What a 2004 Neural Network Taught Me About AI
Rohit reflects on a 2004 IIT Roorkee internship predicting sunspot activity with a neural network, and what it still teaches about modern AI.
Insights category
Personal Reflections
Personal essays on work, change and the connections between a life in technology and a life in medicine.
1 essay in Personal Reflections
5 min read
What a 2004 Neural Network Taught Me About AI
An internship project on sunspots, and the lessons that still apply in the age of large language models
What the essay covers:
- The shape of the work
- Lesson one: a model learns the data, not the world
- Lesson two: a good fit can be a bad sign
- Lesson three: the error bar is part of the answer
- Lesson four: understanding the domain changes the model
- What has changed, and what has not
- Why I still think about sunspots
About this theme
Why Rohit writes about personal reflections
Some lessons do not fit a category. These personal essays connect experiences from technology, business and medicine.
For the full context, read the About Rohit page, or browse all insights.