1% WP42 literature base -- every entry verified against Crossref, arXiv, or the official publisher page on 2026-11-02.2% See literature/VERIFICATION_LOG.csv for the exact URL fetched per key. Keys: authorYEARkeyword.34@article{abadie2023clustering,5 author = {Abadie, Alberto and Athey, Susan and Imbens, Guido W and Wooldridge, Jeffrey M},6 title = {When Should You Adjust Standard Errors for Clustering?},7 journal = {The Quarterly Journal of Economics},8 year = {2023},9 volume = {138},10 number = {1},11 pages = {1--35},12 doi = {10.1093/qje/qjac038},13 url = {https://doi.org/10.1093/qje/qjac038},14 note = {verified: crossref 2026-11-02. Crossref issued 2022-10-06 (online first); print volume 138(1), 2023.}15}1617@incollection{acemoglu2011skills,18 author = {Acemoglu, Daron and Autor, David},19 title = {Skills, Tasks and Technologies: Implications for Employment and Earnings},20 booktitle = {Handbook of Labor Economics},21 year = {2011},22 publisher = {Elsevier},23 doi = {10.1016/S0169-7218(11)02410-5},24 url = {https://doi.org/10.1016/S0169-7218(11)02410-5},25 note = {verified: crossref 2026-11-02}26}2728@article{acemoglu2018race,29 author = {Acemoglu, Daron and Restrepo, Pascual},30 title = {The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment},31 journal = {American Economic Review},32 year = {2018},33 volume = {108},34 number = {6},35 pages = {1488--1542},36 doi = {10.1257/aer.20160696},37 url = {https://doi.org/10.1257/aer.20160696},38 note = {verified: crossref 2026-11-02}39}4041@article{acemoglu2019automation,42 author = {Acemoglu, Daron and Restrepo, Pascual},43 title = {Automation and New Tasks: How Technology Displaces and Reinstates Labor},44 journal = {Journal of Economic Perspectives},45 year = {2019},46 volume = {33},47 number = {2},48 pages = {3--30},49 doi = {10.1257/jep.33.2.3},50 url = {https://doi.org/10.1257/jep.33.2.3},51 note = {verified: crossref 2026-11-02}52}5354@article{acemoglu2020robots,55 author = {Acemoglu, Daron and Restrepo, Pascual},56 title = {Robots and Jobs: Evidence from US Labor Markets},57 journal = {Journal of Political Economy},58 year = {2020},59 volume = {128},60 number = {6},61 pages = {2188--2244},62 doi = {10.1086/705716},63 url = {https://doi.org/10.1086/705716},64 note = {verified: crossref 2026-11-02}65}6667@article{acemoglu2025macroeconomics,68 author = {Acemoglu, Daron},69 title = {The simple macroeconomics of AI},70 journal = {Economic Policy},71 year = {2025},72 volume = {40},73 number = {121},74 pages = {13--58},75 doi = {10.1093/epolic/eiae042},76 url = {https://doi.org/10.1093/epolic/eiae042},77 note = {verified: crossref 2026-11-02}78}7980@techreport{agarwal2023radiology,81 author = {Agarwal, Nikhil and Moehring, Alex and Rajpurkar, Pranav and Salz, Tobias},82 title = {Combining Human Expertise with Artificial Intelligence: Experimental Evidence from Radiology},83 year = {2023},84 number = {31422},85 institution = {National Bureau of Economic Research},86 type = {NBER Working Paper},87 doi = {10.3386/w31422},88 url = {https://doi.org/10.3386/w31422},89 note = {verified: crossref 2026-11-02}90}9192@article{agrawal2019prediction,93 author = {Agrawal, Ajay and Gans, Joshua S. and Goldfarb, Avi},94 title = {Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction},95 journal = {Journal of Economic Perspectives},96 year = {2019},97 volume = {33},98 number = {2},99 pages = {31--50},100 doi = {10.1257/jep.33.2.31},101 url = {https://doi.org/10.1257/jep.33.2.31},102 note = {verified: crossref 2026-11-02}103}104105@misc{anthropic2024computeruse,106 author = {{Anthropic}},107 title = {Introducing computer use, a new Claude 3.5 Sonnet, and Claude 3.5 Haiku},108 year = {2024},109 month = oct,110 howpublished = {Anthropic announcement, 22 October 2024},111 url = {https://www.anthropic.com/news/3-5-models-and-computer-use},112 note = {verified: web (official Anthropic page fetched 2026-11-02; H1 and datePublished 2024-10-22 match)}113}114115@techreport{anthropic2025opus4card,116 author = {{Anthropic}},117 title = {System Card: Claude Opus 4 \& Claude Sonnet 4},118 year = {2025},119 month = may,120 institution = {Anthropic},121 url = {https://assets.claude.com/6d8a8055020700718b0c49369f60816ba2a7c285.pdf},122 note = {verified: web (PDF fetched 2026-11-02; title page reads "System Card: Claude Opus 4 \& Claude Sonnet 4, May 2025"; changelog to 2 Sept. 2025)}123}124125@misc{appel2025uneven,126 author = {Appel, Ruth and McCrory, Peter and Tamkin, Alex and McCain, Miles and Neylon, Tyler and Stern, Michael},127 title = {Anthropic Economic Index report: Uneven geographic and enterprise AI adoption},128 year = {2025},129 howpublished = {arXiv preprint arXiv:2511.15080},130 eprint = {2511.15080},131 archivePrefix = {arXiv},132 url = {https://arxiv.org/abs/2511.15080},133 note = {verified: arxiv 2511.15080}134}135136@article{arntz2017revisiting,137 author = {Arntz, Melanie and Gregory, Terry and Zierahn, Ulrich},138 title = {Revisiting the risk of automation},139 journal = {Economics Letters},140 year = {2017},141 volume = {159},142 pages = {157--160},143 doi = {10.1016/j.econlet.2017.07.001},144 url = {https://doi.org/10.1016/j.econlet.2017.07.001},145 note = {verified: crossref 2026-11-02}146}147148@misc{athey2016econometrics,149 author = {Athey, Susan and Imbens, Guido},150 title = {The Econometrics of Randomized Experiments},151 year = {2016},152 howpublished = {arXiv preprint arXiv:1607.00698},153 eprint = {1607.00698},154 archivePrefix = {arXiv},155 url = {https://arxiv.org/abs/1607.00698},156 note = {verified: arxiv 1607.00698. Published as chapter 3 of the Handbook of Economic Field Experiments, vol. 1 (2017), doi:10.1016/bs.hefe.2016.10.003.}157}158159@article{athey2020allocation,160 author = {Athey, Susan C. and Bryan, Kevin A. and Gans, Joshua S.},161 title = {The Allocation of Decision Authority to Human and Artificial Intelligence},162 journal = {AEA Papers and Proceedings},163 year = {2020},164 volume = {110},165 pages = {80--84},166 doi = {10.1257/pandp.20201034},167 url = {https://doi.org/10.1257/pandp.20201034},168 note = {verified: crossref 2026-11-02}169}170171@article{autor2003skill,172 author = {Autor, D. H. and Levy, F. and Murnane, R. J.},173 title = {The Skill Content of Recent Technological Change: An Empirical Exploration},174 journal = {The Quarterly Journal of Economics},175 year = {2003},176 volume = {118},177 number = {4},178 pages = {1279--1333},179 doi = {10.1162/003355303322552801},180 url = {https://doi.org/10.1162/003355303322552801},181 note = {verified: crossref 2026-11-02}182}183184@article{autor2015jobs,185 author = {Autor, David H.},186 title = {Why Are There Still So Many Jobs? The History and Future of Workplace Automation},187 journal = {Journal of Economic Perspectives},188 year = {2015},189 volume = {29},190 number = {3},191 pages = {3--30},192 doi = {10.1257/jep.29.3.3},193 url = {https://doi.org/10.1257/jep.29.3.3},194 note = {verified: crossref 2026-11-02}195}196197@misc{backlund2025vendingbench,198 author = {Backlund, Axel and Petersson, Lukas},199 title = {Vending-Bench: A Benchmark for Long-Term Coherence of Autonomous Agents},200 year = {2025},201 howpublished = {arXiv preprint arXiv:2502.15840},202 eprint = {2502.15840},203 archivePrefix = {arXiv},204 url = {https://arxiv.org/abs/2502.15840},205 note = {verified: arxiv 2502.15840}206}207208@article{bates2015lme4,209 author = {Bates, Douglas and Mächler, Martin and Bolker, Ben and Walker, Steve},210 title = {Fitting Linear Mixed-Effects Models Using lme4},211 journal = {Journal of Statistical Software},212 year = {2015},213 volume = {67},214 number = {1},215 doi = {10.18637/jss.v067.i01},216 url = {https://doi.org/10.18637/jss.v067.i01},217 note = {verified: crossref 2026-11-02}218}219220@misc{becker2025metr,221 author = {Becker, Joel and Rush, Nate and Barnes, Elizabeth and Rein, David},222 title = {Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity},223 year = {2025},224 howpublished = {arXiv preprint arXiv:2507.09089},225 eprint = {2507.09089},226 archivePrefix = {arXiv},227 url = {https://arxiv.org/abs/2507.09089},228 note = {verified: arxiv 2507.09089}229}230231@misc{bowman2022oversight,232 author = {Bowman, Samuel R. and Hyun, Jeeyoon and Perez, Ethan and Chen, Edwin and Pettit, Craig and Heiner, Scott and Lukošiūtė, Kamilė and Askell, Amanda and Jones, Andy and Chen, Anna and Goldie, Anna and Mirhoseini, Azalia and McKinnon, Cameron and Olah, Christopher and Amodei, Daniela and Amodei, Dario and Drain, Dawn and Li, Dustin and Tran-Johnson, Eli and Kernion, Jackson and Kerr, Jamie and Mueller, Jared and Ladish, Jeffrey and Landau, Joshua and Ndousse, Kamal and Lovitt, Liane and Elhage, Nelson and Schiefer, Nicholas and Joseph, Nicholas and Mercado, Noemí and DasSarma, Nova and Larson, Robin and McCandlish, Sam and Kundu, Sandipan and Johnston, Scott and Kravec, Shauna and Showk, Sheer El and Fort, Stanislav and Telleen-Lawton, Timothy and Brown, Tom and Henighan, Tom and Hume, Tristan and Bai, Yuntao and Hatfield-Dodds, Zac and Mann, Ben and Kaplan, Jared},233 title = {Measuring Progress on Scalable Oversight for Large Language Models},234 year = {2022},235 howpublished = {arXiv preprint arXiv:2211.03540},236 eprint = {2211.03540},237 archivePrefix = {arXiv},238 url = {https://arxiv.org/abs/2211.03540},239 note = {verified: arxiv 2211.03540}240}241242@article{brynjolfsson2017workforce,243 author = {Brynjolfsson, Erik and Mitchell, Tom},244 title = {What can machine learning do? Workforce implications},245 journal = {Science},246 year = {2017},247 volume = {358},248 number = {6370},249 pages = {1530--1534},250 doi = {10.1126/science.aap8062},251 url = {https://doi.org/10.1126/science.aap8062},252 note = {verified: crossref 2026-11-02}253}254255@article{brynjolfsson2018machines,256 author = {Brynjolfsson, Erik and Mitchell, Tom and Rock, Daniel},257 title = {What Can Machines Learn and What Does It Mean for Occupations and the Economy?},258 journal = {AEA Papers and Proceedings},259 year = {2018},260 volume = {108},261 pages = {43--47},262 doi = {10.1257/pandp.20181019},263 url = {https://doi.org/10.1257/pandp.20181019},264 note = {verified: crossref 2026-11-02}265}266267@techreport{brynjolfsson2025canaries,268 author = {Brynjolfsson, Erik and Chandar, Bharat and Chen, Ruyu},269 title = {Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence},270 year = {2025},271 institution = {Stanford Digital Economy Lab},272 type = {Working Paper},273 url = {https://digitaleconomy.stanford.edu/publication/canaries-in-the-coal-mine-six-facts-about-the-recent-employment-effects-of-artificial-intelligence},274 note = {verified: web (official Stanford Digital Economy Lab page fetched 2026-11-02; authors and title match; first version August 2025, revised version dated 2026 on the page)}275}276277@article{brynjolfsson2025generative,278 author = {Brynjolfsson, Erik and Li, Danielle and Raymond, Lindsey},279 title = {Generative AI at Work},280 journal = {The Quarterly Journal of Economics},281 year = {2025},282 volume = {140},283 number = {2},284 pages = {889--942},285 doi = {10.1093/qje/qjae044},286 url = {https://doi.org/10.1093/qje/qjae044},287 note = {verified: crossref 2026-11-02}288}289290@article{cameron2008bootstrap,291 author = {Cameron, A. Colin and Gelbach, Jonah B. and Miller, Douglas L.},292 title = {Bootstrap-Based Improvements for Inference with Clustered Errors},293 journal = {Review of Economics and Statistics},294 year = {2008},295 volume = {90},296 number = {3},297 pages = {414--427},298 doi = {10.1162/rest.90.3.414},299 url = {https://doi.org/10.1162/rest.90.3.414},300 note = {verified: crossref 2026-11-02}301}302303@article{cameron2015practitioner,304 author = {Colin Cameron, A. and Miller, Douglas L.},305 title = {A Practitioner’s Guide to Cluster-Robust Inference},306 journal = {Journal of Human Resources},307 year = {2015},308 volume = {50},309 number = {2},310 pages = {317--372},311 doi = {10.3368/jhr.50.2.317},312 url = {https://doi.org/10.3368/jhr.50.2.317},313 note = {verified: crossref 2026-11-02}314}315316@misc{catalini2026agi,317 author = {Catalini, Christian and Hui, Xiang and Wu, Jane},318 title = {Some Simple Economics of AGI},319 year = {2026},320 howpublished = {arXiv preprint arXiv:2602.20946},321 eprint = {2602.20946},322 archivePrefix = {arXiv},323 url = {https://arxiv.org/abs/2602.20946},324 note = {verified: arxiv 2602.20946}325}326327@techreport{cazzaniga2024genai,328 author = {Cazzaniga, Mauro and Jaumotte, Florence and Li, Longji and Melina, Giovanni and Panton, Augustus and Pizzinelli, Carlo and Rockall, Emma and Mendes Tavares, Marina},329 title = {Gen-AI: Artificial Intelligence and the Future of Work},330 year = {2024},331 number = {SDN/2024/001},332 institution = {International Monetary Fund},333 type = {IMF Staff Discussion Note},334 doi = {10.5089/9798400262548.006},335 url = {https://doi.org/10.5089/9798400262548.006},336 note = {verified: crossref 2026-11-02}337}338339@misc{cemri2025fail,340 author = {Cemri, Mert and Pan, Melissa Z. and Yang, Shuyi and Agrawal, Lakshya A. and Chopra, Bhavya and Tiwari, Rishabh and Keutzer, Kurt and Parameswaran, Aditya and Klein, Dan and Ramchandran, Kannan and Zaharia, Matei and Gonzalez, Joseph E. and Stoica, Ion},341 title = {Why Do Multi-Agent LLM Systems Fail?},342 year = {2025},343 howpublished = {arXiv preprint arXiv:2503.13657},344 eprint = {2503.13657},345 archivePrefix = {arXiv},346 url = {https://arxiv.org/abs/2503.13657},347 note = {verified: arxiv 2503.13657}348}349350@misc{chan2024mlebench,351 author = {Chan, Jun Shern and Chowdhury, Neil and Jaffe, Oliver and Aung, James and Sherburn, Dane and Mays, Evan and Starace, Giulio and Liu, Kevin and Maksin, Leon and Patwardhan, Tejal and Weng, Lilian and Mądry, Aleksander},352 title = {MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering},353 year = {2024},354 howpublished = {arXiv preprint arXiv:2410.07095},355 eprint = {2410.07095},356 archivePrefix = {arXiv},357 url = {https://arxiv.org/abs/2410.07095},358 note = {verified: arxiv 2410.07095}359}360361@misc{chen2023selfdebug,362 author = {Chen, Xinyun and Lin, Maxwell and Schärli, Nathanael and Zhou, Denny},363 title = {Teaching Large Language Models to Self-Debug},364 year = {2023},365 howpublished = {arXiv preprint arXiv:2304.05128},366 eprint = {2304.05128},367 archivePrefix = {arXiv},368 url = {https://arxiv.org/abs/2304.05128},369 note = {verified: arxiv 2304.05128}370}371372@misc{choi2023lawyering,373 author = {Choi, Jonathan H. and Monahan, Amy and Schwarcz, Daniel},374 title = {Lawyering in the Age of Artificial Intelligence},375 year = {2023},376 howpublished = {SSRN working paper},377 doi = {10.2139/ssrn.4626276},378 url = {https://doi.org/10.2139/ssrn.4626276},379 note = {verified: crossref 2026-11-02}380}381382@article{cox1972regression,383 author = {Cox, D. R.},384 title = {Regression Models and Life-Tables},385 journal = {Journal of the Royal Statistical Society Series B: Statistical Methodology},386 year = {1972},387 volume = {34},388 number = {2},389 pages = {187--202},390 doi = {10.1111/j.2517-6161.1972.tb00899.x},391 url = {https://doi.org/10.1111/j.2517-6161.1972.tb00899.x},392 note = {verified: crossref 2026-11-02}393}394395@misc{cui2024developers,396 author = {Cui, Zheyuan and Demirer, Mert and Jaffe, Sonia and Musolff, Leon and Peng, Sida and Salz, Tobias},397 title = {The Effects of Generative AI on High Skilled Work: Evidence from Three Field Experiments with Software Developers},398 year = {2024},399 howpublished = {SSRN working paper},400 doi = {10.2139/ssrn.4945566},401 url = {https://doi.org/10.2139/ssrn.4945566},402 note = {verified: crossref 2026-11-02}403}404405@techreport{dellacqua2023jagged,406 author = {Dell'Acqua, Fabrizio and McFowland III, Edward and Mollick, Ethan R. and Lifshitz-Assaf, Hila and Kellogg, Katherine and Rajendran, Saran and Krayer, Lisa and Candelon, François and Lakhani, Karim R.},407 title = {Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Productivity and Quality},408 year = {2023},409 number = {24-013},410 institution = {Harvard Business School},411 type = {Working Paper},412 doi = {10.2139/ssrn.4573321},413 url = {https://doi.org/10.2139/ssrn.4573321},414 note = {verified: crossref 2026-11-02. SSRN posted content.}415}416417@techreport{dellacqua2025cybernetic,418 author = {Dell'Acqua, Fabrizio and Ayoubi, Charles and Lifshitz-Assaf, Hila and Sadun, Raffaella and Mollick, Ethan R. and Mollick, Lilach and Han, Yi and Goldman, Jeff and Nair, Hari and Taub, Stew and Lakhani, Karim R.},419 title = {The Cybernetic Teammate: A Field Experiment on Generative AI Reshaping Teamwork and Expertise},420 year = {2025},421 number = {25-043},422 institution = {Harvard Business School},423 type = {Working Paper},424 doi = {10.2139/ssrn.5188231},425 url = {https://doi.org/10.2139/ssrn.5188231},426 note = {verified: crossref 2026-11-02. SSRN posted content.}427}428429@article{dietvorst2015aversion,430 author = {Dietvorst, Berkeley J. and Simmons, Joseph P. and Massey, Cade},431 title = {Algorithm aversion: People erroneously avoid algorithms after seeing them err.},432 journal = {Journal of Experimental Psychology: General},433 year = {2015},434 volume = {144},435 number = {1},436 pages = {114--126},437 doi = {10.1037/xge0000033},438 url = {https://doi.org/10.1037/xge0000033},439 note = {verified: crossref 2026-11-02}440}441442@misc{drouin2024workarena,443 author = {Drouin, Alexandre and Gasse, Maxime and Caccia, Massimo and Laradji, Issam H. and Del Verme, Manuel and Marty, Tom and Boisvert, Léo and Thakkar, Megh and Cappart, Quentin and Vazquez, David and Chapados, Nicolas and Lacoste, Alexandre},444 title = {WorkArena: How Capable Are Web Agents at Solving Common Knowledge Work Tasks?},445 year = {2024},446 howpublished = {arXiv preprint arXiv:2403.07718},447 eprint = {2403.07718},448 archivePrefix = {arXiv},449 url = {https://arxiv.org/abs/2403.07718},450 note = {verified: arxiv 2403.07718}451}452453@misc{dziri2023faith,454 author = {Dziri, Nouha and Lu, Ximing and Sclar, Melanie and Li, Xiang Lorraine and Jiang, Liwei and Lin, Bill Yuchen and West, Peter and Bhagavatula, Chandra and Bras, Ronan Le and Hwang, Jena D. and Sanyal, Soumya and Welleck, Sean and Ren, Xiang and Ettinger, Allyson and Harchaoui, Zaid and Choi, Yejin},455 title = {Faith and Fate: Limits of Transformers on Compositionality},456 year = {2023},457 howpublished = {arXiv preprint arXiv:2305.18654},458 eprint = {2305.18654},459 archivePrefix = {arXiv},460 url = {https://arxiv.org/abs/2305.18654},461 note = {verified: arxiv 2305.18654}462}463464@article{eloundou2024gpts,465 author = {Eloundou, Tyna and Manning, Sam and Mishkin, Pamela and Rock, Daniel},466 title = {GPTs are GPTs: Labor market impact potential of LLMs},467 journal = {Science},468 year = {2024},469 volume = {384},470 number = {6702},471 pages = {1306--1308},472 doi = {10.1126/science.adj0998},473 url = {https://doi.org/10.1126/science.adj0998},474 note = {verified: crossref 2026-11-02}475}476477@article{felten2021exposure,478 author = {Felten, Edward and Raj, Manav and Seamans, Robert},479 title = {Occupational, industry, and geographic exposure to artificial intelligence: A novel dataset and its potential uses},480 journal = {Strategic Management Journal},481 year = {2021},482 volume = {42},483 number = {12},484 pages = {2195--2217},485 doi = {10.1002/smj.3286},486 url = {https://doi.org/10.1002/smj.3286},487 note = {verified: crossref 2026-11-02}488}489490@misc{felten2023language,491 author = {Felten, Ed and Raj, Manav and Seamans, Robert},492 title = {How will Language Modelers like ChatGPT Affect Occupations and Industries?},493 year = {2023},494 howpublished = {arXiv preprint arXiv:2303.01157},495 eprint = {2303.01157},496 archivePrefix = {arXiv},497 url = {https://arxiv.org/abs/2303.01157},498 note = {verified: arxiv 2303.01157}499}500501@article{fine1999competing,502 author = {Fine, Jason P. and Gray, Robert J.},503 title = {A Proportional Hazards Model for the Subdistribution of a Competing Risk},504 journal = {Journal of the American Statistical Association},505 year = {1999},506 volume = {94},507 number = {446},508 pages = {496--509},509 doi = {10.1080/01621459.1999.10474144},510 url = {https://doi.org/10.1080/01621459.1999.10474144},511 note = {verified: crossref 2026-11-02}512}513514@article{frey2017future,515 author = {Frey, Carl Benedikt and Osborne, Michael A.},516 title = {The future of employment: How susceptible are jobs to computerisation?},517 journal = {Technological Forecasting and Social Change},518 year = {2017},519 volume = {114},520 pages = {254--280},521 doi = {10.1016/j.techfore.2016.08.019},522 url = {https://doi.org/10.1016/j.techfore.2016.08.019},523 note = {verified: crossref 2026-11-02}524}525526@techreport{gmyrek2023generative,527 author = {Gmyrek, Pawel and Berg, Janine and Bescond, David},528 title = {Generative AI and jobs: A global analysis of potential effects on job quantity and quality},529 year = {2023},530 number = {96},531 institution = {International Labour Organization},532 type = {ILO Working Paper},533 address = {Geneva},534 doi = {10.54394/FHEM8239},535 url = {https://doi.org/10.54394/FHEM8239},536 note = {verified: crossref 2026-11-02}537}538539@techreport{gmyrek2025refined,540 author = {Gmyrek, Pawel and Berg, Janine and Kamiński, Karol and Konopczyński, Filip and Ładna, Agnieszka and Nafradi, Balint and Rosłaniec, Konrad and Troszyński, Marek},541 title = {Generative AI and jobs: A refined global index of occupational exposure},542 year = {2025},543 number = {140},544 institution = {International Labour Organization},545 type = {ILO Working Paper},546 address = {Geneva},547 doi = {10.54394/HETP0387},548 url = {https://doi.org/10.54394/HETP0387},549 note = {verified: crossref 2026-11-02}550}551552@misc{gu2024survey,553 author = {Gu, Jiawei and Jiang, Xuhui and Shi, Zhichao and Tan, Hexiang and Zhai, Xuehao and Xu, Chengjin and Li, Wei and Shen, Yinghan and Ma, Shengjie and Liu, Honghao and Wang, Saizhuo and Zhang, Kun and Wang, Yuanzhuo and Gao, Wen and Ni, Lionel and Guo, Jian},554 title = {A Survey on LLM-as-a-Judge},555 year = {2024},556 howpublished = {arXiv preprint arXiv:2411.15594},557 eprint = {2411.15594},558 archivePrefix = {arXiv},559 url = {https://arxiv.org/abs/2411.15594},560 note = {verified: arxiv 2411.15594}561}562563@misc{guha2023legalbench,564 author = {Guha, Neel and Nyarko, Julian and Ho, Daniel E. and Ré, Christopher and Chilton, Adam and Narayana, Aditya and Chohlas-Wood, Alex and Peters, Austin and Waldon, Brandon and Rockmore, Daniel N. and Zambrano, Diego and Talisman, Dmitry and Hoque, Enam and Surani, Faiz and Fagan, Frank and Sarfaty, Galit and Dickinson, Gregory M. and Porat, Haggai and Hegland, Jason and Wu, Jessica and Nudell, Joe and Niklaus, Joel and Nay, John and Choi, Jonathan H. and Tobia, Kevin and Hagan, Margaret and Ma, Megan and Livermore, Michael and Rasumov-Rahe, Nikon and Holzenberger, Nils and Kolt, Noam and Henderson, Peter and Rehaag, Sean and Goel, Sharad and Gao, Shang and Williams, Spencer and Gandhi, Sunny and Zur, Tom and Iyer, Varun and Li, Zehua},565 title = {LegalBench: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Models},566 year = {2023},567 howpublished = {arXiv preprint arXiv:2308.11462},568 eprint = {2308.11462},569 archivePrefix = {arXiv},570 url = {https://arxiv.org/abs/2308.11462},571 note = {verified: arxiv 2308.11462}572}573574@article{gwet2008interrater,575 author = {Gwet, Kilem Li},576 title = {Computing inter‐rater reliability and its variance in the presence of high agreement},577 journal = {British Journal of Mathematical and Statistical Psychology},578 year = {2008},579 volume = {61},580 number = {1},581 pages = {29--48},582 doi = {10.1348/000711006X126600},583 url = {https://doi.org/10.1348/000711006X126600},584 note = {verified: crossref 2026-11-02}585}586587@misc{hadfield2025economy,588 author = {Hadfield, Gillian K. and Koh, Andrew},589 title = {An Economy of AI Agents},590 year = {2025},591 howpublished = {arXiv preprint arXiv:2509.01063},592 eprint = {2509.01063},593 archivePrefix = {arXiv},594 url = {https://arxiv.org/abs/2509.01063},595 note = {verified: arxiv 2509.01063}596}597598@techreport{hampole2025labor,599 author = {Hampole, Menaka and Papanikolaou, Dimitris and Schmidt, Lawrence and Seegmiller, Bryan},600 title = {Artificial Intelligence and the Labor Market},601 year = {2025},602 number = {33509},603 institution = {National Bureau of Economic Research},604 type = {NBER Working Paper},605 doi = {10.3386/w33509},606 url = {https://doi.org/10.3386/w33509},607 note = {verified: crossref 2026-11-02}608}609610@misc{handa2025economic,611 author = {Handa, Kunal and Tamkin, Alex and McCain, Miles and Huang, Saffron and Durmus, Esin and Heck, Sarah and Mueller, Jared and Hong, Jerry and Ritchie, Stuart and Belonax, Tim and Troy, Kevin K. and Amodei, Dario and Kaplan, Jared and Clark, Jack and Ganguli, Deep},612 title = {Which Economic Tasks are Performed with AI? Evidence from Millions of Claude Conversations},613 year = {2025},614 howpublished = {arXiv preprint arXiv:2503.04761},615 eprint = {2503.04761},616 archivePrefix = {arXiv},617 url = {https://arxiv.org/abs/2503.04761},618 note = {verified: arxiv 2503.04761}619}620621@article{hayes2007reliability,622 author = {Hayes, Andrew F. and Krippendorff, Klaus},623 title = {Answering the Call for a Standard Reliability Measure for Coding Data},624 journal = {Communication Methods and Measures},625 year = {2007},626 volume = {1},627 number = {1},628 pages = {77--89},629 doi = {10.1080/19312450709336664},630 url = {https://doi.org/10.1080/19312450709336664},631 note = {verified: crossref 2026-11-02}632}633634@article{holmstrom1979moral,635 author = {Holmstrom, Bengt},636 title = {Moral Hazard and Observability},637 journal = {The Bell Journal of Economics},638 year = {1979},639 volume = {10},640 number = {1},641 pages = {74},642 doi = {10.2307/3003320},643 url = {https://doi.org/10.2307/3003320},644 note = {verified: crossref 2026-11-02}645}646647@article{holmstrom1991multitask,648 author = {Holmstrom, Bengt and Milgrom, Paul},649 title = {Multitask Principal–Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design},650 journal = {The Journal of Law, Economics, and Organization},651 year = {1991},652 volume = {7},653 number = {special\_issue},654 pages = {24--52},655 doi = {10.1093/jleo/7.special_issue.24},656 url = {https://doi.org/10.1093/jleo/7.special_issue.24},657 note = {verified: crossref 2026-11-02}658}659660@techreport{horton2023silicus,661 author = {Horton, John and Filippas, Apostolos and Manning, Benjamin},662 title = {Large Language Models as Simulated Economic Agents: What Can We Learn from Homo Silicus?},663 year = {2023},664 number = {31122},665 institution = {National Bureau of Economic Research},666 type = {NBER Working Paper},667 doi = {10.3386/w31122},668 url = {https://doi.org/10.3386/w31122},669 note = {verified: crossref 2026-11-02}670}671672@misc{huang2023selfcorrect,673 author = {Huang, Jie and Chen, Xinyun and Mishra, Swaroop and Zheng, Huaixiu Steven and Yu, Adams Wei and Song, Xinying and Zhou, Denny},674 title = {Large Language Models Cannot Self-Correct Reasoning Yet},675 year = {2023},676 howpublished = {arXiv preprint arXiv:2310.01798},677 eprint = {2310.01798},678 archivePrefix = {arXiv},679 url = {https://arxiv.org/abs/2310.01798},680 note = {verified: arxiv 2310.01798}681}682683@techreport{humlum2025still,684 author = {Humlum, Anders and Vestergaard, Emilie},685 title = {Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI},686 year = {2025},687 number = {33777},688 institution = {National Bureau of Economic Research},689 type = {NBER Working Paper},690 doi = {10.3386/w33777},691 url = {https://doi.org/10.3386/w33777},692 note = {verified: crossref 2026-11-02. Previously circulated as "Large Language Models, Small Labor Market Effects" (May 2025).}693}694695@article{ide2025knowledge,696 author = {Ide, Enrique and Talamàs, Eduard},697 title = {Artificial Intelligence in the Knowledge Economy},698 journal = {Journal of Political Economy},699 year = {2025},700 volume = {133},701 number = {12},702 pages = {3762--3800},703 doi = {10.1086/737233},704 url = {https://doi.org/10.1086/737233},705 note = {verified: crossref 2026-11-02}706}707708@misc{immorlica2024agents,709 author = {Immorlica, Nicole and Lucier, Brendan and Slivkins, Aleksandrs},710 title = {Generative AI as Economic Agents},711 year = {2024},712 howpublished = {arXiv preprint arXiv:2406.00477},713 eprint = {2406.00477},714 archivePrefix = {arXiv},715 url = {https://arxiv.org/abs/2406.00477},716 note = {verified: arxiv 2406.00477}717}718719@misc{jimenez2024swebench,720 author = {Jimenez, Carlos E. and Yang, John and Wettig, Alexander and Yao, Shunyu and Pei, Kexin and Press, Ofir and Narasimhan, Karthik},721 title = {SWE-bench: Can Language Models Resolve Real-World GitHub Issues?},722 year = {2023},723 howpublished = {arXiv preprint arXiv:2310.06770},724 eprint = {2310.06770},725 archivePrefix = {arXiv},726 url = {https://arxiv.org/abs/2310.06770},727 note = {verified: arxiv 2310.06770}728}729730@article{kaplan1958nonparametric,731 author = {Kaplan, E. L. and Meier, Paul},732 title = {Nonparametric Estimation from Incomplete Observations},733 journal = {Journal of the American Statistical Association},734 year = {1958},735 volume = {53},736 number = {282},737 pages = {457--481},738 doi = {10.1080/01621459.1958.10501452},739 url = {https://doi.org/10.1080/01621459.1958.10501452},740 note = {verified: crossref 2026-11-02}741}742743@misc{kapoor2024matter,744 author = {Kapoor, Sayash and Stroebl, Benedikt and Siegel, Zachary S. and Nadgir, Nitya and Narayanan, Arvind},745 title = {AI Agents That Matter},746 year = {2024},747 howpublished = {arXiv preprint arXiv:2407.01502},748 eprint = {2407.01502},749 archivePrefix = {arXiv},750 url = {https://arxiv.org/abs/2407.01502},751 note = {verified: arxiv 2407.01502}752}753754@misc{kolt2025governing,755 author = {Kolt, Noam},756 title = {Governing AI Agents},757 year = {2025},758 howpublished = {arXiv preprint arXiv:2501.07913},759 eprint = {2501.07913},760 archivePrefix = {arXiv},761 url = {https://arxiv.org/abs/2501.07913},762 note = {verified: arxiv 2501.07913}763}764765@article{korinek2023generative,766 author = {Korinek, Anton},767 title = {Generative AI for Economic Research: Use Cases and Implications for Economists},768 journal = {Journal of Economic Literature},769 year = {2023},770 volume = {61},771 number = {4},772 pages = {1281--1317},773 doi = {10.1257/jel.20231736},774 url = {https://doi.org/10.1257/jel.20231736},775 note = {verified: crossref 2026-11-02}776}777778@article{krippendorff2004reliability,779 author = {Krippendorff, Klaus},780 title = {Reliability in Content Analysis: Some Common Misconceptions and Recommendations},781 journal = {Human Communication Research},782 year = {2004},783 volume = {30},784 number = {3},785 pages = {411--433},786 doi = {10.1111/j.1468-2958.2004.tb00738.x},787 url = {https://doi.org/10.1111/j.1468-2958.2004.tb00738.x},788 note = {verified: crossref 2026-11-02}789}790791@misc{kwa2025horizon,792 author = {Kwa, Thomas and West, Ben and Becker, Joel and Deng, Amy and Garcia, Katharyn and Hasin, Max and Jawhar, Sami and Kinniment, Megan and Rush, Nate and Von Arx, Sydney and Bloom, Ryan and Broadley, Thomas and Du, Haoxing and Goodrich, Brian and Jurkovic, Nikola and Miles, Luke Harold and Nix, Seraphina and Lin, Tao and Painter, Chris and Parikh, Neev and Rein, David and Sato, Lucas Jun Koba and Wijk, Hjalmar and Ziegler, Daniel M. and Barnes, Elizabeth and Chan, Lawrence},793 title = {Measuring AI Ability to Complete Long Software Tasks},794 year = {2025},795 howpublished = {arXiv preprint arXiv:2503.14499},796 eprint = {2503.14499},797 archivePrefix = {arXiv},798 url = {https://arxiv.org/abs/2503.14499},799 note = {verified: arxiv 2503.14499}800}801802@article{list2019multiple,803 author = {List, John A. and Shaikh, Azeem M. and Xu, Yang},804 title = {Multiple hypothesis testing in experimental economics},805 journal = {Experimental Economics},806 year = {2019},807 volume = {22},808 number = {4},809 pages = {773--793},810 doi = {10.1007/s10683-018-09597-5},811 url = {https://doi.org/10.1007/s10683-018-09597-5},812 note = {verified: crossref 2026-11-02}813}814815@misc{liu2024agentbench,816 author = {Liu, Xiao and Yu, Hao and Zhang, Hanchen and Xu, Yifan and Lei, Xuanyu and Lai, Hanyu and Gu, Yu and Ding, Hangliang and Men, Kaiwen and Yang, Kejuan and Zhang, Shudan and Deng, Xiang and Zeng, Aohan and Du, Zhengxiao and Zhang, Chenhui and Shen, Sheng and Zhang, Tianjun and Su, Yu and Sun, Huan and Huang, Minlie and Dong, Yuxiao and Tang, Jie},817 title = {AgentBench: Evaluating LLMs as Agents},818 year = {2023},819 howpublished = {arXiv preprint arXiv:2308.03688},820 eprint = {2308.03688},821 archivePrefix = {arXiv},822 url = {https://arxiv.org/abs/2308.03688},823 note = {verified: arxiv 2308.03688}824}825826@misc{liu2026claudecode,827 author = {Liu, Jiacheng and Zhao, Xiaohan and Shang, Xinyi and Shen, Zhiqiang},828 title = {Dive into Claude Code: The Design Space of Today's and Future AI Agent Systems},829 year = {2026},830 howpublished = {arXiv preprint arXiv:2604.14228},831 eprint = {2604.14228},832 archivePrefix = {arXiv},833 url = {https://arxiv.org/abs/2604.14228},834 note = {verified: arxiv 2604.14228}835}836837@misc{ma2024spreadsheetbench,838 author = {Ma, Zeyao and Zhang, Bohan and Zhang, Jing and Yu, Jifan and Zhang, Xiaokang and Zhang, Xiaohan and Luo, Sijia and Wang, Xi and Tang, Jie},839 title = {SpreadsheetBench: Towards Challenging Real World Spreadsheet Manipulation},840 year = {2024},841 howpublished = {arXiv preprint arXiv:2406.14991},842 eprint = {2406.14991},843 archivePrefix = {arXiv},844 url = {https://arxiv.org/abs/2406.14991},845 note = {verified: arxiv 2406.14991}846}847848@misc{madaan2023selfrefine,849 author = {Madaan, Aman and Tandon, Niket and Gupta, Prakhar and Hallinan, Skyler and Gao, Luyu and Wiegreffe, Sarah and Alon, Uri and Dziri, Nouha and Prabhumoye, Shrimai and Yang, Yiming and Gupta, Shashank and Majumder, Bodhisattwa Prasad and Hermann, Katherine and Welleck, Sean and Yazdanbakhsh, Amir and Clark, Peter},850 title = {Self-Refine: Iterative Refinement with Self-Feedback},851 year = {2023},852 howpublished = {arXiv preprint arXiv:2303.17651},853 eprint = {2303.17651},854 archivePrefix = {arXiv},855 url = {https://arxiv.org/abs/2303.17651},856 note = {verified: arxiv 2303.17651}857}858859@misc{merrill2026terminalbench,860 author = {Merrill, Mike A. and Shaw, Alexander G. and Carlini, Nicholas and Li, Boxuan and Raj, Harsh and Bercovich, Ivan and Shi, Lin and Shin, Jeong Yeon and Walshe, Thomas and Buchanan, E. Kelly and Shen, Junhong and Ye, Guanghao and Lin, Haowei and Poulos, Jason and Wang, Maoyu and Nezhurina, Marianna and Jitsev, Jenia and Lu, Di and Mastromichalakis, Orfeas Menis and Xu, Zhiwei and Chen, Zizhao and Liu, Yue and Zhang, Robert and Chen, Leon Liangyu and Kashyap, Anurag and Uslu, Jan-Lucas and Li, Jeffrey and Wu, Jianbo and Yan, Minghao and Bian, Song and Sharma, Vedang and Sun, Ke and Dillmann, Steven and Anand, Akshay and Lanpouthakoun, Andrew and Koopah, Bardia and Hu, Changran and Guha, Etash and Dreiman, Gabriel H. S. and Zhu, Jiacheng and Krauth, Karl and Zhong, Li and Muennighoff, Niklas and Amanfu, Robert and Tan, Shangyin and Pimpalgaonkar, Shreyas and Aggarwal, Tushar and Lin, Xiangning and Lan, Xin and Zhao, Xuandong and Liang, Yiqing and Wang, Yuanli and Wang, Zilong and Zhou, Changzhi and Heineman, David and Liu, Hange and Trivedi, Harsh and Yang, John and Lin, Junhong and Shetty, Manish and Yang, Michael and Omi, Nabil and Raoof, Negin and Li, Shanda and Zhuo, Terry Yue and Lin, Wuwei and Dai, Yiwei and Wang, Yuxin and Chai, Wenhao and Zhou, Shang and Wahdany, Dariush and She, Ziyu and Hu, Jiaming and Dong, Zhikang and Zhu, Yuxuan and Cui, Sasha and Saiyed, Ahson and Kolbeinsson, Arinbjörn and Hu, Jesse and Rytting, Christopher Michael and Marten, Ryan and Wang, Yixin and Dimakis, Alex and Konwinski, Andy and Schmidt, Ludwig},861 title = {Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces},862 year = {2026},863 howpublished = {arXiv preprint arXiv:2601.11868},864 eprint = {2601.11868},865 archivePrefix = {arXiv},866 url = {https://arxiv.org/abs/2601.11868},867 note = {verified: arxiv 2601.11868}868}869870@misc{mialon2023gaia,871 author = {Mialon, Grégoire and Fourrier, Clémentine and Swift, Craig and Wolf, Thomas and LeCun, Yann and Scialom, Thomas},872 title = {GAIA: a benchmark for General AI Assistants},873 year = {2023},874 howpublished = {arXiv preprint arXiv:2311.12983},875 eprint = {2311.12983},876 archivePrefix = {arXiv},877 url = {https://arxiv.org/abs/2311.12983},878 note = {verified: arxiv 2311.12983}879}880881@misc{miller2024errorbars,882 author = {Miller, Evan},883 title = {Adding Error Bars to Evals: A Statistical Approach to Language Model Evaluations},884 year = {2024},885 howpublished = {arXiv preprint arXiv:2411.00640},886 eprint = {2411.00640},887 archivePrefix = {arXiv},888 url = {https://arxiv.org/abs/2411.00640},889 note = {verified: arxiv 2411.00640}890}891892@misc{miserendino2025swelancer,893 author = {Miserendino, Samuel and Wang, Michele and Patwardhan, Tejal and Heidecke, Johannes},894 title = {SWE-Lancer: Can Frontier LLMs Earn \$1 Million from Real-World Freelance Software Engineering?},895 year = {2025},896 howpublished = {arXiv preprint arXiv:2502.12115},897 eprint = {2502.12115},898 archivePrefix = {arXiv},899 url = {https://arxiv.org/abs/2502.12115},900 note = {verified: arxiv 2502.12115}901}902903@article{nosek2018preregistration,904 author = {Nosek, Brian A. and Ebersole, Charles R. and DeHaven, Alexander C. and Mellor, David T.},905 title = {The preregistration revolution},906 journal = {Proceedings of the National Academy of Sciences},907 year = {2018},908 volume = {115},909 number = {11},910 pages = {2600--2606},911 doi = {10.1073/pnas.1708274114},912 url = {https://doi.org/10.1073/pnas.1708274114},913 note = {verified: crossref 2026-11-02}914}915916@article{noy2023experimental,917 author = {Noy, Shakked and Zhang, Whitney},918 title = {Experimental evidence on the productivity effects of generative artificial intelligence},919 journal = {Science},920 year = {2023},921 volume = {381},922 number = {6654},923 pages = {187--192},924 doi = {10.1126/science.adh2586},925 url = {https://doi.org/10.1126/science.adh2586},926 note = {verified: crossref 2026-11-02}927}928929@article{olken2015preanalysis,930 author = {Olken, Benjamin A.},931 title = {Promises and Perils of Pre-Analysis Plans},932 journal = {Journal of Economic Perspectives},933 year = {2015},934 volume = {29},935 number = {3},936 pages = {61--80},937 doi = {10.1257/jep.29.3.61},938 url = {https://doi.org/10.1257/jep.29.3.61},939 note = {verified: crossref 2026-11-02}940}941942@misc{otis2024entrepreneurial,943 author = {Otis, Nicholas and Clarke, Rowan and Delecourt, Solène and Holtz, David and Koning, Rembrand},944 title = {The Uneven Impact of Generative AI on Entrepreneurial Performance},945 year = {2024},946 howpublished = {SSRN working paper},947 doi = {10.2139/ssrn.4671369},948 url = {https://doi.org/10.2139/ssrn.4671369},949 note = {verified: crossref 2026-11-02}950}951952@misc{panickssery2024selfpreference,953 author = {Panickssery, Arjun and Bowman, Samuel R. and Feng, Shi},954 title = {LLM Evaluators Recognize and Favor Their Own Generations},955 year = {2024},956 howpublished = {arXiv preprint arXiv:2404.13076},957 eprint = {2404.13076},958 archivePrefix = {arXiv},959 url = {https://arxiv.org/abs/2404.13076},960 note = {verified: arxiv 2404.13076}961}962963@misc{patwardhan2025gdpval,964 author = {Patwardhan, Tejal and Dias, Rachel and Proehl, Elizabeth and Kim, Grace and Wang, Michele and Watkins, Olivia and Fishman, Simón Posada and Aljubeh, Marwan and Thacker, Phoebe and Fauconnet, Laurance and Kim, Natalie S. and Chao, Patrick and Miserendino, Samuel and Chabot, Gildas and Li, David and Sharman, Michael and Barr, Alexandra and Glaese, Amelia and Tworek, Jerry},965 title = {GDPval: Evaluating AI Model Performance on Real-World Economically Valuable Tasks},966 year = {2025},967 howpublished = {arXiv preprint arXiv:2510.04374},968 eprint = {2510.04374},969 archivePrefix = {arXiv},970 url = {https://arxiv.org/abs/2510.04374},971 note = {verified: arxiv 2510.04374}972}973974@misc{peng2023copilot,975 author = {Peng, Sida and Kalliamvakou, Eirini and Cihon, Peter and Demirer, Mert},976 title = {The Impact of AI on Developer Productivity: Evidence from GitHub Copilot},977 year = {2023},978 howpublished = {arXiv preprint arXiv:2302.06590},979 eprint = {2302.06590},980 archivePrefix = {arXiv},981 url = {https://arxiv.org/abs/2302.06590},982 note = {verified: arxiv 2302.06590}983}984985@techreport{pizzinelli2023exposure,986 author = {Pizzinelli, Carlo and Panton, Augustus and Mendes Tavares, Marina and Cazzaniga, Mauro and Li, Longji},987 title = {Labor Market Exposure to AI: Cross-country Differences and Distributional Implications},988 year = {2023},989 number = {WP/23/216},990 institution = {International Monetary Fund},991 type = {IMF Working Paper},992 doi = {10.5089/9798400254802.001},993 url = {https://doi.org/10.5089/9798400254802.001},994 note = {verified: crossref 2026-11-02}995}996997@misc{polo2024tinybenchmarks,998 author = {Polo, Felipe Maia and Weber, Lucas and Choshen, Leshem and Sun, Yuekai and Xu, Gongjun and Yurochkin, Mikhail},999 title = {tinyBenchmarks: evaluating LLMs with fewer examples},1000 year = {2024},1001 howpublished = {arXiv preprint arXiv:2402.14992},1002 eprint = {2402.14992},1003 archivePrefix = {arXiv},1004 url = {https://arxiv.org/abs/2402.14992},1005 note = {verified: arxiv 2402.14992}1006}10071008@misc{rein2025hcast,1009 author = {Rein, David and Becker, Joel and Deng, Amy and Nix, Seraphina and Canal, Chris and O'Connel, Daniel and Arnott, Pip and Bloom, Ryan and Broadley, Thomas and Garcia, Katharyn and Goodrich, Brian and Hasin, Max and Jawhar, Sami and Kinniment, Megan and Kwa, Thomas and Lajko, Aron and Rush, Nate and Sato, Lucas Jun Koba and Von Arx, Sydney and West, Ben and Chan, Lawrence and Barnes, Elizabeth},1010 title = {HCAST: Human-Calibrated Autonomy Software Tasks},1011 year = {2025},1012 howpublished = {arXiv preprint arXiv:2503.17354},1013 eprint = {2503.17354},1014 archivePrefix = {arXiv},1015 url = {https://arxiv.org/abs/2503.17354},1016 note = {verified: arxiv 2503.17354}1017}10181019@techreport{restrepo2025missed,1020 author = {Restrepo, Pascual},1021 title = {We Won't be Missed: Work and Growth in the AGI World},1022 year = {2025},1023 number = {34423},1024 institution = {National Bureau of Economic Research},1025 type = {NBER Working Paper},1026 doi = {10.3386/w34423},1027 url = {https://doi.org/10.3386/w34423},1028 note = {verified: crossref 2026-11-02}1029}10301031@misc{shinn2023reflexion,1032 author = {Shinn, Noah and Cassano, Federico and Berman, Edward and Gopinath, Ashwin and Narasimhan, Karthik and Yao, Shunyu},1033 title = {Reflexion: Language Agents with Verbal Reinforcement Learning},1034 year = {2023},1035 howpublished = {arXiv preprint arXiv:2303.11366},1036 eprint = {2303.11366},1037 archivePrefix = {arXiv},1038 url = {https://arxiv.org/abs/2303.11366},1039 note = {verified: arxiv 2303.11366}1040}10411042@article{simmons2011falsepositive,1043 author = {Simmons, Joseph P. and Nelson, Leif D. and Simonsohn, Uri},1044 title = {False-Positive Psychology},1045 journal = {Psychological Science},1046 year = {2011},1047 volume = {22},1048 number = {11},1049 pages = {1359--1366},1050 doi = {10.1177/0956797611417632},1051 url = {https://doi.org/10.1177/0956797611417632},1052 note = {verified: crossref 2026-11-02}1053}10541055@misc{sinha2025horizon,1056 author = {Sinha, Akshit and Arun, Arvindh and Goel, Shashwat and Staab, Steffen and Geiping, Jonas},1057 title = {The Illusion of Diminishing Returns: Measuring Long Horizon Execution in LLMs},1058 year = {2025},1059 howpublished = {arXiv preprint arXiv:2509.09677},1060 eprint = {2509.09677},1061 archivePrefix = {arXiv},1062 url = {https://arxiv.org/abs/2509.09677},1063 note = {verified: arxiv 2509.09677}1064}10651066@misc{sreenivas2026contracting,1067 author = {Sreenivas, Nanda Kishore and Larson, Kate},1068 title = {Contracting for LLM Delegation: Moral Hazard in Technology and Effort Choice},1069 year = {2026},1070 howpublished = {arXiv preprint arXiv:2608.18232},1071 eprint = {2608.18232},1072 archivePrefix = {arXiv},1073 url = {https://arxiv.org/abs/2608.18232},1074 note = {verified: arxiv 2608.18232}1075}10761077@misc{stroebl2024flaws,1078 author = {Stroebl, Benedikt and Kapoor, Sayash and Narayanan, Arvind},1079 title = {The Limits of Inference Scaling Through Resampling},1080 year = {2024},1081 howpublished = {arXiv preprint arXiv:2411.17501},1082 eprint = {2411.17501},1083 archivePrefix = {arXiv},1084 url = {https://arxiv.org/abs/2411.17501},1085 note = {verified: arxiv 2411.17501}1086}10871088@misc{svanberg2024costeffective,1089 author = {Svanberg, Maja and Li, Wensu and Fleming, Martin and Goehring, Brian and Thompson, Neil},1090 title = {Beyond AI Exposure: Which Tasks are Cost-Effective to Automate with Computer Vision?},1091 year = {2024},1092 howpublished = {SSRN working paper},1093 doi = {10.2139/ssrn.4700751},1094 url = {https://doi.org/10.2139/ssrn.4700751},1095 note = {verified: crossref 2026-11-02}1096}10971098@misc{tomlinson2025working,1099 author = {Tomlinson, Kiran and Jaffe, Sonia and Wang, Will and Counts, Scott and Suri, Siddharth},1100 title = {Working with AI: Measuring the Applicability of Generative AI to Occupations},1101 year = {2025},1102 howpublished = {arXiv preprint arXiv:2507.07935},1103 eprint = {2507.07935},1104 archivePrefix = {arXiv},1105 url = {https://arxiv.org/abs/2507.07935},1106 note = {verified: arxiv 2507.07935}1107}11081109@inproceedings{trivedi2024appworld,1110 author = {Trivedi, Harsh and Khot, Tushar and Hartmann, Mareike and Manku, Ruskin and Dong, Vinty and Li, Edward and Gupta, Shashank and Sabharwal, Ashish and Balasubramanian, Niranjan},1111 title = {AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents},1112 booktitle = {Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},1113 year = {2024},1114 doi = {10.18653/v1/2024.acl-long.850},1115 url = {https://doi.org/10.18653/v1/2024.acl-long.850},1116 note = {verified: crossref 2026-11-02}1117}11181119@article{vaccaro2024combinations,1120 author = {Vaccaro, Michelle and Almaatouq, Abdullah and Malone, Thomas},1121 title = {When combinations of humans and AI are useful: A systematic review and meta-analysis},1122 journal = {Nature Human Behaviour},1123 year = {2024},1124 volume = {8},1125 number = {12},1126 pages = {2293--2303},1127 doi = {10.1038/s41562-024-02024-1},1128 url = {https://doi.org/10.1038/s41562-024-02024-1},1129 note = {verified: crossref 2026-11-02}1130}11311132@inproceedings{vania2021irt,1133 author = {Vania, Clara and Htut, Phu Mon and Huang, William and Mungra, Dhara and Pang, Richard Yuanzhe and Phang, Jason and Liu, Haokun and Cho, Kyunghyun and Bowman, Samuel R.},1134 title = {Comparing Test Sets with Item Response Theory},1135 booktitle = {Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)},1136 year = {2021},1137 doi = {10.18653/v1/2021.acl-long.92},1138 url = {https://doi.org/10.18653/v1/2021.acl-long.92},1139 note = {verified: crossref 2026-11-02}1140}11411142@inproceedings{wang2024fair,1143 author = {Wang, Peiyi and Li, Lei and Chen, Liang and Cai, Zefan and Zhu, Dawei and Lin, Binghuai and Cao, Yunbo and Kong, Lingpeng and Liu, Qi and Liu, Tianyu and Sui, Zhifang},1144 title = {Large Language Models are not Fair Evaluators},1145 booktitle = {Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},1146 year = {2024},1147 doi = {10.18653/v1/2024.acl-long.511},1148 url = {https://doi.org/10.18653/v1/2024.acl-long.511},1149 note = {verified: crossref 2026-11-02}1150}11511152@misc{wang2024officebench,1153 author = {Wang, Zilong and Cui, Yuedong and Zhong, Li and Zhang, Zimin and Yin, Da and Lin, Bill Yuchen and Shang, Jingbo},1154 title = {OfficeBench: Benchmarking Language Agents across Multiple Applications for Office Automation},1155 year = {2024},1156 howpublished = {arXiv preprint arXiv:2407.19056},1157 eprint = {2407.19056},1158 archivePrefix = {arXiv},1159 url = {https://arxiv.org/abs/2407.19056},1160 note = {verified: arxiv 2407.19056}1161}11621163@misc{wang2024openhands,1164 author = {Wang, Xingyao and Li, Boxuan and Song, Yufan and Xu, Frank F. and Tang, Xiangru and Zhuge, Mingchen and Pan, Jiayi and Song, Yueqi and Li, Bowen and Singh, Jaskirat and Tran, Hoang H. and Li, Fuqiang and Ma, Ren and Zheng, Mingzhang and Qian, Bill and Shao, Yanjun and Muennighoff, Niklas and Zhang, Yizhe and Hui, Binyuan and Lin, Junyang and Brennan, Robert and Peng, Hao and Ji, Heng and Neubig, Graham},1165 title = {OpenHands: An Open Platform for AI Software Developers as Generalist Agents},1166 year = {2024},1167 howpublished = {arXiv preprint arXiv:2407.16741},1168 eprint = {2407.16741},1169 archivePrefix = {arXiv},1170 url = {https://arxiv.org/abs/2407.16741},1171 note = {verified: arxiv 2407.16741}1172}11731174@misc{wijk2024rebench,1175 author = {Wijk, Hjalmar and Lin, Tao and Becker, Joel and Jawhar, Sami and Parikh, Neev and Broadley, Thomas and Chan, Lawrence and Chen, Michael and Clymer, Josh and Dhyani, Jai and Ericheva, Elena and Garcia, Katharyn and Goodrich, Brian and Jurkovic, Nikola and Karnofsky, Holden and Kinniment, Megan and Lajko, Aron and Nix, Seraphina and Sato, Lucas and Saunders, William and Taran, Maksym and West, Ben and Barnes, Elizabeth},1176 title = {RE-Bench: Evaluating frontier AI R\&D capabilities of language model agents against human experts},1177 year = {2024},1178 howpublished = {arXiv preprint arXiv:2411.15114},1179 eprint = {2411.15114},1180 archivePrefix = {arXiv},1181 url = {https://arxiv.org/abs/2411.15114},1182 note = {verified: arxiv 2411.15114}1183}11841185@misc{xie2024osworld,1186 author = {Xie, Tianbao and Zhang, Danyang and Chen, Jixuan and Li, Xiaochuan and Zhao, Siheng and Cao, Ruisheng and Hua, Toh Jing and Cheng, Zhoujun and Shin, Dongchan and Lei, Fangyu and Liu, Yitao and Xu, Yiheng and Zhou, Shuyan and Savarese, Silvio and Xiong, Caiming and Zhong, Victor and Yu, Tao},1187 title = {OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments},1188 year = {2024},1189 howpublished = {arXiv preprint arXiv:2404.07972},1190 eprint = {2404.07972},1191 archivePrefix = {arXiv},1192 url = {https://arxiv.org/abs/2404.07972},1193 note = {verified: arxiv 2404.07972}1194}11951196@misc{xu2024agentcompany,1197 author = {Xu, Frank F. and Song, Yufan and Li, Boxuan and Tang, Yuxuan and Jain, Kritanjali and Bao, Mengxue and Wang, Zora Z. and Zhou, Xuhui and Guo, Zhitong and Cao, Murong and Yang, Mingyang and Lu, Hao Yang and Martin, Amaad and Su, Zhe and Maben, Leander and Mehta, Raj and Chi, Wayne and Jang, Lawrence and Xie, Yiqing and Zhou, Shuyan and Neubig, Graham},1198 title = {TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks},1199 year = {2024},1200 howpublished = {arXiv preprint arXiv:2412.14161},1201 eprint = {2412.14161},1202 archivePrefix = {arXiv},1203 url = {https://arxiv.org/abs/2412.14161},1204 note = {verified: arxiv 2412.14161}1205}12061207@misc{yang2024sweagent,1208 author = {Yang, John and Jimenez, Carlos E. and Wettig, Alexander and Lieret, Kilian and Yao, Shunyu and Narasimhan, Karthik and Press, Ofir},1209 title = {SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering},1210 year = {2024},1211 howpublished = {arXiv preprint arXiv:2405.15793},1212 eprint = {2405.15793},1213 archivePrefix = {arXiv},1214 url = {https://arxiv.org/abs/2405.15793},1215 note = {verified: arxiv 2405.15793}1216}12171218@misc{yao2024taubench,1219 author = {Yao, Shunyu and Shinn, Noah and Razavi, Pedram and Narasimhan, Karthik},1220 title = {$\tau$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains},1221 year = {2024},1222 howpublished = {arXiv preprint arXiv:2406.12045},1223 eprint = {2406.12045},1224 archivePrefix = {arXiv},1225 url = {https://arxiv.org/abs/2406.12045},1226 note = {verified: arxiv 2406.12045}1227}12281229@misc{yuan2026osworld2,1230 author = {Yuan, Mengqi and Zhou, Zilong and Xiong, Xinzhuang and Wu, Weiming and Sun, Jiayang and Song, Jiamin and Cui, Kaiqian and Wang, Bowen and Wu, Haoyuan and Li, Yitong and Lu, Dunjie and Lu, Haikong and Zhen, Qi and Wang, Xinyuan and Deng, Jiaqi and Yang, Yuhao and Chen, Cheng and Zheng, Boyuan and Su, Alex and Yu, Xiao and Zou, Hao and Agashe, Saaket and Lu, Xing Han and Kaur, Manpreet and Qi, Zhengyang and Chen, Vincent Sunn and Sala, Frederic and Liu, Dayiheng and Lin, Junyang and Yu, Zhou and Su, Yu and Reddy, Siva and Wang, Xin Eric and Qi, Peng and Xie, Tianbao and Yu, Tao},1231 title = {OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks},1232 year = {2026},1233 howpublished = {arXiv preprint arXiv:2606.29537},1234 eprint = {2606.29537},1235 archivePrefix = {arXiv},1236 url = {https://arxiv.org/abs/2606.29537},1237 note = {verified: arxiv 2606.29537}1238}12391240@misc{zheng2023judge,1241 author = {Zheng, Lianmin and Chiang, Wei-Lin and Sheng, Ying and Zhuang, Siyuan and Wu, Zhanghao and Zhuang, Yonghao and Lin, Zi and Li, Zhuohan and Li, Dacheng and Xing, Eric P. and Zhang, Hao and Gonzalez, Joseph E. and Stoica, Ion},1242 title = {Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena},1243 year = {2023},1244 howpublished = {arXiv preprint arXiv:2306.05685},1245 eprint = {2306.05685},1246 archivePrefix = {arXiv},1247 url = {https://arxiv.org/abs/2306.05685},1248 note = {verified: arxiv 2306.05685}1249}12501251@misc{zhou2024webarena,1252 author = {Zhou, Shuyan and Xu, Frank F. and Zhu, Hao and Zhou, Xuhui and Lo, Robert and Sridhar, Abishek and Cheng, Xianyi and Ou, Tianyue and Bisk, Yonatan and Fried, Daniel and Alon, Uri and Neubig, Graham},1253 title = {WebArena: A Realistic Web Environment for Building Autonomous Agents},1254 year = {2023},1255 howpublished = {arXiv preprint arXiv:2307.13854},1256 eprint = {2307.13854},1257 archivePrefix = {arXiv},1258 url = {https://arxiv.org/abs/2307.13854},1259 note = {verified: arxiv 2307.13854}1260}12611262@misc{zhou2025lost,1263 author = {Zhou, Hongli and Huang, Hui and Zhao, Ziqing and Han, Lvyuan and Wang, Huicheng and Chen, Kehai and Yang, Muyun and Bao, Wei and Dong, Jian and Xu, Bing and Zhu, Conghui and Cao, Hailong and Zhao, Tiejun},1264 title = {Lost in Benchmarks? Rethinking Large Language Model Benchmarking with Item Response Theory},1265 year = {2025},1266 howpublished = {arXiv preprint arXiv:2505.15055},1267 eprint = {2505.15055},1268 archivePrefix = {arXiv},1269 url = {https://arxiv.org/abs/2505.15055},1270 note = {verified: arxiv 2505.15055}1271}12721273@misc{zhu2025rigorous,1274 author = {Zhu, Yuxuan and Jin, Tengjun and Pruksachatkun, Yada and Zhang, Andy and Liu, Shu and Cui, Sasha and Kapoor, Sayash and Longpre, Shayne and Meng, Kevin and Weiss, Rebecca and Barez, Fazl and Gupta, Rahul and Dhamala, Jwala and Merizian, Jacob and Giulianelli, Mario and Coppock, Harry and Ududec, Cozmin and Sekhon, Jasjeet and Steinhardt, Jacob and Kellermann, Antony and Schwettmann, Sarah and Zaharia, Matei and Stoica, Ion and Liang, Percy and Kang, Daniel},1275 title = {Establishing Best Practices for Building Rigorous Agentic Benchmarks},1276 year = {2025},1277 howpublished = {arXiv preprint arXiv:2507.02825},1278 eprint = {2507.02825},1279 archivePrefix = {arXiv},1280 url = {https://arxiv.org/abs/2507.02825},1281 note = {verified: arxiv 2507.02825}1282}1283