Short essays on things that seem interesting
I was still an AI skeptic back in 2025, as you can see from my previous post back in December. The pace of improvement has gotten exponentially better since then, at least from my experience and recollection. I expected LLMs to improve the state of the art of AI-assisted coding eventually, but not this quickly. I don't want to say LLMs made me eat my own words, but it seems like they did.
This website is built using mkdocs-material, which I chose because it’s easy to maintain and comes with a built-in blog plugin. Overall, it’s been a smooth experience until today.
It's amazing how popular Robotics as become a regular topic in mainstream media. It is no longer science fiction. Autonomous cars are roaming the streets in SF and nobody bats an eye. Humanoid robots startups are floating like flies.
When I first started working with ROS back in 2018, I was just getting my feet wet in the world of mobile robotics. ROS has been around since 2011. Since then, RViz was the de facto 3D visualization tool, and rqt_plot was the go-to option for visualizing time series data from ROS topics. These tools were essential, if somewhat limited.
Recently I came across Professor Andrew Lo's MIT OCW videos on Finance theory on youtube. In the video talking about stock options, he used the analogy of thinking of higher education as a option, or a call option to be exact. You paid tuition as premium. At the end of the 4 years, the degree may earn you a job with a good starting salary, just as the value of the option become as the stock price hit the strike price. The most you can lose is the tuition aka option premium.
I came upon a news video talking about weather forecast and the effect of Machine Learning on the speed and accuracy of ML compared to traditional weather model (I'm guessing numerical physics model). I gotta say I'm quite impressed by how popular ML is getting especially getting a piece on the news.