![](\"https:\/\/fortune.com\/img-assets\/wp-content\/uploads\/2025\/12\/54971199697_468678389e_o-e1765423303599.jpg?w=2048\")
<\/p>\n<\/p>\n
If a team of human engineers built a web browser that only half-worked, it wouldn\u2019t get people talking. But when Michael Truell, CEO of coding startup Cursor, posted on X last week that a swarm of AI agents had built a browser that, he wrote, \u201ckind of works\u201d\u2014while running uninterrupted for a week without any human intervention\u2014it went viral across the tech world, with over 6 million views.\u00a0<\/p>\n
Why the buzz? Two big reasons: For one thing, AI\u2019s attention span has historically been short. In the early days of ChatGPT, models could stay on task for only a few seconds. That horizon stretched to minutes for better models, then to hours. The Cursor project claims to be one of the first times an AI system has sustained a complex, open-ended software project for an entire week without human guidance.<\/p>\n
In addition, single AI agents are limited to focused, small tasks. But getting hundreds of agents to coordinate on a big project has still seemed futuristic. That\u2019s why Cursor wanted to see how far they could push autonomous coding\u2014on a project that could take months for a human team\u2014by having an \u201corchestra\u201d of AI agents working as a team. Could an AI system be persistent enough, and work together well enough, to explore code, break work into parts, debug itself, and keep moving forward for days without drifting away from the task at hand?\u00a0<\/p>\n
The researchers found that the answer was mostly yes. Cursor\u2019s experiment orchestrated hundreds of agents into something like a software team. It had \u201cplanners,\u201d \u201cworkers,\u201d and \u201cjudges\u201d coordinating across millions of lines of code. This hints at what both Cursor and OpenAI say is a near future in which AI doesn\u2019t just assist employees, but takes on entire projects. That would fundamentally reshape how complex work gets done\u2014first in software development, but then in other professions.<\/p>\n
There have been AI swarm experiments for a couple of years now. But today, Cursor says, models are smarter and can stay coherent for much longer. The models can be run at a far larger scale, with a custom layer that orchestrates hundreds of agents and keeps them from descending into chaos.\u00a0<\/p>\n
Jonas Nelle, an engineer at Cursor working on long-running AI agents, told Fortune <\/em>that as AI models keep getting better, engineers and researchers need to revisit their assumptions every few months about what the AI models can do. While he admitted he \u201cwouldn\u2019t download it and delete Chrome today,\u201d the browser project was \u201ccertainly better than anything models previously would have been able to do.\u201d\u00a0<\/p>\n These long-running agents are an important frontier, added Bill Chen, an OpenAI engineer who stress-tests and evaluates the real-world behavior of the company\u2019s models. The length of a task, and the fact that an AI system can accomplish the task autonomously and coherently is a \u201cvery good indicator of how intelligent and how general a system is,\u201d he said. The Cursor project, which was powered by OpenAI\u2019s GPT-5.2, is \u201ca direct result of us really continuously pushing forward the boundaries of model capabilities.\u201d In the future, he said, there will be even longer horizon tests.\u00a0\u00a0<\/p>\n Still, these are not production-ready systems. Besides being buggy and incomplete, a project running swarms of agents for days or weeks is expensive. While prices have fallen steeply over the past year, long-running jobs with hundreds of AI agents can still rack up costs. <\/p>\n There are also security issues. An autonomous system raises worries about vulnerabilities, data leaks, and much more, and requires many new layers of control and auditability.<\/p>\n But Chen said he foresees a near future where something like this could be ready \u201cfor broad consumption and at a not prohibitive cost. Progress has been continuous so far, he explained, and there have been important unlocks every step of the way. For now, he said, the excitement is driven by the fact that this is a real, practical example of model capability, \u201cversus how this model performs on academic and public evaluations and benchmarks.\u201d<\/p>\n The shift has surprised even longtime AI observers. In a recent post, independent researcher Simon Willison predicted that by 2029, someone would build a full web browser largely using AI\u2014and that it wouldn\u2019t even be surprising. \u201cRolling a new web browser is one of the most complicated software projects I can imagine,\u201d he wrote. Cursor may have accelerated that timeline. \u201cI may have been off by three years,\u201d Willison said. \u201cI have to admit I\u2019m very surprised to see something this capable emerge so quickly.\u201d<\/p>\n This speaks to what OpenAI and others have talked about as a \u201ccapabilities overhang\u201d\u2014the idea that the most sophisticated AI models can do much more than what\u2019s publicly deployed, but the right combination of tools, product design, and drops in cost can suddenly make them usable at scale. So while tools like the Cursor browser aren\u2019t quite ready for primetime, the trajectory is clear.\u00a0<\/p>\n This story was originally featured on Fortune.com<\/p>\n #Cursor #swarm #agents #powered #OpenAI #build #run #web #browser #weekwith #human #Heres #developers #buzzing<\/p>\n","protected":false},"excerpt":{"rendered":" If a team of human engineers b… <\/p>\n","protected":false},"author":1,"featured_media":2657,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[2],"tags":[2083,5539,82,10238,1827,1578,2867,3595,2905,703,10236,65,10234,10235,4253,10237],"_links":{"self":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/posts\/15207"}],"collection":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=15207"}],"version-history":[{"count":0,"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/posts\/15207\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/media\/2657"}],"wp:attachment":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=15207"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=15207"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=15207"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}AI agent swarms are not ready for business use<\/h2>\n