Artificial general intelligence (AGI) is an established field of research. Yet some have questioned if the term still has meaning. AGI has been subject to so much hype and speculation it has become something of a Rorschach test. Melanie Mitchell argues the debate will only be settled through long term, scientific investigation. To that end here is a short, accessible and provocative overview of AGI. I compare definitions of intelligence, settling on intelligence in terms of adaptation and AGI as an artificial scientist. Taking my cue from Sutton's Bitter Lesson I describe two foundational tools used to build adaptive systems: search and approximation. I compare pros, cons, hybrids and architectures like o3, AlphaGo, AERA, NARS and Hyperon. I then discuss overall meta-approaches to making systems behave more intelligently. I divide them into scale-maxing, simp-maxing, w-maxing based on the Bitter Lesson, Ockham's and Bennett's Razors. These maximise resources, simplicity of form, and the weakness of constraints on functionality. I discuss examples including AIXI, the free energy principle and The Embiggening of language models. I conclude that though scale-maxed approximation dominates, AGI will be a fusion of tools and meta-approaches. The Embiggening was enabled by improvements in hardware. Now the bottlenecks are sample and energy efficiency.
Trevor Fitzpatrick, Seamus Kelly, Patrick Carey
et al.
The emergence of Generative AI (Gen AI) has motivated an interest in understanding how it could be used to enhance productivity across various tasks. We add to research results for the performance impact of Gen AI on complex knowledge-based tasks in a public sector setting. In a pre-registered experiment, after establishing a baseline level of performance, we find mixed evidence for two types of composite tasks related to document understanding and data analysis. For the Documents task, the treatment group using Gen AI had a 17% improvement in answer quality scores (as judged by human evaluators) and a 34% improvement in task completion time compared to a control group. For the Data task, we find the Gen AI treatment group experienced a 12% reduction in quality scores and no significant difference in mean completion time compared to the control group. These results suggest that the benefits of Gen AI may be task and potentially respondent dependent. We also discuss field notes and lessons learned, as well as supplementary insights from a post-trial survey and feedback workshop with participants.
Rémi Toupin, Geoff Krause, Poppy Nicolette Riddle
et al.
In recent years, ocean governance has called for strategic action and science‐informed policy to work towards the sustainable development of the ocean, most notably as part of the UN Decade of Ocean Science for Sustainable Development (2021–2030). This common framework identifies the integration of scientific knowledge in governance as a key process to deliver solutions responding to the current challenges, opportunities, and transformations posed by global change in the oceans. This article presents a methodological approach for identifying ocean‐related research outputs and documenting research‐based knowledge integration in documents that inform ocean governance. Specifically, this study builds on an analysis of the references included in the UN Second World Ocean Assessment report to (a) identify and describe the research outputs cited in the distinct chapters of the report, (b) identify research outputs relevant to ocean governance through the analysis of citations from and to references included in the UN Second World Ocean Assessment, (c) compare both datasets to examine the position of the literature cited in the report within a broader ecosystem of ocean‐related research, and (d) present a method to identify topically relevant research that could be integrated in future ocean assessments. Our findings show distinct referencing practices across chapters and expert groups and a higher reliance on high‐profile sources in the report compared to a broader dataset of ocean research outputs. Moreover, this study highlights an innovative approach to identifying ocean research based on knowledge syntheses and considers discussion points about integrating research‐based knowledge in documents informing ocean governance.
Taiwo Bintu Ayinde, Charles F. Nicholson, Benjamin Ahmed
Abstract Achieving Net Zero Emissions in vegetable production systems is a critical challenge in dryland climates of low- and middle-income countries, yet limited data exists to assess the feasibility of such systems. This study employs life cycle inventory methods to evaluate key performance metrics, including yield per land area, production costs, cumulative energy demand (CED), global warming potential (GWP), and water use (WU) for Controlled Environment Agriculture (CEA) in screen houses and field-based tomato production systems in Northern Nigeria. The findings reveal that CEA, despite its high production cost of ₦24,070.80 per m², achieves the highest yield of 28.57 kg per m². Additionally, CEA demonstrates superior efficiency, exhibiting the lowest C ED (0.025 MJ/kg) and GWP (0.76 kg CO₂-eq/kg). In contrast, rainfed field production, while having the lowest cost (₦58.45 per m²), results in the lowest yield (0.08 kg/m²) and the highest GWP (34,545.8%). Irrigated field production performs moderately, with a production cost of ₦150.38 per m², a yield of 0.22 kg per m², and a GWP of 12,572.4%. A key factor influencing yield variation across production systems is the difference in tomato varieties cultivated in open-field and CEA environments. CEA relies on hybrid varieties optimized for controlled conditions, whereas open-field farming utilizes varieties adapted to outdoor environmental fluctuations, contributing to disparities in yield potential. This study highlights the trade-offs between cost, yield, energy efficiency, and environmental impact across different production models. The results underscore the advantages of adopting more efficient and controlled cultivation methods like CEA, offering potential pathways for sustainable and environmentally responsible agricultural practices in regions facing climate and resource constraints.
Societal Impact Statement Lack of plant awareness represents a significant phenomenon characterized by the underestimation of plants, with clear implications for sustainability. This study explores the potential of key competencies in education for sustainable development as an effective framework for mitigating this phenomenon. Through conceptual analysis, these competencies emerge as a valuable tool for enhancing plant awareness. This has significant implications for both the educational community and the general public, as it offers an additional pathway for fostering plant awareness, which can ultimately lead to increased public pressure and stronger mobilization by policymakers on critical issues such as biodiversity conservation and climate change. Summary The phenomenon of “plant blindness” or “lack of plant awareness” has received much attention from researchers over the last years. Recognizing education as both a contributing factor to and a potential solution for this issue, this study explores key competencies in education for sustainable development as a framework to enhance plant awareness. A conceptual analysis was conducted to identify thematic relationships between this framework and plant awareness. The analysis suggests that enhancing systems thinking and integrated problem‐solving competencies can help learners better recognize and understand the importance of plants for both human welfare and planet Earth. Through the development of critical thinking, normative, and self‐awareness competencies, learners are encouraged to question existing personal and societal perspectives on plants, thereby reshaping their perception of flora. Moreover, the application of anticipatory, strategic, and collaboration competencies allows learners to explore the intrinsic values of the plant world more deeply, fostering respect and empathy, which can lead to a broader shift in attitudes toward flora. Through the integration of these elements into plant education, botany classes can become more engaging and relevant to real‐world issues. This approach can help bridge the gap between traditional science education and the development of pro‐conservation behaviors, while also enriching the evaluation methods used to assess plant awareness. Additionally, key competencies in education for sustainable development within plant education can promote not only a less utilitarian perspective of plants as organisms but also a more holistic approach to science education, reducing its emphasis on instrumentalization.
The status of the equivalence principle in modified symmetric teleparallel gravity is examined. In this theory, minimum length geodesics are distinct from autoparallel geodesics, that is, the ``shortest'' paths are not the ``straightest'' paths. We show that a standard argument that singles out metric geodesics in general relativity does not apply in modified symmetric teleparallel gravity. This is because the latter theory does not obey the equivalence principle in the sense of Weinberg. We argue, however, that the structure of the theory makes it inevitable that a freely falling test particle follows a shortest path, a geodesic of the metric. The geodesic equation that governs the motion of a freely falling test particle involves the Levi-Civita connection, not some other connection obtained by solving the connection field equations of the theory. This also has bearing on whether, under appropriate conditions, modified symmetric teleparallel gravity is fully equivalent to general relativity.
Simplicity is held by many to be the key to general intelligence. Simpler models tend to "generalise", identifying the cause or generator of data with greater sample efficiency. The implications of the correlation between simplicity and generalisation extend far beyond computer science, addressing questions of physics and even biology. Yet simplicity is a property of form, while generalisation is of function. In interactive settings, any correlation between the two depends on interpretation. In theory there could be no correlation and yet in practice, there is. Previous theoretical work showed generalisation to be a consequence of "weak" constraints implied by function, not form. Experiments demonstrated choosing weak constraints over simple forms yielded a 110-500% improvement in generalisation rate. Here we show that all constraints can take equally simple forms, regardless of weakness. However if forms are spatially extended, then function is represented using a finite subset of forms. If function is represented using a finite subset of forms, then we can force a correlation between simplicity and generalisation by making weak constraints take simple forms. If function is determined by a goal directed process that favours versatility (e.g. natural selection), then efficiency demands weak constraints take simple forms. Complexity has no causal influence on generalisation, but appears to due to confounding.
This paper leverages various philosophical and ontological frameworks to explore the concept of embodied artificial general intelligence (AGI), its relationship to human consciousness, and the key role of the metaverse in facilitating this relationship. Several theoretical frameworks underpin this exploration, such as embodied cognition, Michael Levin's computational boundary of a "Self," and Donald D. Hoffman's Interface Theory of Perception, which lead to considering human perceived outer reality as a symbolic representation of alternate inner states of being, and where AGI could embody a different form of consciousness with a larger computational boundary. The paper further discusses the necessary architecture for the emergence of an embodied AGI, how to calibrate an AGI's symbolic interface, and the key role played by the Metaverse, decentralized systems and open-source blockchain technology. The paper concludes by emphasizing the importance of achieving a certain degree of harmony in human relations and recognizing the interconnectedness of humanity at a global level, as key prerequisites for the emergence of a stable embodied AGI.
La búsqueda de articular el espacio y el tiempo forma parte constitutiva de la Geografía en tanto tradición disciplinar. En este sentido, el pasado puede ser considerado como una dimensión explicativa del presente tanto en lo que tiene que ver con el devenir de las ideas geográficas como con las transformaciones territoriales. Este texto pone el foco en tres formas de abordar la relación entre las geografías y los pasados: los procesos de formación territorial, la relación entre paisaje y patrimonio, y la decolonialidad del saber y la comprensión de la circulación de ideas geográficas.
PRIMA (The PRobe for-Infrared Mission for Astrophysics) is a concept for a far-infrared (IR) observatory. PRIMA features a cryogenically cooled 1.8 m diameter telescope and is designed to carry two science instruments enabling ultra-high sensitivity imaging and spectroscopic studies in the 24 to 235 microns wavelength range. The resulting observatory is a powerful survey and discovery machine, with mapping speeds better by 2 - 4 orders of magnitude with respect to its far-IR predecessors. The bulk of the observing time on PRIMA should be made available to the community through a General Observer (GO) program offering 75% of the mission time over 5 years. In March 2023, the international astronomy community was encouraged to prepare authored contributions articulating scientific cases that are enabled by the telescope massive sensitivity advance and broad spectral coverage, and that could be performed within the context of GO program. This document, the PRIMA General Observer Science Book, is the edited collection of the 76 received contributions.
To make accurate inferences in an interactive setting, an agent must not confuse passive observation of events with having intervened to cause them. The $do$ operator formalises interventions so that we may reason about their effect. Yet there exist pareto optimal mathematical formalisms of general intelligence in an interactive setting which, presupposing no explicit representation of intervention, make maximally accurate inferences. We examine one such formalism. We show that in the absence of a $do$ operator, an intervention can be represented by a variable. We then argue that variables are abstractions, and that need to explicitly represent interventions in advance arises only because we presuppose these sorts of abstractions. The aforementioned formalism avoids this and so, initial conditions permitting, representations of relevant causal interventions will emerge through induction. These emergent abstractions function as representations of one`s self and of any other object, inasmuch as the interventions of those objects impact the satisfaction of goals. We argue that this explains how one might reason about one`s own identity and intent, those of others, of one`s own as perceived by others and so on. In a narrow sense this describes what it is to be aware, and is a mechanistic explanation of aspects of consciousness.
Igor Žiberna, Danijel Ivajnšič, Eva Konečnik Kotnik
This research analyses changes of built-up and related areas in Slovenia in the period 2000-2023. Results are prepared on the level of statistical regions and municipalities. Special emphasis is orientated towards original categories of land use, on which built-up and related areas are located today. Finally, we discuss the expansion of built-up and related areas into arable areas, especially those that are extremely important for agriculture and food production.
Human societies have been reshaping the geomorphology of landscapes for thousands of years, producing anthropogenic geomorphic features ranging from earthworks and reservoirs to settlements, roads, canals, ditches and plough furrows that have distinct characteristics compared with landforms produced by natural processes. Physical geographers have long recognized the widespread importance of these features in altering landforms and geomorphic processes, including hydrologic flows and stores, to processes of soil erosion and deposition. In many of the same landscapes, archaeologists have also utilized anthropogenic geomorphic features to detect and analyse human societal activities, including symbolic formations, agricultural systems, settlement patterns and trade networks. This paper provides a general framework aimed at integrating geophysical and archaeological approaches to observing, identifying and interpreting the full range of anthropogenic geomorphic features based on their structure and functioning, both individually and as components of landscape-scale management strategies by different societies, or “sociocultural fingerprints”. We then couple this framework with new algorithms developed to detect anthropogenic geomorphic features using precisely detailed three-dimensional reconstructions of landscape surface structure derived from LiDAR and computer vision photogrammetry. Human societies are now transforming the geomorphology of landscapes at increasing rates and scales across the globe. To understand the causes and consequences of these transformations and contribute to building sustainable futures, the science of physical geography must advance towards empirical and theoretical frameworks that integrate the natural and sociocultural forces that are now the main shapers of Earth’s surface processes.