# MIT Technology Review’s The Download reports on Manus and Waabi’s virtual robotrucks [Canonical reading](https://escapemeridian.com/readings/mit-technology-review-s-the-download-reports-on-manus-and-waabi-e034f370) On March 12, 2025, MIT Technology Review’s The Download reports that its staff obtained access to Manus, an AI agent launched the previous week, but this newsletter excerpt does not say what its test found. It says Manus claims to be the world’s first general AI agent and can act autonomously across a wide range of tasks. The distinction matters: the source reports the company’s description and the reviewers’ access, not the test’s results. The newsletter also reports Waabi’s claim that its virtual simulation can demonstrate the safety of its driverless trucks without running them for miles on real roads. Waabi says its digital twin—a virtual version of its trucks—uses real sensor data and matches the trucks’ performance on real roads almost exactly. The newsletter presents this as Waabi’s argument for simulation over accumulating real-world miles; it gives no independent assessment of that comparison. Other items describe a University of Cambridge team’s artificial leaf producing ethylene and ethane, and Boston Metal running its largest reactor yet to make steel. The newsletter says Boston Metal still has milestones to reach before its process operates at a scale that could affect the steel industry. The Manus test results, and the basis for Waabi’s safety claim, remain unspecified in this excerpt. 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 MIT Technology Review · Article Original source date: Mar 12, 2025 (2025-03-12T12:10:00Z). Date basis: Date from the publisher’s own records, through its API. Added to Escape Meridian: 2026-11-14T22:28:23.593Z. First observed: 2026-10-25T10:36:04.664Z. Retrieved: 2026-10-25T10:37:49.951Z. - [The Download: testing new AI agent Manus, and Waabi’s virtual robotruck ambitions](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Limits Text as retrieved on October 25, 2026; originally published March 12, 2025. The excerpt does not give the findings of MIT Technology Review’s Manus test. The excerpt gives no independent assessment or detailed evidence for Waabi’s safety claim. ## Open question What did MIT Technology Review’s test of Manus find, and what evidence does Waabi use to support its claim that its simulation can demonstrate truck safety? ## Source-stated MIT Technology Review says it obtained access to Manus, but this newsletter excerpt does not provide the test’s findings. > MIT Technology Review was able to obtain access to Manus. Here’s what we made of it. - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated The newsletter says Manus claims to be the world’s first general AI agent and to act autonomously across a wide range of tasks. > Manus claims to be the world's first general AI agent, building off multiple AI models and agents to act autonomously on a wide range of tasks. - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated Waabi says its virtual simulation is accurate enough to demonstrate the safety of its driverless trucks without running them for miles on real roads. > Waabi says its super-realistic virtual simulation is now accurate enough to prove the safety of its driverless big rigs without having to run them for miles on real roads. - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated Waabi says its digital twin uses real sensor data and performs almost exactly like real trucks on real roads. > Waabi says they now match almost exactly - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated The newsletter says the University of Cambridge team’s device produces ethylene and ethane. > The group’s device produces ethylene and ethane - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated MIT Technology Review reports that Boston Metal ran its largest reactor yet to make steel, while noting that the company has further milestones before reaching a scale that could affect the steel industry. > there are still a lot of milestones left before reaching the scale needed to make a dent in the steel industry - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source and contributor context MIT Technology Review · 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. ## AI agent An AI system that chooses and carries out actions toward a goal, often using software tools. - [Definition reference](https://www.ibm.com/think/topics/ai-agents) ## 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: 7cf6818d-f81c-8606-5fe4-e6c0900229a2. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-14T22:23:40.542Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model, second\_opinion\_categorical. Observation: 2a5765478627b902456cd513f39f736147dc6b7057f564b1431a227bc58fc584. Source text hash: eaa63fcb3b94c575950c7b2ecb28649aae8e0fd48f1b8f2707f1e16b67a0da1e. ## Revision and source-check history ### Version 1 — 2026-11-14T22:28:23.593Z On March 12, 2025, MIT Technology Review’s The Download reports that its staff obtained access to Manus, an AI agent launched the previous week, but this newsletter excerpt does not say what its test found. It says Manus claims to be the world’s first general AI agent and can act autonomously across a wide range of tasks. The distinction matters: the source reports the company’s description and the reviewers’ access, not the test’s results. The newsletter also reports Waabi’s claim that its virtual simulation can demonstrate the safety of its driverless trucks without running them for miles on real roads. Waabi says its digital twin—a virtual version of its trucks—uses real sensor data and matches the trucks’ performance on real roads almost exactly. The newsletter presents this as Waabi’s argument for simulation over accumulating real-world miles; it gives no independent assessment of that comparison. Other items describe a University of Cambridge team’s artificial leaf producing ethylene and ethane, and Boston Metal running its largest reactor yet to make steel. The newsletter says Boston Metal still has milestones to reach before its process operates at a scale that could affect the steel industry. The Manus test results, and the basis for Waabi’s safety claim, remain unspecified in this excerpt. ## Original source MIT Technology Review · Article Original source date: Mar 12, 2025 (2025-03-12T12:10:00Z). Date basis: Date from the publisher’s own records, through its API. Added to Escape Meridian: 2026-11-14T22:28:23.593Z. First observed: 2026-10-25T10:36:04.664Z. Retrieved: 2026-10-25T10:37:49.951Z. - [The Download: testing new AI agent Manus, and Waabi’s virtual robotruck ambitions](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Limits Text as retrieved on October 25, 2026; originally published March 12, 2025. The excerpt does not give the findings of MIT Technology Review’s Manus test. The excerpt gives no independent assessment or detailed evidence for Waabi’s safety claim. ## Open question What did MIT Technology Review’s test of Manus find, and what evidence does Waabi use to support its claim that its simulation can demonstrate truck safety? ## Source-stated MIT Technology Review says it obtained access to Manus, but this newsletter excerpt does not provide the test’s findings. > MIT Technology Review was able to obtain access to Manus. Here’s what we made of it. - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated The newsletter says Manus claims to be the world’s first general AI agent and to act autonomously across a wide range of tasks. > Manus claims to be the world's first general AI agent, building off multiple AI models and agents to act autonomously on a wide range of tasks. - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated Waabi says its virtual simulation is accurate enough to demonstrate the safety of its driverless trucks without running them for miles on real roads. > Waabi says its super-realistic virtual simulation is now accurate enough to prove the safety of its driverless big rigs without having to run them for miles on real roads. - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated Waabi says its digital twin uses real sensor data and performs almost exactly like real trucks on real roads. > Waabi says they now match almost exactly - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated The newsletter says the University of Cambridge team’s device produces ethylene and ethane. > The group’s device produces ethylene and ethane - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source-stated MIT Technology Review reports that Boston Metal ran its largest reactor yet to make steel, while noting that the company has further milestones before reaching a scale that could affect the steel industry. > there are still a lot of milestones left before reaching the scale needed to make a dent in the steel industry - [Supporting source](https://www.technologyreview.com/2025/03/12/1113172/the-download-testing-new-ai-agent-manus-and-waabis-virtual-robotruck-ambitions/) ## Source and contributor context MIT Technology Review · 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. ## AI agent An AI system that chooses and carries out actions toward a goal, often using software tools. - [Definition reference](https://www.ibm.com/think/topics/ai-agents) ## 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: 7cf6818d-f81c-8606-5fe4-e6c0900229a2. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-14T22:23:40.542Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model, second\_opinion\_categorical. Observation: 2a5765478627b902456cd513f39f736147dc6b7057f564b1431a227bc58fc584. Source text hash: eaa63fcb3b94c575950c7b2ecb28649aae8e0fd48f1b8f2707f1e16b67a0da1e. ### uncertain — 2026-11-14T22:23:40.542Z — anthropic/claude-opus-5.5 Publication: 7cf6818d-f81c-8606-5fe4-e6c0900229a2 attribution — pass: This is a single MIT Technology Review newsletter issue, The Download, with item blurbs written in the publication's own words, so it counts as one piece. The reading attributes Manus's self-description and Waabi's claims to those companies and the reporting to the newsletter. > This is today's edition of The Download, our weekday newsletter summary — pass: The reading is understandable without specialist background. 'AI agent' is defined beside the reading and 'digital twin' is explained inline. It correctly notes that the Manus test findings are absent from the excerpt. One phrase is slightly ambiguous: 'claims to be the world's first general AI agent and can act autonomously' could be read as the newsletter's own assertion. The claims list attributes it correctly, so this does not prevent understanding. date — pass: March 12, 2025 matches the UTC day of originalPublished, which is consistent with the URL. The undated Deep Dive links that mention later events (for example, 22 nations calling for a global AI body) appear to be later page additions. The reading neither uses nor relies on them. The modifiedAt timestamp precedes the publication instant on the same day, and no later edit is flagged. support — uncertain: Most claims match the source. The closing sentence, 'the basis for Waabi's safety claim... remain unspecified,' sits uneasily with the source, which does state a basis: comparing the twin's performance with real trucks on real roads. The reading's own second paragraph also describes that basis. The question's second half has the same problem. Smallest repair: say the excerpt gives no detailed data or independent check for the comparison, as the limitations already do. > measures how the twin's performance compares to that of real trucks on real roads media — pass: No media was selected, and no verified candidates were supplied, so nothing was misattributed. ## 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/e034f3702c87edb269e63f1436a875f7b02f3790e6c1536b2b1326ba3a4d1476/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/mit-technology-review-s-the-download-reports-on-manus-and-waabi-e034f370.json)