority already delegated. But requiring human approval during routine execution makes little technical sense, and in customer care and sales, I would argue it can actually make the experience worse.

Beijing is steadily reducing its dependence on foreign suppliers across the broader chip supply chain. That progress is uneven, and some critical gaps will remain for years. Still, the direction is clear. China is building domestic capabilities in the equipment, materials, components and manufacturing processes needed to support a far more self-sufficient semiconductor ecosystem.  Missing that key point could become a major strategic setback for the Western tech ecosystem Our review at Sinolytics examined 20 enabling technologies across the semiconductor supply chain. Three years ago, roughly 60 percent of those technologies were bottlenecks, meaning China lacked substantial domestic capabilities. Today, that share has fallen to about 40 percent. If current trends continue, it could decline to roughly 15 percent within the next three years. The figures are estimates rather than guarantees, but they show how quickly the industrial base is changing.

Dix laboratoires internationaux s'associent pour tenter de modéliser le fonctionnement du cerveau humain grâce à l'intelligence artificielle. Financé à hauteur de 15 millions de dollars par Meta, ce vaste projet prévoit de collecter 15 000 heures d'activité cérébrale, avec l'espoir de faire progresser le diagnostic et le traitement des maladies neurologiques.

Japan is fundamentally reshaping its space strategy by transitioning from government-led research missions to building a commercial space economy. The cornerstone of this transformation is the ¥1 trillion Space Strategy Fund, a ten-year programme established in 2024 that positions JAXA as a major technology-funding agency supporting private companies, start-ups and research institutions. The initiative aims to expand Japan's domestic space market from approximately ¥4 trillion to ¥8 trillion by the early 2030s. Rather than financing isolated projects, the programme cultivates an integrated industrial ecosystem spanning launch systems, satellite constellations, communications, Earth observation and advanced technologies. Strategic investments include small synthetic-aperture-radar satellites, optical communications systems and high-resolution Earth observation platforms. The economic value extends beyond hardware to encompass a broader data economy, with satellites generating commercial applications in agriculture, logistics, insurance, disaster management and infrastructure monitoring. Japan is already demonstrating practical success, with the Japan Meteorological Agency operationalising data from the IBUKI-GW satellite in September 2026 for commercial applications including retail demand forecasting. By developing competitive satellite platforms and monetising the services and information they generate, Japanese companies position themselves to capture significant international market opportunities.

Axelera AI, a European leader in AI acceleration hardware, announced the launch of its Europa architecture, a next-generation AI Processing Unit designed to extend edge-computing advantages to demanding enterprise and data center workloads. The architecture enables organizations to run production AI on controlled infrastructure while addressing rising cloud costs and power constraints. Europa delivers up to 6x more tokens per second per watt than GPU-based solutions, offering superior energy efficiency for inference workloads, which represent the bulk of AI compute demands. The company has established partnerships with Dell and Supermicro for validated systems and maintains a sales pipeline exceeding $1.5 billion, with deployments across more than 600 global customers in defense, robotics, drones, retail, and security sectors. Europa particularly addresses Physical AI applications and regulated industries including financial services, healthcare, and government, where data sovereignty and compliance requirements necessitate on-premises deployment. Dell Technologies collaborates with Axelera and systems integrator E4 to support EU AI Factories in Italy and Luxembourg, emphasizing how the solution provides performance, control, and sovereignty while reducing infrastructure costs and operational constraints inherent to cloud-dependent AI deployment models.

History demonstrates that nations adopting military technologies rapidly gain strategic advantages, as evidenced by drone deployment in Russia's Ukraine conflict. Artificial intelligence represents the next transformative transition, yet Europe risks falling behind despite its strategic autonomy ambitions. A new category of AI-focused defence firms, termed "neo-primes," has emerged to provide services in intelligence, targeting, command and control, and logistics—contrasting with traditional large defence contractors. The United States leads this sector with successful companies like Palantir and Anduril, creating European security concerns. The Dutch State Secretary for Defence has called for a Palantir alternative, while UK parliamentary reports identify foreign technology dependence as strategically unacceptable. Restrictions imposed in June 2026 on foreign access to advanced US AI models, including Anthropic's Claude Mythos 5 and Fable 5, highlighted how frontier AI capabilities face US regulatory control, constituting potential vulnerability for Europe. The Draghi Report identified the technology sector's role in Europe's productivity gap with China and the US. Technological sovereignty now requires developing indigenous neo-primes capable of scaling into credible armed forces partners. Europe must establish an independent AI technological foundation across the three-layer stack: infrastructure and data, foundation and specialized models, and applications.

The European Union has launched a comprehensive Tech Sovereignty Package designed to strengthen European technological independence and competitiveness. The package encompasses two legislative proposals—the Chips Act 2.0 and the Cloud and AI Development Act (CADA)—alongside two strategic documents addressing open-source development and AI implementation in the energy sector. Currently under review by the European Parliament and Council, final approval is not anticipated before late 2027. Building on the original 2023 Chips Act, which mobilized over €52 billion in investment, Chips Act 2.0 seeks to further incentivize private investment by streamlining permitting procedures and improving access to energy, land, and financing. The CADA proposal tackles barriers to scaling European cloud and computing capacity through three strategic pillars, implementing reciprocity requirements for non-EU providers participating in EU procurement. Recognizing that data centers consume 2.5 percent of EU electricity—reaching 20 percent in major hubs like Ireland—the Strategic Roadmap for Digitalisation and AI in Energy aims to deploy sovereign AI solutions for critical decarbonization objectives, including electricity grid optimization. This initiative reflects the EU's commitment to reducing dependency on external technology providers while maintaining leadership in semiconductor manufacturing, exemplified by companies like ASML and research centers such as Belgium's IMEC.

Every bathroom visit, decoded automatically - no wearable, no manual logging, no routine to change.

Entre les startups et les grands groupes, c’est une histoire d’amour qui peine à se concrétiser. Les deux camps se tournent autour, discutent, mais ça finit encore trop rarement en union. Pourtant, les jeunes pousses ne cessent de clamer haut et fort qu’elles ont besoin des grandes entreprises pour grandir, tandis que ces dernières assurent voir dans les startups un réservoir d’innovations pour accélérer leur transformation numérique. Si les intérêts semblent alignés sur le papier, la réalité est hélas plus nuancée.

The packaging industry faces a critical transition toward sustainable materials as fossil-derived carbon currently accounts for approximately 85% of organic chemicals and materials, according to the Nova Institute. Ward Heij of Colesco Circular & Climate Credit Impact Fund discusses how renewable carbon—sourced from recycled waste streams, captured CO2, and biogenic biomass—offers viable alternatives to reduce fossil dependency. While biobased packaging solutions already exist across polymers, fibers, and coatings and are technically proven, the real challenge lies in scaling these innovations commercially. Successful deployment requires three essential elements: long-term customer demand certainty, robust regulatory frameworks, and accessible financing mechanisms. Many companies fail during the transition from demonstration to commercial deployment, despite having viable technology. By shifting carbon sources through biological cycles alongside technical improvements, the packaging industry can meaningfully advance toward circular economy principles and reduce reliance on fossil resources, creating significant business opportunities for companies that can navigate these commercial hurdles.

Energy communities (ECs) represent citizen-driven initiatives that advance clean energy transitions while enhancing energy security across Europe. A peer-reviewed study examining EC perspectives in northern Sweden employs the Multi-Level Perspective framework to assess implementation challenges and opportunities in this region. The analysis identifies significant barriers including limited national regulation specific to ECs, complex concession and taxation frameworks, lower electricity prices, extended heating seasons, and reduced solar irradiation that compromise the economic viability of solar photovoltaic installations. These factors particularly constrain EC development in high-latitude regions. However, recent EU directives, including the Clean Energy for All Europeans Package (2019) and directive 2018/2001/EU, which established Renewable Energy Communities and Citizens Energy Communities, create a supportive landscape for expansion. The research concludes that while ECs offer flexibility and contribute to climate and energy security objectives, their broader adoption in high-latitude regions requires clearer regulatory frameworks, targeted financial incentives, adapted business models, robust governance structures, and enhanced institutional support to overcome current implementation obstacles.

A specialized chemical manufacturing plant utilizing palm oil and palm kernel oil feedstocks faced severe operational challenges despite access to abundant raw materials in Southeast Asia. A turnaround team was deployed to stabilize cash flow and achieve profitability within a three-year timeframe, yet ultimately failed to meet this target, resulting in the facility entering court-supervised judicial management proceedings in mid-2026. Critically, the plant's difficulties were not attributable to feedstock scarcity—both palm oil and palm kernel oil remain plentiful and readily available throughout the region. Rather, the operational collapse stemmed from near-term structural, financial, and logistical bottlenecks that proved insurmountable within the compressed timeline. This case underscores that while the broader transition toward pure bioprocessing remains decades away, plant-based chemical production currently faces significant practical challenges beyond raw material availability, highlighting the complex infrastructure and operational requirements necessary for viable bio-based chemical manufacturing.

This systematic review compiles the current approaches for the development of versatile delivery systems for small drug molecules. In addition, the review identifies the key challenges that need to be addressed along the translational pathway from fundamental and applied research to clinical application. Simplifying the design of multifunctional systems to facilitate their manufacturing, integrating machine learning algorithms to predict their properties and optimize system composition, and developing standardized protocols for preclinical evaluation using animal models represent key promising directions for future development in this area. The knowledge covered in this review may be useful for researchers in planning future studies and selecting appropriate methods and approaches for the development and characterization of polymer-based delivery systems.

Venture capitalist Balaji Srinivasan announced a signed memorandum of understanding with Kazakhstan in July 2026 to relocate his Network School, a "frontier community for techno-optimists" incubator established in 2024, from Forest City, Malaysia to Astana. The move followed disputes with Malaysian authorities regarding allegations of housing Israeli passport holders and licensing breaches at the hotel-based operation. While investigations did not confirm passport violations, minor technicalities prompted the school's closure. Srinivasan responded by pivoting to Kazakhstan, characterizing it as an undervalued nation offering Singapore and Dubai-level competency with affordable land and energy resources. A photograph documented Srinivasan meeting with Zhaslan Madiyev, Kazakhstan's Minister of Artificial Intelligence and Digital Development, as the school's signage was reinstalled with state sponsorship. The development reflects Srinivasan's evolution from his roles as general partner at Andreessen Horowitz and CTO of Coinbase to his current position as a global angel investor with expanding international projects, positioning him as an influential figure promoting techno-optimist entrepreneurial ventures across geopolitical boundaries.

This study examines how coordinated deployment of public data openness and supercomputing infrastructure strengthens agricultural supply chain resilience in China. Using provincial panel data from 31 Chinese provinces spanning 2011–2024, researchers exploited the staggered rollout of public data platforms and supercomputing centers as quasi-natural experiments. They constructed a synergistic difference-in-differences model, validated through event-study parallel trend tests, placebo simulations, propensity score matching, and double/debiased machine learning. Results demonstrate that policy coordination substantially enhances agricultural supply chain resilience through two primary mechanisms: industrial structure rationalization and agricultural electrification enhancement. The synergistic approach reduces informational barriers between agricultural and related industries, optimizes factor reallocation, and accelerates adoption of precision agriculture technologies. Heterogeneity analysis reveals policy effects are substantially stronger in central and western regions, suggesting digital infrastructure offsets traditional regional disparities. This research is significant as it clarifies how coordinated digital transformation policies enhance food security and agricultural development resilience amid escalating external shocks including climate extremes, market volatility, and geopolitical uncertainties.

Anthropic's Threat Intelligence team has released a comprehensive report documenting the detection and disruption of malicious Claude AI usage between December 2025 and August 2026. The report reveals that threat actors—including state-sponsored groups, financially motivated criminals, commercial spyware vendors, and politically motivated individuals—attempted to exploit Claude's Haiku, Sonnet, and Opus models across seven harm categories: cyber operations, influence operations, surveillance, scams and fraud, biological misuse, conventional weapons development, and distillation. Notable cases included fake dating app networks designed for fraud and surveillance systems targeting dissidents. No misuse involved the Fable or Mythos-class models, except for one distillation case. Each incident was disrupted, and learnings were incorporated into strengthened safeguards while intelligence was shared with authorities and industry partners. Anthropic emphasizes that as AI capabilities advance, proactive disclosure and cross-industry coordination become essential for identifying threat patterns, informing policy responses, and collectively strengthening defenses against emerging malicious applications.

A 20-second voice recording can already reveal something about your risk of type 2 diabetes. London startup Thymia trained a model using speech from more than 21,000 people, validated it on another 7,319, and then compared its predictions with HbA1c blood tests in 801 participants. In that blood-tested group, the model gave a person with diabetes a higher risk score than someone without it about 75% of the time, and at one screening threshold it flagged 82% of diabetes cases. It is still a preprint and nowhere near replacing a blood test, but the remarkable part is that a metabolic disease leaves enough information inside 20 seconds of ordinary speech for an AI model to pick it up.

Data monetization is a CFO priority, because it can diversify revenue streams, expand margins and multiples—and change your capital allocation priorities. Yet, for both analysis and execution, it often needs a new toolkit. Compared to other products, data usually has higher upfront fixed costs, fewer variable costs, and longer-lived revenue streams. It can also pose new regulatory, reputational, and governance risks that CFOs have to manage.

The U.S. Department of Energy is partnering with nuclear power plants to produce clean hydrogen at scale, addressing a critical gap in American energy production. Currently, approximately 95% of U.S. hydrogen comes from natural gas through steam-methane reforming, a carbon-intensive process. The DOE is investing billions to leverage existing nuclear assets for clean hydrogen generation through electrolysis, which splits water into pure hydrogen and oxygen using heat and electricity. High-temperature electrolyzers prove particularly efficient for this application. A single 1,000-megawatt nuclear plant could produce up to 150,000 tons of hydrogen annually, enabling new markets in fertilizer production, oil refining, steel manufacturing, and fuel cell vehicles. The DOE's Office of Critical Minerals and Energy Innovation and Office of Nuclear Energy have initiated three demonstration projects. Notably, Constellation constructed a low-temperature electrolysis system at Nine Mile Point nuclear plant, marking America's first nuclear-powered clean hydrogen production facility. The project began generating hydrogen in February 2023 and will supply additional grid power by 2025. This initiative represents a strategic approach to decarbonizing multiple industrial sectors while revitalizing nuclear facilities' economic viability.

Inhaled therapeutics are emerging as a promising strategy to combat antimicrobial resistance by delivering medicines directly to the lungs, potentially improving treatment efficacy for respiratory infections while minimizing systemic exposure. According to a 2024 Lancet analysis, bacterial antimicrobial resistance caused an estimated 4.71 million deaths globally in 2021, with 1.14 million directly attributable to AMR, highlighting the urgency of addressing this public health crisis. While developing new antibiotics remains essential, researchers are increasingly investigating whether optimizing delivery methods for existing therapies could enhance effectiveness. Pulmonary drug delivery advances in antibiotics, vaccines, and biologics are being driven by innovations in particle engineering, aerosol science, and inhaler technologies. Unlike traditional oral or intravenous administration, which distributes medicine systemically before reaching infected lung tissue, targeted inhaled delivery places therapeutic agents precisely where needed. This approach requires specialized expertise in formulation design, aerosol characterization, device selection, and manufacturing to ensure optimal lung deposition. By refocusing on delivery optimization alongside antibiotic discovery, the global health community can better utilize existing treatment arsenals while combating the growing threat of antimicrobial resistance.