Original title: The Human Advantage in the Robotics Revolution
Original author: Sumir Meghani, Instawork Robotics Labs (IRL)
Translation: Peggy, BlockBeats
Editor's note: While most people are still discussing "Will robots replace human jobs," this article argues that humans will not only not be replaced but are becoming an indispensable key infrastructure in the "physical AI system."
The current core bottleneck in the industry lies not in algorithms or hardware, but in "data and deployment capability." Robots need to train by observing skilled humans performing tasks in real environments, but high-quality, diverse physical world data is extremely scarce, resulting in what is termed the "100,000-year data gap." This has brought to light a long-overlooked capability—the skilled, schedulable, and verifiable human workforce.
Within this framework, the role of humans is redefined: as the "data source" for training machines, providing standardized and labelable operational processes; as the "field nodes" that support system operation, taking on maintenance, repair, and remote control; and ultimately entering a "human-machine collaboration market" connected by platforms, becoming a necessary condition for the scaling of robots.
In fact, technological change does not eliminate labor but rather restructures labor division. From ATMs to the internet, every technological leap has been accompanied by employment anxiety, but what often changes is not "whether there are jobs" but "how jobs are redefined." In this round of technological cycles represented by humanoid robots, the same path is being replayed: tasks are disassembled, capabilities are standardized, positions are reorganized, and new occupations are generated.
The real opportunity lies not in "replacing humans" but in who can build that bridge to transform human capabilities into scalable data, operational systems, and collaborative networks.
Below is the original text:
A year ago, I posed a somewhat unusual question for the labor market: What will happen to the "pros" on our platform when robots arrive?
Our vision is to create economic opportunities for pros and partners worldwide. Today, over ten million pros rely on us for their livelihoods, and many of them are already pondering the same question. We have a deep responsibility toward this and must provide answers.
At the same time, we have observed an unexpected phenomenon: some robotics companies have begun to appear on our application platform, collaborating with our pros. They need individuals with expertise in robot training tasks and need to enter various diverse business scenarios—the very environments where future robots will be deployed. What they depend on is the workforce system we have been building.
At that moment, everything suddenly became clear: Instawork can provide human labor for the "physical AI economy."
The "100,000-Year Problem"
Ken Goldberg succinctly summarized this issue as the "100,000-year data gap": on one side is the vast amount of data used to train language models; on the other is the extremely limited and highly specialized data needed to train robots to perform delicate operations in the physical world.
Note: Ken Goldberg is a highly influential scholar in the field of robotics and artificial intelligence, as well as an artist and interdisciplinary researcher
It is this gap that, despite billions of dollars continuously flowing into robotics companies, we still have not seen humanoid robots cleaning hotel rooms or unloading goods in warehouses... at least not yet.
Our estimates suggest that the entire industry collected approximately 100,000 hours of training data in 2024; this number will rise to one million hours by 2025; and by 2026, it is expected to reach 20 million hours. This represents exponential growth, but even so, it only covers 0.04% of bridging this gap.
An increasing number of companies are joining this race, attempting to build humanoid or general-purpose robots: foundational model labs are developing vision-language-action (VLA) models, hardware companies are constructing physical machines, and participants in the intermediate sector are continuing to emerge. Capital investment has reached tens of billions of dollars. Yet all these participants face the same bottleneck: data.
However, the key is that we have actually seen this scene before.
When ATMs first appeared, almost everyone predicted that bank tellers would disappear. But the opposite happened— the number of tellers actually increased. ATMs reduced branch costs, allowing banks to open more branches; and the role of tellers shifted from counting cash to maintaining customer relationships.
This pattern recurs in every significant technological change: the Industrial Revolution, electrification, the internet. New technologies do not obliterate jobs; they reshape them and create more new opportunities.
A new wave is coming, but this time, it looks more like us: with arms, legs, and eyes.
The Three Acts of Physical AI
Act One: Training Robots
Over the past year, I proactively reached out to some of the brightest minds in the global robotics learning field—from researchers and lab heads to entrepreneurs developing dexterous robotic hands and complete humanoid robots. They generously shared their time and insights, impressing me greatly. Frankly, we weren’t originally part of this industry; however, the more I listened, the more clearly I saw the space where Instawork could fit in.
One viewpoint was repeatedly mentioned: robots learn by observing skilled humans performing delicate physical tasks in real environments. This means everything from properly cutting vegetables to navigating busy warehouses and arranging hotel bedding according to brand standards. The challenge lies in the fact that collecting such data with high quality is extremely difficult—you cannot just randomly have someone wear a camera and start recording. The data must cover diverse environments, tasks, and hand movements; more critically, the individuals performing these tasks must be genuinely skilled. Otherwise, robots trained with "poor cutting skills" will only learn "poor cutting skills" (which is not good for anyone).
This is fundamentally an operational labor problem: how to recruit skilled workers, train them, ensure output quality, and manage a distributed labor network across different regions and scenarios—this is precisely what we have been doing. We have over ten million validated skilled pros, covering hundreds of task types; established deep relationships with partners, enabling access to real business scenarios; and have data on who has stable attendance and consistently produces high-quality work. This combination is something no data collection company can replicate from scratch. In fact, many labs have approached us voluntarily, and we are currently collaborating with most leading teams in this field.
Act Two: The Rise of Robot "Handlers"
One thing that is often overlooked is that robots also need humans.
An executive from a leading robotics company told me they have a critical component that needs to be replaced every 4-6 months—which is frequent enough that dedicated technicians cannot be justified, yet high enough that any downtime incurs noticeable losses. With the rise of autonomous driving, delivery robots, and various automated deployments, more companies face similar issues: expansion requires on-site support, but equipping dedicated personnel in every market is not economically feasible.
We have already launched pilot projects with several robotics companies, covering services like battery replacement, component substitution, and robot maintenance. Meanwhile, we have established a certification system for hourly workers—arguably the first attempt in the industry. In just the initial weeks, over 20,000 pros have been certified.
On the data collection side, certified pros learn how to operate wearable recording devices, capture high-quality video, and annotate sensor data—when a robotics lab needs to record hours of bed-making processes in real hotel suites, they get professionals rather than "newbies learning on the job." On the technical support side, certified pros will master hardware diagnostics, safety protocols, and maintenance processes specific to particular robotic systems.
Imagine a scenario where a logistics company deploys an automated fleet of robots across more than a dozen warehouses. At two a.m., a robot in the Memphis warehouse miscalculates navigation, or a sensor module needs to be replaced on an device in Phoenix. No need to wait for factory techs to arrive days later; a certified Instawork pros can arrive within hours to resolve the issue. Simultaneously, we are also developing VR-based remote control training to support labs in expanding data collection without being limited to purely on-site recording.
If tens of billions of AI devices are expected to be deployed over the next decade, the opportunity lies not just in maintaining them but in creating entirely new job categories: robotics technicians, fleet operators, remote control experts, and perhaps even new positions we have yet to name.
Act Three: The Human-Machine Collaboration Market
Last year, I had lunch with the CEO of a large global hotel chain. They are seriously considering how to enhance the consistency of room service through automation. Numerous robotics companies seek to deploy products in their hotels, but they find it challenging to assess—which is merely a "demonstration effect," and which results in genuine "operational outcomes." We are very familiar with these scenarios, processes, and pain points—because we have already been providing services in these locations.
We are building a "robot service market"—connecting robotics companies with businesses ready to deploy automation. We are serving both supply and demand sides simultaneously, meaning we are not merely "matchmaking," but can genuinely drive implementation.
The future is not about "robots replacing humans" but about "robots collaborating with humans." This is precisely what the Instawork Robotics Lab aims to achieve: three capabilities, one platform—training robots, supporting their operation in the real world, and connecting them to genuine business scenarios that need them.
Bridges
In every major technological change, the question has never been whether new jobs will emerge—the answer is always affirmative. The real question is: who will build the bridge connecting the present to the future.
We believe that at every stage of this process, skilled humans are needed—from training the first generation of robots, to deploying large-scale systems, and to designing future human-machine collaboration processes. We hope that the pros on our platform can span the entire process.
In the "physical AI revolution," Instawork hopes to be that bridge: accumulating deep experience in the most influential industries; already providing training data for robotics labs; already cultivating certified talent for data collection and on-site operations; and also building a market that connects robotics with business needs.
We look forward to the next stage.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
