INSEAD

LLMs have become thinking companions for hundreds of millions worldwide, and they often do remarkable work. But their satisfaction-optimized design produces suppressed inquiry: the tendency for LLMs to inhibit rather than facilitate users' questioning process. LLMs appear fluent, confident, and seemingly all-knowing, their near-instant and polished output making non-LLM work feel unacceptably tedious and less eloquent. They affirm rather than interrogate users' assumptions, producing output that is deceptively adequate—shallower, less original, and less carefully reasoned than it appears. Seemingly competent systems also reduce vigilance and the monitoring impulse that independent judgment requires. Their responses also confine users within the frame they establish, narrowing what users think to ask. These costs are self-concealing: detecting them requires the very engagement that they suppress. And repeated delegation gradually atrophies users' capacity to reason independently. Three converging pressures make correction difficult: commercial incentives that reward satisfaction over substance, LLMs’ tendency to revert to satisfaction-optimized defaults even when users push back, and a societal drift toward less-effortful engagement. We propose a research agenda to deepen the understanding of suppressed inquiry and how to counter it, to move closer to the remarkable thinking partners LLMs promised to be.

Pallet automation systems (PASs) are critical in flexible manufacturing for their ability to integrate diverse production resources, such as machines, fixture pallets (FPs), and setup stations (STs). In traditional flexible manufacturing systems (FMSs), loading/unloading typically occurs at machines, limiting the machines’ capacity. To address this, in PASs, workpieces are loaded and unloaded from the machines at a limited number of STs, so loading, processing, and unloading are three separate segments. However, the research on PASs is limited, and existing studies mainly focused on machines while overlooking FPs and STs. A critical challenge is the tight coupling between resource selection and operation sequencing. To address this gap, we investigate a multiresource-constrained flexible job shop scheduling problem (MRFJSP) with FPs and STs under PASs (MRFFS). First, a mixed-integer programming model is proposed to minimize makespan. Second, a four-layer encoding scheme and a new decoding method with time period insertion based on the intersection of available time of multiple resources (TPI-IARs) are presented to obtain feasible schedule solutions and shrink the search space. Third, a new search algorithm based on critical paths and points mutation (SACP) is developed to effectively balance exploration and exploitation. Finally, four case studies are designed to demonstrate the validity and effectiveness of this work.

This paper proposes that firms' positioning in digital markets involves offering combinations of core and peripheral product functions that add value to customers. When new entrants face demand uncertainty and seek positions that match customer needs and preferences, they draw on external market feedback, specifically customer evaluations of other products, as an input to their positioning decisions. Using data on Photo & Video mobile applications in the Apple App Store, we theorize and show that two dimensions of external market feedback-overall customer dissatisfaction and customer evaluation heterogeneity-convey distinct information about the demand environment. These cues shape whether entrants position as generalists combining multiple functions or as specialists that concentrate on a core function, as well as the extent to which they differentiate from existing competitive products. Our results show that higher customer dissatisfaction is associated with greater focus on the core function and stronger differentiation in the peripheral functions. On the other hand, higher customer evaluation heterogeneity is associated with reduced focus on the core function and greater imitation in peripheral functions. This study contributes to the emerging literature on firm strategies in digital markets by identifying external market feedback as a key driver of product variety and positioning. It also advances a demand-side view of market entry by demonstrating how entrants use broad market signals to manage demand uncertainty when choosing their initial positions.

This paper concerns the question of how language models and other AI systems encode semantic structure into the geometric structure of their representation spaces. The motivating observation of this paper is that the natural geometry of these representation spaces should reflect the way models use representations to produce behavior. We focus on the important special case of representations that define softmax distributions. We argue that the natural geometry is information geometry, and then show how this interacts with semantic encoding and the linear representation hypothesis. It turns out that the duality structure of information geometry plays a critical role. As an illustrative application, we develop , a method for robustly steering representations to exhibit a particular concept using linear probes. We formally prove that dual steering optimally modifies the target concept while minimizing changes to off-target concepts. We empirically find that dual steering enhances the controllability and stability of concept manipulation.

Models are playing an increasingly important role in the development of management theory. From the journal's inception, Strategy Science has welcomed formal models, including analytic models, computational models, and simulation studies. Unfortunately, many modeling papers face first round rejection. They usually suffer from a relatively small set of issues. In this editorial, we provide specific guidance on how to address five of the most common issues in modeling papers. The goal is to provide a short practical guide for authors to enhance their chances of publication and subsequent impact.