I remember, it was midnight, and August 22nd was turning into August 23rd. I was just rolling back into H-town from Tennessee, where I saw the total solar eclipse. I checked wunderground.com and saw a 100% chance of rain sustained for over 72 hours, beginning in a couple of days. I lol'd. I told my friends in the car "Wow, I've never seen these forecast models completely break before! Hilarious, someone's code has a bug."
Then I got home, I think I listened to music for an hour or so before I remembered to check the forecast again. That was when I saw that there was a storm developing over the Yucatan peninsula that everyone agreed would head this way and just hang out about with its eye about 100 miles southwest of here for several days. That night, I made a facebook post warning everyone, I think it got about 6 likes.
So I guess I'm not really impressed with these people who just translate NOAA or European forecasts into words. It's cool, they give it a voice, but I feel like they're just rehashing other people's work into a weatherman blog. I guess that's a form of work, but I've never been terribly impressed. But I do respect that they (at least used to) make an effort to respond to people's comments and especially questions.
At least the article concedes that meteorology is all about pattern recognition. So much of my understanding of space weather is based on checking all the various plots on solarcycle24.com every day for over a decade, now.
IMHO, the real heroes are the people programming the models. And they're almost always the most reclusive people you'll ever meet.