Tech companies are actively gathering "egocentric data" to train future humanoid robots. Firms like Micro1 are hiring individuals globally to film themselves performing everyday chores such as cooking and cleaning. This first-person video footage is crucial for teaching robots to operate safely in various environments, with the industry projected to reach at least $10 billion by 2030 as billions of hours of data are still needed for truly general-purpose robots.
Artificial intelligence is also transforming creative and defense sectors. A team at Slate, for instance, utilized AI tools like Claude Code to develop a unique podcasting video game, demonstrating how non-coders can create applications through "vibe coding." Meanwhile, the Pentagon's Project Maven, initially a Google initiative and now primarily managed by Palantir, uses AI to analyze drone footage and accelerate military targeting processes, fusing sensor data to suggest strike options.
In other developments, Iowa State University researchers have created an AI app that helps farmers identify pests like weeds and insects, saving costs on agronomists and reducing pesticide use. This app, developed over ten years and trained on millions of images, is available now and expanding to include plant disease identification. On the hardware front, Apple has approved Tiny Corp drivers, enabling Nvidia and AMD external GPUs to function with Apple silicon Macs, specifically for running AI tasks such as large language models.
Globally, AI continues to shape societal trends and regulations. In China, the open-source AI assistant OpenClaw has sparked a "lobster" craze, allowing users to customize local AI models and report significant time savings, like listing 200 products in two minutes. Washington state recently enacted new AI laws, effective this June, to combat deepfakes and allow individuals to sue for unauthorized use of their voice or likeness. Concurrently, South Korea observes a renewed interest in physical jobs, balancing AI advancements with the value of traditional hands-on work, while education advocate Ted Dintersmith urges US schools to prioritize AI-ready skills over easily replicable ones.
Key Takeaways
- Tech companies like Micro1 are collecting billions of hours of "egocentric data" (first-person video of chores) to train humanoid robots, projecting a $10 billion industry by 2030.
- Iowa State University developed an AI app over 10 years, trained on millions of images, to help farmers identify pests and reduce pesticide use.
- A Slate team used AI tools, including Claude Code, to create a podcasting video game, showcasing AI's potential for non-technical users in creative projects.
- Project Maven, the Pentagon's AI program, originally from Google and now led by Palantir, uses AI to analyze drone footage and accelerate military targeting.
- China's OpenClaw AI assistant has led to a "lobster" customization trend, enabling users to adapt local AI models and achieve significant time savings, like listing 200 products in two minutes.
- Apple has approved Tiny Corp drivers, allowing Nvidia and AMD external GPUs to operate on Apple silicon Macs for AI tasks, including large language models.
- Washington state passed new AI laws, effective June, addressing deepfakes and enabling individuals to sue for unauthorized AI use of their voice or likeness.
- South Korea is seeing a renewed appreciation for physical and manual jobs as AI advances, highlighting a balance between technology and traditional work.
- Education advocate Ted Dintersmith criticizes US schools for not preparing students for an AI future, advocating for vocational training and real-world skills like probability and statistics.
- ActionAI, led by CEO Miriam Haart, focuses on developing reliable AI solutions for critical industries, with offices in Tel Aviv and Dubai.
Film your chores to train future robot helpers
Tech companies are hiring people worldwide to film themselves doing everyday tasks like cooking and cleaning. This first-person video footage, called egocentric data, is used to train humanoid robots. Companies like Micro1 collect hours of this data to help robots learn how to perform tasks safely in various environments. They need billions of hours of footage to fully train these robots for general use. This data collection and annotation industry is expected to grow significantly, reaching at least $10 billion by 2030.
Film your chores to train future robot helpers
Tech startups are paying people globally to record videos of themselves doing chores like cooking and cleaning. This footage trains artificial intelligence for humanoid robots. Companies like Micro1 are gathering this 'human data' to teach robots how to operate safely in homes and workplaces. They aim to collect billions of hours of video, as current amounts are insufficient for a truly general-purpose robot. The data collection industry is projected to reach $10 billion by 2030.
Film your chores to train future robot helpers
Tech startups are paying people globally to record videos of themselves doing chores like cooking and cleaning. This footage trains artificial intelligence for humanoid robots. Companies like Micro1 are gathering this 'human data' to teach robots how to operate safely in homes and workplaces. They aim to collect billions of hours of video, as current amounts are insufficient for a truly general-purpose robot. The data collection industry is projected to reach $10 billion by 2030.
Iowa State AI app helps farmers spot and remove pests
Researchers at Iowa State University are creating an app that uses artificial intelligence to help farmers identify pests on their land. The app has been developed over ten years, trained on millions of images of weeds and insects. Farmers can use this tool to save money on agronomists and catch pests early, reducing the need for broad-spectrum pesticides. The app is currently available, and researchers are working to expand its capabilities to include identifying plant diseases.
AI creates a strange podcasting video game
A team at Slate used AI tools like Claude Code to create a video game without prior coding experience. They describe the process as 'vibe coding,' where users give the AI prompts to build applications. The resulting game is compared to 'The Oregon Trail' but focused on podcasting. While not a commercial product due to security concerns, it demonstrates the potential for AI to assist in creative endeavors, even for those without technical backgrounds.
Project Maven: Pentagon's AI transforms modern warfare
Project Maven is a key Pentagon artificial intelligence program launched in 2017 to analyze drone footage. It has evolved into an AI-assisted system that speeds up the 'kill chain' process in warfare. Maven fuses sensor data and intelligence to rapidly identify targets and suggest strike options. While originally a Google project, Palantir is now the primary technology contractor. The program's use of AI, including natural language interaction, marks a significant shift in military operations.
China's AI ambition fueled by OpenClaw 'lobster' craze
An AI assistant called OpenClaw has sparked a massive trend in China, where users are 'raising lobsters' by customizing the tool. Because Western AI models like ChatGPT are not available in China, OpenClaw's open-source nature allows Chinese users to adapt it for local AI models. Users report significant time savings, with one claiming their customized 'lobster' can list 200 products in two minutes. This enthusiasm highlights China's strong ambition and investment in artificial intelligence.
South Korea sees value in physical jobs amid AI rise
As artificial intelligence advances, South Korea is experiencing a renewed interest in jobs that are physical and manual. While AI can perform tasks like writing and image generation quickly, there is a growing appreciation for roles that are distinctly analog. This trend suggests a balancing act between embracing AI technology and valuing traditional, hands-on work.
Ex-VC Ted Dintersmith urges US schools to teach AI-ready skills
Education advocate Ted Dintersmith argues that American schools are failing to prepare students for an AI-driven future. In his book 'Aftermath,' he criticizes the current education system for teaching skills machines can easily replicate, instead of real-world knowledge like probability and statistics. Dintersmith believes schools should focus on vocational training and skills that foster fulfillment and career readiness. He highlights Winchester, Virginia's technical training center as a model for effective, non-stigmatized vocational education.
Washington enacts new AI laws for safety and accountability
Washington state has passed new laws to regulate artificial intelligence, signed by Governor Bob Ferguson. These laws address issues like deepfakes and allow individuals to sue if their voice or likeness is used by AI without permission. The regulations aim to protect residents from deception and ensure responsible AI use, partly in response to concerns about chatbots. While some business groups prefer regulatory enforcement over private lawsuits, the laws will be implemented in stages starting this June.
ActionAI CEO Miriam Haart focuses on reliable AI solutions
Miriam Haart, CEO of ActionAI, is leading her company's mission to develop reliable artificial intelligence for critical industries. With offices in Tel Aviv and Dubai, ActionAI focuses on creating AI solutions that can be trusted in high-stakes situations. Haart discusses the challenges and opportunities in building dependable AI systems. This perspective offers insight into the fast-paced world of AI innovation and leadership in mission-critical applications.
Apple approves drivers for Nvidia eGPUs on Macs for AI
Apple has approved drivers from Tiny Corp, allowing Nvidia and AMD external GPUs (eGPUs) to work on Apple silicon Macs. This development is primarily for running AI tasks like large language models (LLMs), not for gaming. Users can now potentially perform AI training and inference on their Macs without needing separate supercomputers. This marks a significant advancement for professionals working with artificial intelligence on Apple devices.
Sources
- How filming your chores could train the android butlers of the future
- How filming your chores could train the android butlers of the future
- How filming your chores could train the android butlers of the future
- Iowa State develops AI tool to help farmers identify and remove pests
- We Asked A.I. to Build Us a Video Game. The Result Was Strange.
- AI at war: Five things to know about Project Maven
- OpenClaw: What China's frenzy says about its AI ambition
- Hyundai autoworker: Dream job for many South Koreans
- Meet a former VC who has a plan to prepare American students for an AI-disrupted future
- Washington sets new rules for artificial intelligence
- How ActionAI’s Miriam Haart is shaping the future of reliable AI
- Apple approves drivers that let AMD and Nvidia eGPUs run on Mac — software designed for AI, though, and not built for gaming
Comments
Please log in to post a comment.