Science 2.0 Research Program
The Science 2.0 Research Program is an initiative designed to rethink how science is evaluated, shared, and pursued—centered on intrinsic motivation, epistemic integrity, and collaborative trust. In light of accelerating AI capabilities and restructuring in federal research funding frameworks, this program cultivates the innovation of enhanced models for academic pathways. One significant area of this program aims to develop a model that helps researchers reclaim autonomy from funder priorities and publisher expectations defined by externally imposed evaluative subjectivity.
An effort within the direction is focused on redesigning peer review as a trust-building mechanism, addressing how credibility is established in scientific communities. This direction explores systems designed to increase transparency, reduce evaluative bias, and align feedback with author-defined values and epistemic goals, rather than prestige-driven incentives. Rather than reinforcing hierarchical gatekeeping in a competition-based selection, these models emphasize reason-giving, clarity of critique, and continuous engagement—treating peer review as a structured, open-ended dialogue that advances discovery on the researcher’s own terms. The approach aims to make research evaluation more accountable, more scalable, and more faithful to the real dynamics of inquiry.
Beyond this area, the program supports broader inquiries into perennial questions about how the structures of scientific enterprise and system might evolve by leveraging emerging technologies and transforming systemic pressures from detrimental to beneficial. These include new models of researcher collectives, meta-research efforts, and the use of AI as a cognitive amplifier to support critical thinking, synthesis, and strategic planning—freeing researchers to concentrate on inquiry, leadership, and scientific understanding. This program offers a framework for a more open and effective research ecosystem.
We welcome sponsors and collaborators to power this mission.
Collaborating with Gim Soon Wan (IEEE) and Zacharie Mbaitiga (NIT-Ok).
With valuable contributions from Charles Yokoyama (Fujita Health University), Micah Altman (MIT), and Pierre Azoulay (MIT).
CephNeuroAI Initiative
The CephNeuroAI Initiative is a research-driven program aimed at developing a practical and effective model for convergence research, centered on the underexplored yet highly promising topic of uncovering the circuit-level and computational foundations of cephalopod behavior. Cephalopods—such as octopuses and squid—possess a unique combination of distributed nervous systems, locally autonomous neural processing, and advanced adaptive behaviors, making them an ideal model for exploring how cognition and decision-making emerge in both biological and artificial systems. By bridging neuroscience, artificial intelligence, biotechnology, biomedicine, social sciences, philosophy of mind, and additional science-related disciplines, the initiative provides a platform to examine the nature of intelligence beyond traditional vertebrate models, focusing on how decentralized systems integrate sensory, motor, and cognitive functions.
CephNeuroAI centers on three core objectives. First, it develops wireless, multiplexed neurotechnologies that allow precise and reversible modulation of neural activity in freely behaving cephalopods—overcoming the limitations of existing methods. Second, the initiative designs parametric behavioral tasks to evaluate how centralized and distributed circuits interact during real-time decision-making. Third, it constructs mechanistic computational models that simulate hypothesized circuit operations while capturing latent cognitive variables. Together, these components form a structured, interdisciplinary approach for decoding the neural logic behind cephalopod intelligence.
The initiative’s broader impact lies in establishing a comparative framework for adaptive biocomputation, expanding the boundaries of systems neuroscience and informing the development of distributed control systems in AI and robotics. CephNeuroAI serves as a collaborative hub for researchers across science, engineering, and philosophy, fostering a transdisciplinary culture of inquiry and innovation. By studying cephalopods as living models of non-hierarchical cognition, the initiative challenges conventional assumptions about control, agency, and intelligence—contributing new perspectives to both fundamental science, applied technologies, and philosophy of mind.
We welcome sponsors and collaborators to power this mission.
Collaborating with Horst Obenhaus (NTNU), Joshua Rosenthal (MBL), and Stephen Senft (MBL).
With valuable contributions from Matthias Michel (MIT).
EpistemeOS Strategy Program
The EpistemeOS Strategy Program is guided by a long-horizon vision: researchers can safely and creatively “drive” ever-stronger means—AI and beyond—upholding epistemic responsibility, agency, or meaning, grounded in the mindset and skills required to lead oneself and others. As scientific work accelerates and becomes increasingly mediated by powerful capabilities, progress often depends less on raw performance than on steering: clarifying one’s commitments, exercising judgment under uncertainty, choosing problems worth pursuing, coordinating people and methods, and preserving truthfulness when incentives, complexity, and speed pull in different directions. EpistemeOS program treats this steering capacity as both a research subject and a trainable discipline, aiming to develop strategy operating principles that remain transferable across fields, institutions, and eras.
A direction explores research strategy as an operating system for inquiry—particularly when goals, signals, and degrees of control are in flux—by characterizing how researchers coordinate their reserach experineces, values, and philosophy, with input for strategy planning and decision-making from both human and AI experts, including contextual influences on judgment. This direction aims to develop models and practices for converting ambiguity— including periods of reduced clarity or controllability— into structured progress. This direction also examines, under such conditions, the practical feasibility of decision traceability and failure-to-learning loops throughout the research process. The program’s intent is to strengthen researchers’ practical clarity, support structures, and confidence while enabling contributions that remain meaningful to knowledge and community.
Beyond this, through dialogue, exchange, and perspective-taking, the program pursues an integrated agenda on strategy execution and coordination across research roles and career stages, including AI-augmented workflows and practices to release cognitive bandwidth for rigorous, self-directed inquiry. In this way, the program works to develop and refine transferable approaches that can be adapted to diverse research contexts.
We welcome sponsors and collaborators to power this mission.
Collaborating with Arda Ecevit (NexStrat) and Şölen Soya (MIT SFMBA’23).
Past Events
2025
2025 Gold Key Award Presentation — Jennifer A. Lewis
MIT Sigma Xi. November 3rd, 2025 | Cambridge, MA, USA
Sponsored by Sigma Xi, The Scientific Research Honor Society (ΣΞ)
CephNeuroAI Interdisciplinary Research Dialogue: MIT × MBL
CephNeuroAI Initiative, Division of Convergence Research, MIT Sigma Xi. May 30th, 2025 | Cambridge, MA, USA
Sponsored by MIT School of Science—SQoL Grant (AY24-25 Spring #004302)