Code & Clarity

Amazon just announced it will close Mechanical Turk to new customers starting July 30, 2026. Current users can keep working, but the platform won’t get new features or major updates. The move signals a slow wind-down for the 21-year-old service. Mechanical Turk started as a place for companies to pay people small sums for quick, repetitive tasks like labeling images or transcribing audio. Over time, it became a key tool for training AI, with workers annotating data to help machine learning models improve. Yet the platform also faced criticism for low pay and poor working conditions, sparking debates about the ethics of gig work. The service even played a role in the Cambridge Analytica scandal, where data collected via Mechanical Turk helped build voter profiles. Ironically, as AI advanced, many workers on the platform began using AI tools themselves to complete tasks faster, raising questions about the reliability of the data being generated. Launched in 2005, Mechanical...

India's gig economy has grown significantly in recent years, with online food delivery and home services becoming increasingly popular. A startup called Human Archive is tapping into this trend, partnering with companies to collect data on everyday tasks that can be used to train robots. The company is having workers wear special caps with cameras to collect egocentric video data, which can then be used to train AI models. The startup has already deployed over 1,000 headsets across multiple locations and has raised $8.2 million in funding from investors. Human Archive's approach is unique in that it is collecting data from multiple sensors, including cameras, tactile gloves, and full-body motion capture suits. This data can be used to fine-tune AI models and demonstrate their effectiveness on robots. The company is working with multiple partners, including smaller startups, to offer discounted services to customers in exchange for data collection. Human Archive...

Get ready for a fascinating twist in the world of self-driving cars and ride-sharing. Uber recently revealed a bold long-term plan: it wants to turn its millions of human drivers’ cars into a vast network of data-gathering sensors. This data would then be fed to companies developing autonomous vehicles and other artificial intelligence models that need to understand the real world. Praveen Neppalli Naga, Uber’s chief technology officer, shared this vision at a tech event in San Francisco. He described it as a natural evolution of a program Uber introduced earlier this year called AV Labs. Currently, AV Labs uses a small fleet of Uber-operated cars specially equipped with sensors. However, the ambition is much bigger. Naga confirmed that the company eventually aims to equip its massive global network of drivers with these sensor kits. Imagine the sheer scale of data collection if even a fraction of Uber's millions of drivers worldwide were contributing. This move aims ...