# Rodney Brooks argues deployment at scale takes decades [Canonical reading](https://escapemeridian.com/readings/rodney-brooks-argues-deployment-at-scale-takes-decades-9fff163c) Rodney Brooks published this piece on April 17, 2025. He argues that technologies take longer than expected to reach practical deployment, so research needs sustained, steady funding. He traces the point through his early work on visual map making for mobile robots. Brooks says he and colleagues developed loop closing, which he describes as the basis for SLAM (Simultaneous Localization and Mapping). He says their first geometric inference methods were inadequate. Later work made the idea practical-ish, while further development took decades, he says. Brooks says it became practical at large scale with cheap GPUs, alongside demand for deep learning inference. He connects that progress to the possibility of his current company. His closing warning is pointed: when a barely working lab demo depends on PhD students babysitting it, he says, laugh at a claim that it will change the world in two years. The example supports Brooks’s argument that a promising demonstration is not the same as deployment at scale. He gives no details about his current company’s applications or the funding needed to move other technologies into deployment. Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims. ## Original source Rodney Brooks · Article Original source date: Apr 17, 2025 (2025-04-17T05:29:57.000Z). Date basis: Date supplied by the publisher’s feed. Added to Escape Meridian: 2026-11-14T13:33:24.815Z. First observed: 2026-10-23T17:53:53.260Z. Retrieved: 2026-10-23T17:55:42.643Z. - [Deployment at scale takes decades](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Limits Text as retrieved on October 23, 2026; originally published April 17, 2025. Brooks does not give details about his current company’s applications or specify what sustained funding would be needed for other technologies to reach deployment at scale. ## Open question What sustained funding and further work would be needed for other technologies to move from a barely working lab demo to deployment at scale? ## Source-stated Brooks argues that deployment takes longer than people expect and calls for sustained, steady funding. > The short answer is that it takes longer than anyone thinks to get to deployment, and so sustained steady funding is necessary. - [Supporting source](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Source-stated Brooks says loop closing is the basis for SLAM and recounts that early geometric inference methods were inadequate. > We were both surprised by the others’ paper as we had invented the same thing, loop closing, which is the basis for SLAM (Simultaneous Localization and Mapping). - [Supporting source](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Source-stated Brooks says work on making the idea practical continued for decades, and that it is now practical at large scale with cheap GPUs. > For the next 20 years, hundreds (certainly, perhaps more) people worked on really making it practical. But still, early this century, my hopeful “visual” map making was not really practical, 1-D Lidar worked-ish. Now with cheap GPUs by the hundreds, on cheap silicon, driven by the need for deep - [Supporting source](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Source-stated Brooks warns readers to laugh at claims that a barely working lab demo will change the world in two years. > If you see someone with a new technology that is a barely working lab demo with PhD students baby-sitting it behind the scenes, and the they say it is going to change the world in two years, just laugh. - [Supporting source](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Source and contributor context Rodney Brooks · Publisher The publisher is identified here. No individual author has been inferred. ## Source media Media retrieval status: unchecked. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Graphics processing unit A chip that performs many calculations at once; the main hardware used to train and run AI models. - [Definition reference](https://en.wikipedia.org/wiki/Graphics_processing_unit) ## Deep learning Machine learning that uses many layers of simple calculations; the basis of most current AI. - [Definition reference](https://en.wikipedia.org/wiki/Deep_learning) ## The World The World is a public radio news program publishing international reporting on politics, society, labor and technology. - [Explore this context](/entities/the-world) - [Reference](https://theworld.org/) ## Editorial provenance Version 1. Publication ID: 31aa3f24-e005-b5fc-85c4-ea2af1ed780c. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-14T13:29:04.336Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: 5d5cfdd60fb6d499c89efa95fd84eb07171b944f23245776c187efdd82541fb3. Source text hash: 46c1fbeef9cc214db02cd7e6237ae9961aab3e1dbb0124451bb7378a2104a018. ## Revision and source-check history ### Version 1 — 2026-11-14T13:33:24.815Z Rodney Brooks published this piece on April 17, 2025. He argues that technologies take longer than expected to reach practical deployment, so research needs sustained, steady funding. He traces the point through his early work on visual map making for mobile robots. Brooks says he and colleagues developed loop closing, which he describes as the basis for SLAM (Simultaneous Localization and Mapping). He says their first geometric inference methods were inadequate. Later work made the idea practical-ish, while further development took decades, he says. Brooks says it became practical at large scale with cheap GPUs, alongside demand for deep learning inference. He connects that progress to the possibility of his current company. His closing warning is pointed: when a barely working lab demo depends on PhD students babysitting it, he says, laugh at a claim that it will change the world in two years. The example supports Brooks’s argument that a promising demonstration is not the same as deployment at scale. He gives no details about his current company’s applications or the funding needed to move other technologies into deployment. ## Original source Rodney Brooks · Article Original source date: Apr 17, 2025 (2025-04-17T05:29:57.000Z). Date basis: Date supplied by the publisher’s feed. Added to Escape Meridian: 2026-11-14T13:33:24.815Z. First observed: 2026-10-23T17:53:53.260Z. Retrieved: 2026-10-23T17:55:42.643Z. - [Deployment at scale takes decades](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Limits Text as retrieved on October 23, 2026; originally published April 17, 2025. Brooks does not give details about his current company’s applications or specify what sustained funding would be needed for other technologies to reach deployment at scale. ## Open question What sustained funding and further work would be needed for other technologies to move from a barely working lab demo to deployment at scale? ## Source-stated Brooks argues that deployment takes longer than people expect and calls for sustained, steady funding. > The short answer is that it takes longer than anyone thinks to get to deployment, and so sustained steady funding is necessary. - [Supporting source](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Source-stated Brooks says loop closing is the basis for SLAM and recounts that early geometric inference methods were inadequate. > We were both surprised by the others’ paper as we had invented the same thing, loop closing, which is the basis for SLAM (Simultaneous Localization and Mapping). - [Supporting source](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Source-stated Brooks says work on making the idea practical continued for decades, and that it is now practical at large scale with cheap GPUs. > For the next 20 years, hundreds (certainly, perhaps more) people worked on really making it practical. But still, early this century, my hopeful “visual” map making was not really practical, 1-D Lidar worked-ish. Now with cheap GPUs by the hundreds, on cheap silicon, driven by the need for deep - [Supporting source](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Source-stated Brooks warns readers to laugh at claims that a barely working lab demo will change the world in two years. > If you see someone with a new technology that is a barely working lab demo with PhD students baby-sitting it behind the scenes, and the they say it is going to change the world in two years, just laugh. - [Supporting source](https://rodneybrooks.com/deployment-at-scale-takes-decades/) ## Source and contributor context Rodney Brooks · Publisher The publisher is identified here. No individual author has been inferred. ## Source media Media retrieval status: unchecked. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Graphics processing unit A chip that performs many calculations at once; the main hardware used to train and run AI models. - [Definition reference](https://en.wikipedia.org/wiki/Graphics_processing_unit) ## Deep learning Machine learning that uses many layers of simple calculations; the basis of most current AI. - [Definition reference](https://en.wikipedia.org/wiki/Deep_learning) ## The World The World is a public radio news program publishing international reporting on politics, society, labor and technology. - [Explore this context](/entities/the-world) - [Reference](https://theworld.org/) ## Editorial provenance Version 1. Publication ID: 31aa3f24-e005-b5fc-85c4-ea2af1ed780c. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-14T13:29:04.336Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: 5d5cfdd60fb6d499c89efa95fd84eb07171b944f23245776c187efdd82541fb3. Source text hash: 46c1fbeef9cc214db02cd7e6237ae9961aab3e1dbb0124451bb7378a2104a018. ### uncertain — 2026-11-14T13:29:04.336Z — anthropic/claude-opus-5.5 Publication: 31aa3f24-e005-b5fc-85c4-ea2af1ed780c attribution — pass: This is Rodney Brooks's own blog essay. The reading consistently attributes the argument, the history and the closing warning to Brooks. > I published an invited, unrefereed, paper “Visual Map Making for a Mobile Robot”. summary — uncertain: Loop closing is the central example, but the reading never says in plain language what it is. Brooks explains it as a robot recognizing a place it has been before and correcting its drifted position estimates. The reading also omits the 1985 starting point and Brooks's claim that SLAM now underpins drones, agricultural robots and mobile robots. That omission weakens the stakes of the decades-long arc. 'He and colleagues developed' blurs that Brooks and the Toulouse team invented it independently. Smallest repair: add a one-clause loop-closing gloss, the 1985 date and the current applications. > then it recognizes it is at a place it was in before and significantly corrects its estimate of where it must have been for all its intermediate observations date — pass: The reading's April 17, 2025 date matches the UTC calendar day of evidence.originalPublished. The reading discloses the later retrieval date. It uses no hindsight framing or later material. support — pass: Each claim matches the source: steady funding, loop closing as the basis for SLAM, inadequate early inference, and roughly 20 more years of work. The reading also matches Brooks on GPUs made cheap by demand for deep learning inference, his company and the laugh warning. The question follows from the source and does not presuppose an answer. > Now with cheap GPUs by the hundreds, on cheap silicon, driven by the need for deep learning inference, it is practical at large scale. media — pass: No media was selected, and no verified candidates were supplied. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/9fff163cae628280cea90fe6148f34faef741b6883815bd07a96dc191aa55c41/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/rodney-brooks-argues-deployment-at-scale-takes-decades-9fff163c.json)