Reasoning models are large language models that use extra "thought tokens" before answering, providing both higher accuracy and explicit reasoning for their response. A major question has been whether language model reasoning generalizes beyond mathematics, programming, and logic, where most previous work has focused. We demonstrate that reasoning models can be post-trained in scientific domains without additional domain pretraining, and require substantially less data compared to contemporary domain-specific models. We report ether0, a 24B parameter LLM (based on Mistral-Small-24B) that can reason in natural language and respond with chemical structures. This reasoning model was trained with reinforcement learning on 577,790 experimentally-grounded chemistry tasks involving synthesized organic molecules. Our model outperforms all previous general-purpose chemistry models, frontier models, and humans, and is more data efficient relative to specialized models. We anticipate that this method can be applied to train highly data-efficient language models specialized for predictive and generative tasks across a wide variety of scientific domains.

To perform effectively in markets that are becoming increasingly turbulent and volatile, organisations should form information system linkages with their business partners. This study seeks to produce a ‘vision’ for the future of such linkages. A number of recent developments in information systems and technology (IS/IT) appear to promise the ability to make improvements in this domain. These are electronic hubs, web services, widespread adoption of common enterprise resource planning (ERP) systems and enterprise portals. This study, which is carried out by means of the Delphi technique, explores the expected future role and use of these information systems. The expert participants in the study believe that the focus of enterprise systems is shifting from an internal to an external orientation. They suggest that ERP systems may be reaching a structural limit concerning their capabilities and adjunct technologies will be required to integrate multiple inter-organisational operations. The participants expect that the three other systems and technologies considered can provide those necessary adjunct technologies, either used singly, or more likely in combination. Finally, we find that the transition to a greater external orientation is expected to include an extension in the types of organisations that will be integrated electronically into networks.

The electronic marketplace domain has recently witnessed the joining together of a number of previously independent marketplaces and the formation of collaborative alliances between others. This study seeks to determine the nature or type of cooperative arrangements or alliances that are currently being forged between electronic marketplaces, and the strategic rationale that is leading to this observed alliance formation. The findings of the study are based upon an examination of statements made to the press by marketplaces undertaking strategic alliances. This published data was supplemented with face-to-face interviews with managers at three relevant electronic marketplaces. Three distinct types of alliance are observed; the merger between previously separate marketplaces, the acquisition of one marketplace by another and the formation of an interoperability agreement between two marketplaces. Three rationales for alliance formation were observed: an increase in the number of buyers or suppliers in a given market sector that can access the marketplace (an increase in scale of operations), an increase in the breadth or depth of services that are offered to users of the marketplace (an increase in the scope of operations) and providing the ability to exchange information across multiple tiers of a supply chain. A broader discussion of the findings is given and suggestions for further research are made.