We design, fund, and operate high-impact programs that prototype a new model for how science is done, shared, and scaled for real-world utility.
The Scientific Contribution Graph: Automated Literature-based Technological Roadmapping at Scale
Sir Isaac Newton famously wrote, “If I have seen further, it is by standing on the shoulders of giants”. Scientific contributions are rarely developed in isolation, but build upon prior contributions, such as problem framings, experimental methods, and empirical findings. Understanding these prerequisite relationships is important for studying scientific progress, and for automated scientific discovery systems that must reason about which existing capabilities can be used to develop new ones (e.g. Lu et al., 2024; Jansen et al., 2025b; Baek et al., 2025).
Linking the world's research to the code it runs on - OpenAlex blog
Research relies on software. Software written by scientists, for science, runs through the entire modern research stack: NumPy and SciPy, R and ggplot2, Jupyter, BLAST, ImageJ, AlphaFold. Yet in the scholarly record, that software is nearly invisible. Software is not usually cited formally in publications and is usually just mentioned in the text, which means […]
Hyperproblems: New Ways of Doing and Communicating Science - Hyperproblems
Hyperproblems: Hyperproblems are scientific challenges whose scale, complexity, novelty and interdependence overwhelm traditional research models, requiring…
Kevin Weil 🇺🇸 on Twitter / X
💥 Today we’re introducing Prism—a free, AI-native workspace for scientists to write and collaborate on research, powered by GPT-5.2.Accelerating science requires progress on two fronts:1. Frontier AI models that use scientific tools and can tackle the hardest problems2.… pic.twitter.com/cnLysHixuQ— Kevin Weil 🇺🇸 (@kevinweil) January 27, 2026
Banff 2025, Open Science Meeting — Continuous Science Foundation
We’ve got to a big idea—and now it’s time to get practical. CSF is looking for curious, committed individuals to help us tackle some of the toughest (and most exciting) questions around how to make Composable Science real.
Ten quick tips to SNIFF out sustainable and secure scientific software
Modern computational biology depends heavily on open-source software tools, analysis pipelines, and containerized workflows developed and shared by the research community. While there is extensive guidance (including Quick Tips and Simple Rules articles) on how to build robust and sustainable scientific software, far less has been written for researchers in the role of software users evaluating whether an existing tool is reliable, secure, and sustainable enough for their work. Here we present ten quick tips to help researchers critically assess the tools they adopt. Our tips are organized around a framework that centers on key evaluation features: source, network, interaction, fit, and fragility (SNIFF). These dimensions prompt researchers to consider who maintains a tool and why, whether it is embedded in a broader ecosystem, how actively its developers and users engage, whether it matches the intended use case and licensing requirements, and how robust its dependencies and security practices are. By applying these tips, researchers can make more informed decisions, reduce the risk of relying on abandoned or insecure software, and contribute to a more sustainable scientific software ecosystem.
Seth Bannon on Twitter / X
There are some exceptionally good ideas here for reforming and accelerating science.Some of my favorite proposals:Apply the scientific method to science funding (track applications, reviewers, funding decisions, and outcomes, then experimentally test which grantmaking methods… https://t.co/Q0JmqTBPYm— Seth Bannon (@sethbannon) July 22, 2026
SCI — Software Carbon Intensity | Green Software Foundation
The global standard for calculating and reducing the carbon emissions of your software. ISO/IEC 21031:2024.
Games Demystified: Super Mario Galaxy
Want to see how Mario Galaxy created its unique gravity-based physics effects for the in-game planets? Developer Alessi analyzes and reproduces the same concepts with a playable game prototype and source code.
Kosmos: An AI Scientist for Autonomous Discovery
Today, we are announcing Kosmos, our next-generation AI Scientist. Kosmos is a major upgrade on Robin, our previous AI Scientist. You can read about it in our technical report, here. Kosmos is available to use from day one on our platform, here.
ATScience 2026 Agenda - ATScience Notes
Join us for a full-day of scientific explorations on the AT Protocol
science-live-platform/zotero at main · ScienceLiveHub/science-live-platform
Science Live Platform - Transform research into connected knowledge - ScienceLiveHub/science-live-platform
Hypercerts: Recognizing and Rewarding Impact – ATProto Implementation
Hi all, excited to share some of what we’ve been building and to introduce hypercerts to the ATProto community. Where we started The hypercerts project began at Protocol Labs with a simple goal: improve how we fund public goods. Many of the contributions society relies on most—open-source software, scientific R&D, and ecological regeneration—remain underfunded because their value is hard to see, coordinate around, and reward. What are hypercerts Hypercerts address this by serving as digital im...
Stargazer: A Scalable Model-Fitting Benchmark Environment for AI Agents under Astrophysical Constraints
The rise of autonomous AI agents suggests that dynamic benchmark environments with built-in feedback on scientifically grounded tasks are needed to evaluate the capabilities of these agents in research work. We introduce Stargazer, a scalable environment for evaluating AI agents on dynamic, iterative physics-grounded model-fitting tasks using inference on radial-velocity (RV) time series data. Stargazer comprises 120 tasks across three difficulty tiers, including 20 real archival cases, covering diverse scenarios ranging from high-SNR single-planet systems to complex multi-planetary configurations requiring involved low-SNR analysis. Our evaluation of eight frontier agents reveals a gap between numerical optimization and adherence to physical constraints: although agents often achieve a good statistical fit, they frequently fail to recover correct physical system parameters, a limitation that persists even when agents are equipped with vanilla skills. Furthermore, increasing test-time compute yields only marginal gains, with excessive token usage often reflecting recursive failure loops rather than meaningful exploration. Stargazer presents an opportunity to train, evaluate, scaffold, and scale strategies on a model-fitting problem of practical research relevance today. Our methodology to design a simulation-driven environment for AI agents presumably generalizes to many other model-fitting problems across scientific domains. Source code and the project website are available at https://github.com/Gudmorning2025/Stargazer and https://gudmorning2025.github.io/Stargazer, respectively.
