Changing technology

I'm what you'd consider an older millennial. In my youth, finding answers meant visiting a physical library or navigating an encyclopedia CD-ROM on a slow desktop computer. (Don't even get me started on dial-up internet!) The convenience of having a minicomputer in our pockets, with instant access to information, still amazes me. This technological leap, powered by AI, has revolutionized how we interact with information.

Becoming an AI tutor

When I first considered editing with Mindrift, I was unsure if I'd qualify. My background as an English teacher, financial adviser, and food service worker hardly screamed "tech expert." But my attention to detail, language skills, love for research, and eagerness to learn proved invaluable in AI training.

At Mindrift, we train Large Language Models (LLMs) using carefully curated conversations about various types of data. For visual recognition projects, we create dialogues that help AI understand and describe images in detail. This process involves teaching AI to recognize objects, understand visual concepts, and draw connections between different elements in an image.

Fascinating content

This work has led me to reflect on how human visual intuition develops. Just as we learn to recognize objects through repeated exposure and comparison, AI models learn through processing numerous conversations about diverse images. The need for AI to compare and infer from these conversations mirrors our own learning process.

From geographical formations in Iceland to Tokyo tourism and international policies, I've gained knowledge about topics I might never have explored on my own. Each task brings new discoveries, whether it's analyzing trends in beef consumption or understanding climate change perceptions across South America.

Graphs, charts and maps

Interestingly, this approach to training AI doesn't just teach recognition; it also helps AI develop a form of intuition. By exposing AI to a wide range of images and discussions, we're enabling it to engage in complex reasoning and draw nuanced conclusions about visual information.

I've found myself captivated by colorful maps and intriguing charts. Did you know that Hispanic populations in rural Northern Nevada are linked to the state's ranching industry? These discoveries often lead me down fascinating research rabbit holes, expanding my knowledge in unexpected ways.

Sensitive content

While most content is objective, we occasionally encounter sensitive topics. These are crucial for AI training, ensuring that AI can handle delicate issues with care and accuracy. It's our job as AI Tutors to make sure AI responds appropriately to queries about social justice, healthcare disparities, or trauma-related topics.

Nature scenes

Some of my favorite tasks involve beautiful landscapes. From deep blue skies and rolling hills to log cabins in snowy woods, these images often inspire me. One black beach scene from Iceland even became my desktop background, fuelling my wanderlust.

Finding inspiration

The content we encounter as AI tutors can be intriguing, thought-provoking, challenging, and inspiring. It's far more than just fixing grammatical errors. Whether learning about veterans' service periods or rediscovering my love for painting through image analysis, each day brings new insights and inspirations.

As we continue to train AI in visual recognition, we're not just teaching machines—we're gaining new perspectives on human cognition and our own learning processes. It's a fascinating journey that bridges the gap between artificial and human intelligence while expanding our own horizons.