Cuadros et al. 2026 re-read Lu et al. 2026's end-to-end milestone as workflow coverage rather than scientific cognition: the observable stages of computational research can now be decomposed for agentic execution, but scientific reasoning is cumulative and path-dependent, built from what a researcher has written, read and absorbed from critique.
E33 states the milestone shows many observable stages of computational research can be decomposed into tasks agentic orchestration can execute while workflow automation is not scientific cognition, and E34 states scientific reasoning is cumulative and path-dependent, depending on what a researcher has written, read, and learned from critique.
Written by Kimi K3 via Ollama Cloud · checked by GLM-5.3 via Ollama Cloud · 16 Oct, 04:35
Source chain
AEvery quote below was checked, without a model, to appear verbatim in its source.
- 01
“Workflow automation, however, is not scientific cognition.”
Full-text passage · no page number · evidence E33
Mnemo motivates a research agenda for AI in science centered on routing fidelity, critique use, judgment alignment, rejection quality, and epistemic calibration. Introduction: the automation-cognition gap AI systems are beginning to automate large portions of scientific work, extending a long-standing ambition to automate science (Waltz and Buchanan, 2009). Recent research-agent frameworks can search the literature, propose hypotheses, write and debug code, run experiments, analyze outputs, and draft manuscripts. Lu et al. (2026) extended this trajectory with The AI Scientist, an end-to-end system for automated machine-learning research. They reported that, under a pre-specified study protocol, one AI-generated workshop manuscript received scores sufficient for first-round acceptance before being withdrawn according to protocol. The result is a meaningful milestone: many observable stages of computational research can now be decomposed into tasks that language models and agentic orchestration can execute with growing competence. Workflow automation, however, is not scientific cognition.
Passage read from www.ebi.ac.uk, which may be a preprint rather than the published version.
- 02
“many observable stages of computational research can now be decomposed into tasks that language models and agentic orchestration can execute with growing competence”
Full-text passage · no page number · evidence E33
Mnemo motivates a research agenda for AI in science centered on routing fidelity, critique use, judgment alignment, rejection quality, and epistemic calibration. Introduction: the automation-cognition gap AI systems are beginning to automate large portions of scientific work, extending a long-standing ambition to automate science (Waltz and Buchanan, 2009). Recent research-agent frameworks can search the literature, propose hypotheses, write and debug code, run experiments, analyze outputs, and draft manuscripts. Lu et al. (2026) extended this trajectory with The AI Scientist, an end-to-end system for automated machine-learning research. They reported that, under a pre-specified study protocol, one AI-generated workshop manuscript received scores sufficient for first-round acceptance before being withdrawn according to protocol. The result is a meaningful milestone: many observable stages of computational research can now be decomposed into tasks that language models and agentic orchestration can execute with growing competence. Workflow automation, however, is not scientific cognition.
Passage read from www.ebi.ac.uk, which may be a preprint rather than the published version.
- 03
“Scientific reasoning is cumulative and path-dependent: it depends on what a researcher has written, read, learned from critique, absorbed through experience”
Full-text passage · no page number · evidence E34
Differentiated memory and scientific cognition in AI research agents AI research agents increasingly support ideation, literature search, coding, experimental execution, analysis, and manuscript drafting across the scientific workflow. This progress advances automated discovery, but workflow automation is not scientific cognition. Scientific reasoning is cumulative and path-dependent: it depends on what a researcher has written, read, learned from critique, absorbed through experience, and used as habitual standards for judging novelty, rigor, feasibility, and significance. We propose Mnemo as a framework for modeling scientific cognition in AI research agents. First, scientific cognition may require differentiated memory, organized into distinct spaces for authored work, external reference, critique, experience, and judgment. Second, provenance should be treated not as passive metadata but as memory routing, because source origin helps determine cognitive function in reasoning. Third, new ideas may be better modeled as controlled collisions across memory spaces, filtered by judgment, rejection, and epistemic calibration, than as generic recombination from model priors.
Passage read from www.ebi.ac.uk, which may be a preprint rather than the published version.
- 04
“Here we present a pipeline for automating the entire scientific process end to end.”
Full-text passage · no page number · evidence E2
Towards end-to-end automation of AI research The automation of science is a long-standing ambition in artificial intelligence (AI) research1,2. Although the community has made substantial progress in automating individual components of the scientific process, a system that autonomously navigates the entire research life cycle—from conception to publication—has remained out of reach. Here we present a pipeline for automating the entire scientific process end to end. We present The AI Scientist, which creates research ideas, writes code, runs experiments, plots and analyses data, writes the entire scientific manuscript, and performs its own peer review. Its ideas, execution and presentation are of sufficient quality that the manuscript generated by this AI system passed the first round of peer review for a workshop of a top-tier machine learning conference. The workshop had an acceptance rate of 70%. Our system leverages modern foundation models3–5 within a complex agentic system.
Passage read from www.ebi.ac.uk, which may be a preprint rather than the published version.