analiza aiqbench
This study reveals a critical bottleneck in AI advancement: massive breakthroughs like the Transformer era have paradoxically narrowed the field's intellectual diversity by concentrating attention on winning paradigms rather than expanding the space of ideas. The analysis of decades of research data shows that while carrying capacity for new concepts increases, the rate of idea exhaustion accelerates, suggesting that current incentive structures and peer review policies may inadvertently stifle future innovation. Institutions are urged to shift focus from rewarding concentration on dominant models to protecting diverse attention across the broader landscape of computer science.
Stronger artificial intelligence breakthroughs hasten their own exhaustion
arXiv:2610.27296v1 Announce Type: new Abstract: Progress in artificial intelligence is widely assumed to scale with effort: more researchers, papers and investment should accelerate discovery. Theories of technological growth instead locate the difficulty of discovery in the space of ideas. Here we show that neither scale nor space is enough: breakthroughs depend on diverse attention, which collective concentration can deplete. Across 36 years of computer science, we analyze 7.2 million papers, 73,089 peer reviews and 3.7 million researchers and their topic shifts. The field expanded while its attention narrowed sharply after the Transformer. Over half of topic-shifting researchers moved into AI, especially mid-career ones, and reviewer confidence fell by a full point on a 5-point scale. Each successive breakthrough, from AlexNet to large language models, raised the field's carrying capacity for ideas but was followed by steeper exhaustion of those ideas. Bigger breakthroughs capture attention more completely, depleting the diversity their successors may draw upon. Structural biology shows the same cycle after cryo-EM and AlphaFold. Simulations show how well-intentioned and widely used policies such as reciprocal reviewing empower prolific authors and further narrow what the field accepts. We call for institutions that protect diverse attention in artificial intelligence rather than reward concentration on the winning paradigm.
Abstract:Progress in artificial intelligence is widely assumed to scale with effort: more researchers, papers and investment should accelerate discovery. Theories of technological growth instead locate the difficulty of discovery in the space of ideas. Here we show that neither scale nor space is enough: breakthroughs depend on diverse attention, which collective concentration can deplete. Across 36 years of computer science, we analyze 7.2 million papers, 73,089 peer reviews and 3.7 million researchers and their topic shifts. The field expanded while its attention narrowed sharply after the Transformer. Over half of topic-shifting researchers moved into AI, especially mid-career ones, and reviewer confidence fell by a full point on a 5-point scale. Each successive breakthrough, from AlexNet to large language models, raised the field's carrying capacity for ideas but was followed by steeper exhaustion of those ideas. Bigger breakthroughs capture attention more completely, depleting the diversity their successors may draw upon. Structural biology shows the same cycle after cryo-EM and AlphaFold. Simulations show how well-intentioned and widely used policies such as reciprocal reviewing empower prolific authors and further narrow what the field accepts. We call for institutions that protect diverse attention in artificial intelligence rather than reward concentration on the winning paradigm.
Artificial Intelligence (cs.AI); Computers and Society (cs.CY); General Economics (econ.GN)
arXiv:2610.27296 [cs.AI]
(or arXiv:2610.27296v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.27296 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 301 cuvinte
Subiecte conexe
Learning to Play: Reinforcement Learning and Games
arXiv · 15 oct. 2026
Integrating AI for Human-Centric Breast Cancer Diagnostics: A Multi-Scale and Multi-View Swin Transformer Framework
arXiv · 2 nov. 2026
Deep reinforcement learning for process design: Review and perspective
arXiv · 9 iun. 2026
Counterexample-driven autonomous research with application to excess radiated energy in a Z-pinch
arXiv · 4 nov. 2026