At the London Tech Week event in early June, Nvidia CEO Jensen Huang praised the UK as the ‘envy of the world’ when it comes to AI researchers, but he also criticised it as the largest AI ecosystem in the world without its own infrastructure. The criticism is somewhat self-serving: when the UK does get around to building out that infrastructure, it’s certain to consist largely of chips sold by Huang’s company. It’s also unsurprising: Huang has been pitching the idea of ‘sovereign AI’ since at least 2023, conscious that nation states are the next deep pockets to target after the hyperscalers and generously funded model builders. In a world where the only real contenders in the race for AI supremacy are the US and China, we look at how the pursuit of AI sovereignty is playing out across the rest of the planet.

Pseudosection modelling of a relict garnet-core in Palaeoproterozoic rocks from the Gridino area in the southern Belomorian belt of Karelia reveals peak-pressure eclogite-facies conditions of 610–650°C, 18–20 kbar for two retro-eclogite samples and 610–665°C, 23–26 kbar for a rare Mg-rich biotite-orthopyroxene eclogite, suggesting low initial metamorphic field gradients of 6.6–10°C/km. This confirms an earlier finding in Karelia and, considering other Palaeoproterozoic eclogite occurrences worldwide, that ‘cold’ subduction conditions, characteristic of modern-style subduction, occurred during the Palaeoproterozoic, ∼2 Ga ago, for the first time in Earth history. However, compositions of most other phases in the retro-eclogite were reset by diffusion, deformation and recrystallisation during subsequent pressure release and heating to variable degrees, a reason for earlier overestimations of temperatures. By contrast, peak-pressure conditions for a biotite paragneiss (640–740°C, 15–18 kbar) that occurs close to the biotite-orthopyroxene eclogite locality already show an early resetting of its initial assemblage. High-pressure granulite-facies peak-temperature conditions of the retro-eclogite at 712± 5°C, 9–12 kbar (along a field gradient of 20°C/km) were determined by Zr-in-rutile thermometry and quartz-in-garnet elastic barometry. These conditions were dated by a Rb/Sr mineral isochron for the biotite-orthopyroxene eclogite at 1830±20 Ma for the first time. Using existing ages for the peak-pressure conditions, possible slow overall exhumation rates of <0.9 mm/y between eclogite and the granulite-facies stages could be determined that are compatible with erosion as the main exhumation mechanism. The peak-temperature conditions were possibly established by thermal relaxation during early exhumation. However, a younger Rb/Sr mineral isochron for the biotite paragneiss indicates a characteristic Sr-isotopic disequilibrium distribution caused by diffusion during slow cooling between ∼1800 and 1750 Ma during later exhumation.

In vitro systems involving microbial fermentation typically require freshly obtained inocula, such as rumen fluid or faeces. The objective of this study was to test whether preserved faeces can be used instead of fresh faeces in the Hohenheim gas test (HGT). Fresh faeces from sheep (control, C) was compared with seven differently preserved faeces by using nine different feeds and studying in vitro gas production (GP) (n = 6–9 per treatment). Preservation involved freezing at −20°C (FR), shock-freezing with liquid nitrogen (N) and additional freezing at −20°C (FRN), FRN followed by defrosting (FRNdef), shock-freezing with liquid N and freeze-drying (FDN), freeze-drying (FD) and freeze-drying with storage for 3 weeks (FD3W) or 6 months (FD6M). Metaproteomics was used to analyse microbiome composition and function in treatments C, FR, FRN, FD, and FDN (n = 3 per treatment). On average across all feeds, the potential GP with FR and FRN (61 mL/200 mg DM) was comparable to that of C (62 mL/200 mg DM), whereas values for FRNdef, FDN, FD, FD3W, and FD6M were 85, 78, 76, 78 and 71% of C, respectively. All estimated GP kinetic parameters were affected by feed and preservation interactions (P<0.001). Microbiomes from C, FR, and FRN differed from those of FD and FDN based on the relative abundance of the core proteins (P<0.001). FD and FDN showed a significant decline of Bacteroidota, functional redundancy values, and specific proteins such as carbohydrate esterases (CE) (P<0.05) and glycoside hydrolases (GH) (P<0.01). Overall, frozen faeces closely resembled fresh faeces and can serve as a viable alternative inoculum source in the HGT. This may reduce animal numbers used for scientific purposes, but preservation and storage must be strictly standardised to maintain an active microbiome for GP-based in vitro tests.

This paper investigates time-varying risk sharing between annuity buyer and provider. It explores Pareto optimal (PO) and viable Pareto optimal (VPO) risk-sharing designs, in which the share of the reserve deviation transferred to the policyholder varies over time. The optimization problem, based on a weighted average of mean-variance preferences, results in a complex quartic objective function. Such optimization problems are difficult to solve, and checking their convexity is known to be NP-hard. A heuristic method is introduced to simplify the problem, providing a closed-form solution that closely approximates the numerical results. The paper also highlights factors influencing the existence of VPO designs, with age playing a critical role, thereby suggesting the suitability of these designs as retirement products.

Translation is key to the political economy of neorural revival in contemporary Italy. Drawing on fieldwork with neorural farmers, I show how translations across semiotic domains and displays of linguistic and pragmatic untranslatability simultaneously produce capitalist value and temporary disruptions of the subsumption of life under capital. To understand this apparent paradox, I analyze the complex relationship between contemporary neorural revivalists and mid-twentieth-century neodialect poets. Driven by a reaction against the post-war encompassment of regional linguistic varieties within a national standard, the metapragmatics of untranslatability developed by the neodialect literary movement has indirectly provided contemporary neoruralists with semiotic resources to conjure profitable forms of agrolinguistic incommensurability. However, unlike the poets’ nostalgic and anticapitalist sabotage of the collusion between centripetal linguistic standardization and intensive agribusiness scalability, the farmers’ interactional disruptions of pragmatic regimentation and seamless intertranslatability are both a project of capitalist valorization and an exit strategy from unfulfilling wage-labor arrangements.

Personalised nutrition (PN) has emerged as an approach to optimise individual health outcomes through more targeted and tailored dietary recommendations based on unique genetic, phenotypic, medical, lifestyle and contextual factors. The application of artificial intelligence (AI) presents an opportunity to achieve personalised nutrition advice at a scale that has population impact. This review introduces a nutrition audience to different AI applications and offers insights into the concepts of AI that might be relevant to the field of nutrition research. The current and future uses of AI in PN are discussed, as well as the potential benefits and challenges to their application. AI-driven solutions have the potential to improve health and reduce the risk of disease because they can consider more information about an individual in making recommendations. However, challenges such as data interoperability, ethical considerations, and model interpretability remain an issue limiting widespread use at this point. This review will provide a foundational understanding of the application of AI within PN and help to identify opportunities to leverage the potential of AI in transforming dietary guidance and enhancing health outcomes through innovative solutions.

International treaties commonly request States to submit periodic reports on measures adopted to facilitate compliance with relevant obligations, permitting them to identify shortcomings and develop appropriate policies, promote transparency and facilitate the exchange of good practices. International humanitarian law (IHL) might appear at odds with this approach as its core instruments do not establish a periodic reporting procedure; indeed, only limited reporting activities have been required from States party to the Geneva Conventions and their Additional Protocols. The present paper challenges this perspective, exploring mandatory periodic national reporting activities provided by other treaties forming part of the IHL framework, as in relation to cultural property and weapons systems, as well as more informal reporting mechanisms on IHL developed outside treaty regimes, including those addressing organized armed groups. Taking stock of existing approaches and practices, the paper identifies relevant trends, opportunities and challenges for IHL reporting activities.

Ion-acoustic waves in a dusty plasma are investigated where it is assumed that the ions follow a Cairns distribution and the electrons are Boltzmann distributed. Two theoretical methods are applied: Sagdeev pseudopotential analysis (SPA) and reductive perturbation theory (RPT). Since SPA incorporates all nonlinearities of the model it is the most accurate but deriving soliton profiles requires numerical integration of Poisson’s equation. By contrast, RPT is a perturbation method which at second order yields the Gardner equation incorporating both the quadratic nonlinearity of the Korteweg–de Vries (KdV) equation and the cubic nonlinearity of the modified KdV equation. For consistency with the perturbation scheme the coefficient of the quadratic term needs to be at least an order of magnitude smaller than the coefficient of the cubic term. Solving the Gardner equation yields an analytic expression of the soliton profile. Selecting an appropriate set of compositional parameters, the soliton solutions obtained from SPA and RPT are analysed and compared.

Investigating risk factors for mpox’s infectious period is vital for preventing this emerging disease, yet evidence remains scarce. This study aimed to identify risk factors associated with the duration of mpox infectiousness among mpox cases in Vietnam. The primary outcome was the duration of the mpox infectiousness, defined between symptom onset and the first negative test result for the mpox virus. Fine and Gray’s regression models were employed to assess the associations between the infectious period and several risk factors while accounting for competing risks of death by mpox. Most mpox cases recovered within 30 days. Patients with HIV or treated at multiple facilities for mpox had lower incidence rates of cleared infection compared to those who were HIV-negative or treated at a single facility. In regression models, patients with mpox symptoms of rash or mucosal lesions (sub-distribution hazard ratios = 0.62, 95% confidence interval = 0.46–0.83), ulcers (0.57, 0.41–0.80), or fever (0.62, 0.46–0.83) had significantly prolonged infectious periods than those without such symptoms. Our findings provided insights for managing mpox cases, especially those vulnerable to prolonged infectious periods in settings with sporadic cases reported.

I consider the conditions for defining a mineral by its dominant end-member formula. One can calculate the end-member proportions of the end-members of a mineral provided that the end-members are linearly independent (i.e. they are phase components of the mineral); the result includes the dominant end-member of the mineral. If the end-members used in this calculation are not linearly independent, the corresponding set of simultaneous equations is indeterminate. One may remove an end-member from the system, removing the linear dependence; however, any end-member formula may be removed, leaving various sets of end-members that function as phase components. Each set of end-members produces a different solution for the end-member proportions. Each set of positive end-member proportions may (or may not) result in a different dominant end-member; however, within the compositional limits of the species, the same end-member is dominant over all others calculated with different combinations of component end-members. Problems previously encountered in attempting to calculate the dominant end-member formula were due to (1) using mineral formulae that do not accord with the requirements of stoichiometry, and (2) using end-members that are not components of the system. Where the set of end-members chosen to relate mineral composition to end-member proportions contains an end-member that is a linear combination of the other end-members, one must calculate the end-member proportions for all distinct subsets of linearly independent end-members. The dominant end-member over all sets of end-member proportions with all proportions positive is the dominant end-member. Thus for any mineral formula, the dominant end-member formula may be identified and serves to uniquely characterize and identify the mineral. The arguments used here are illustrated by reference to the minerals of the garnet supergroup.

High-order harmonic generation (HHG) in noble gases driven by femtosecond lasers is currently a feasible solution to obtain ultrafast pulses in the extreme ultraviolet (EUV) wavelength range. Implementation of high-flux EUV sources requires driving HHG using an ultrafast laser source in the visible wavelength range with MHz repetition rate. In this paper, we employ a multi-pass cell followed by chirped mirrors to compress 1-MHz, 200-W, 300-fs pulses at 1.03 μm to a duration of 35 fs. The resulting 186-W compressed pulses are focused onto 0.5-mm thick beta barium borate crystal to drive second-harmonic generation and produce positively chirped pulses at 520 nm. These green pulses are de-chirped to 26 fs in duration with an average power of 64 W, which, to the best of our knowledge, represents the highest average power of green pulses with a duration below 100 fs.

Exhibition trophies have become invisible to most people reading about and looking at images of the great world’s fairs. This is not surprising; trophies have fallen out of our awareness because they, and the criticisms they provoked, have received surprisingly little scholarly attention. This article reveals not only this largely overlooked form, but also just how much cultural work they were doing and why so many people found them disturbing. Exhibition trophies became a solution to the nineteenth-century design problem of representing progress, imperial power, extractive superabundance, control of the natural world, and industrial capacity. Nineteenth-century exhibitors and collectors made trophies out of a wide array of commodities, animals, raw materials, manufactured goods, weapons, and “primitive” objects. But by carrying with them ancient connotations of high-minded victory and violence, exhibition trophies also inspired criticisms that got to the heart of modern forms of conquest. Divisive in the middle of the nineteenth century, trophies were ubiquitous by the turn of the twentieth. Meanwhile a new, rival way of displaying imperial power emerged that challenged ethnographic trophies in particular: the new science of anthropology. This article begins to recover this lost form and its implications—from disquiet to the acceptance of abundance (even overabundance) as a collective goal.

The questions of whether first language (L1) speakers and second language (L2) learners can both predict what follows based on given linguistic cues and what factors may influence this predictive processing are still underexplored. Prior research has focused on the success or failure of predictions in real-time processing, paying relatively less attention to the speed of prediction. This study addresses these gaps by investigating the role of word co-occurrence frequency and proficiency in L1 and L2 predictive processing, using the Korean classifier system. In a webcam-based visual-world eye-tracking experiment, both L1-Korean speakers and L2-Korean learners showed sound predictive processing, with the frequency of co-occurrence between classifiers and nouns playing a crucial role. Higher co-occurrence frequency expedited predictive processing for L1-Korean speakers and boosted the ability to make online predictions for L2-Korean learners. The study also revealed a proficiency effect, where more advanced L2-Korean learners made predictions regardless of co-occurrence frequency, unlike their less advanced counterparts. Our findings suggest that predictive mechanisms in L1 and L2 operate in a qualitatively similar way. In addition, the use of webcam eye-tracking is expected to create a more inclusive and equitable research environment for (applied) psycholinguistics.

Next-generation X-ray satellite telescopes such as XRISM, NewAthena and Lynx will enable observations of exotic astrophysical sources at unprecedented spectral and spatial resolution. Proper interpretation of these data demands that the accuracy of the models is at least within the uncertainty of the observations. One set of quantities that might not currently meet this requirement is transition energies of various astrophysically relevant ions. Current databases are populated with many untested theoretical calculations. Accurate laboratory benchmarks are required to better understand the coming data. We obtained laboratory spectra of X-ray lines from a silicon plasma at an average spectral resolving power of $\sim$7500 with a spherically bent crystal spectrometer on the Z facility at Sandia National Laboratories. Many of the lines in the data are measured here for the first time. We report measurements of 53 transitions originating from the K-shells of He-like to B-like silicon in the energy range between $\sim$1795 and 1880 eV (6.6–6.9 Å). The lines were identified by qualitative comparison against a full synthetic spectrum calculated with ATOMIC. The average fractional uncertainty (uncertainty/energy) for all reported lines is ${\sim}5.4 \times 10^{-5}$. We compare the measured quantities against transition energies calculated with RATS and FAC as well as those reported in the NIST ASD and XSTAR’s uaDB. Average absolute differences relative to experimentally measured values are 0.20, 0.32, 0.17 and 0.38 eV, respectively. All calculations/databases show good agreement with the experimental values; NIST ASD shows the closest match overall.

Manipulation and deception were not born with AI: online architecture of choice can be harmful when it contains dark patterns or deceptive designs. These techniques deceive or manipulate users through interfaces that have the substantial effect of subverting or altering users’ agency, decision-making, or choice as part of their online activities. But AI has the potential to further enhance this manipulation increase its sophistication and scale. This article presents the principle of ‘Fairness by Design’ as a potential solution as well as a set of interface prototypes inspired by it and developed within Amurabi’s R&D Lab. These solution prototypes are called ‘Fair Patterns’. Fair patterns make it possible to implement the principles of transparency, trust, and autonomy by providing the right level of information at the right moment in the user journey, in clear language and without cognitive overload.

Fanfaniite, Ca4Mn2+Al4(PO4)6(OH)4·12H2O, from the Hühnerkobel pegmatite mine, Bavaria, has been characterised by chemical analyses and synchrotron single-crystal diffraction. The average crystal structure was refined in space group C2/c (cell parameters a = 10.055(2), b = 24.132(5), c = 6.2590(10) Å, β = 91.35(3)°) to compare with reported monoclinic structures of other calcioferrite-group minerals with general formula Ca4AB4(PO4)6(OH)4·12H2O, A = Mn2+, Fe2+, Mg, B = Al, Fe3+. The average structure contains disordered half-occupied A sites and associated coordinated water molecules. The diffraction data for fanfaniite contains weak reflections that violate the c-glide condition, as also reported for montgomeryite, and in addition contains extremely weak, diffuse reflections requiring a doubling of a, as reported for kingsmountite. Structure refinements were conducted for the noncentrosymmetric C2 model used for montgomeryite and for the P$\bar 1$ model used for kingsmountite. The fanfaniite diffraction data is better explained by the triclinic model with doubled a cell parameter, although the extent of ordering of the A-site cations is considerably lower (56%) than reported for kingsmountite (85%). If the C2 model contributes, it can only be at the scale of the unit cell.