Request for Content: Latent Space Issue #1

This is a request for content. It’s time to broaden the set of intellectual goods that scientists can publish and get credit for. While traditional academic publishing is restricted to original research articles, reviews, and textbooks, there are a number of other intellectual goods we already produce that are unrecognized as valuable outputs and remain largely unpublished except as blogs. These include hypothesis registration, missing dataset justification, gap analysis, technology development roadmaps, and retrospective development stories that document how ambitious projects have actually unfolded, what was learned along the way, and how those involved have had to update their mental models of their fields. Each of these explores the idea of the adjacent possible, the set of potential things that are visible from current data but not yet realized. The adjacent possible is latent in every action we undertake as scientists: designing experiments, choosing questions to address, grubbing for grant money, yet is rarely made explicit or rewarded as a serious intellectual good. Although there’s nothing formally stopping you from publishing a technology development roadmap as a literature review with Nature Publishing Group, you also know it would be a hard sell and the standards of formal review papers represent an enormous amount of work and a higher barrier to entry. There’s plenty of space for intellectual output between a technical blog post and a formal traditional article that’s absolutely worth exploring, so the following is an experiment in DeSci publishing to launch a quarterly journal, Latent Space, exploring the adjacent possible and testing new ways of using bounties, writing in public, and rewarding new kinds of content. 

For its first issue, Latent Space hereby requests applications for the following Request for Content (RFC). Applications will be made to topic-specific bounties and the winner of each will commit to writing the full piece within the next 90 days, where completion and publication will be rewarded with a second bounty and all the rewards of authorship inherent to publishing on ResearchHub. Anyone can apply to any topic because publication is the ultimate proof of work and ResearchHub’s open peer review will judge quality and find anything you’ve missed, although you’re encouraged to work with themes that you know you can bring a high level of rigor to. For RFC2024.1, Latent Space is soliciting technology development roadmaps, missing dataset justifications, and/or gap analyses for any of the following topics:

1. Viable cell and gene therapy business models for rare conditions and personalized medicine
2. Dispositive evidentiary standards for superconducting material testing
3. High-priority bioanalytes for improving dynamic variation resolution (e.g., diurnal fluctuations in prolactin)
4. Sector-specific integration maps for high-impact use of AI/ML tools even at 5% hallucination rates
5. Economic estimates of lost productivity accruing from multi-party environmental review of new development projects in energy, transportation, and/or housing
6. Prediction and justification of what co-developments will be necessary to achieve >50% overall response rates in adoptive T-cell immunotherapy for solid cancers
7. Specific projects only feasible with exabyte-scale computing resources for data integration, modeling, and/or AI/ML development

Applicants will nominate their content for acceptance by answering the corresponding bounty with a full outline of the content they plan to produce and one completed section thereof within 30 days of the RFC announcement. Applicants may work alone or in teams of any size they want, although one lead author will be ultimately responsible for content completion and handling splitting of the bounty based on co-author contribution.

Content may be structured as author teams see fit but should meet the following requirements for each of the content classes below:

Technology development roadmaps

• Provide a clear and comprehensive overview of the current state of the technology including active development fronts vs abandoned fonts
• Identify key challenges and current barriers to further development and adoption, including quantitative performance needs where possible
• Outline a step-by-step plan for advancing the technology over a specified time frame (e.g., 5-10 years) with identification of key dependencies for each branch
• Estimate resources (financial, human, and infrastructural) required for each stage of development
• Include visual aids such as timelines, flowcharts, or infographics to enhance clarity and understanding

Missing dataset justifications

• Don’t worry about whether it is a fishing expedition, just make sure it’s a good fishing expedition going after high value fish
• Identify an outstanding specific research question or problem that cannot be adequately addressed with existing datasets
• Describe the ideal dataset that would be needed to fully address the research question or at least identify specific avenues for inquiry that enable new development
• Justify the need for this missing dataset by explaining how it would advance knowledge in the field, including potential benefits, revision of adjacent knowledge, and testing of long standing assumptions
• If possible, identify why this dataset has not previously been generated

Gap analyses

• Provide a high-level review of the current state of knowledge in a specific research area: what is the shape of the frontier?
• Identify key gaps in understanding or areas where further research is needed
• Prioritize the identified gaps based on their potential impact on advancing the field
• Propose specific research questions or hypotheses that could be addressed to fill each gap
• Discuss how technical, infrastructure, or resource issues have gated development of effective solutions that could close the extant gap and how they could be overcome