Did not expect that I could overdo it with one Python function called around 270 times...

Trying to get ChatGPT to write decent code given a slightly hypothetical situation and...

I have owned this book for many years and am just now noticing that the binding has a typo (it's Scikit-Learn, only one T).

Stack Overflow: "A user has requested your help!" Me💭: Oh FFS, why. And why bother me when I'm trying to see the answer that I was searching for? SO: It's on a library you've used once, over six years ago, never deeper than a tutorial. 🤦‍♂️🙄

Me: I had better back this up before running. Server:

I really wish LinkedIn didn't normalize writing with one sentence per line.

'Vacations.' The word is 'vacations.'

I guess I can throw this away 😂😂😂

It's starting to look like the fallout from this latest TRB volunteer reorganization is going to be more papers reviewed by fewer people...

Well, this is a new one:

sending data as a PDF

Less than 10% of papers in transportation research make either their data or code available, per this presenter at MOMO. 😬😬😬

Saying that if we had 10 gigawatts for AI we can cure cancer is like saying that if we had a giant canvas we can paint a masterpiece.

In a world full of AI making professional headshots, be the one to hire an actual photographer.

I almost made the mistake of using a similar word - biweekly - in a scope. Decided to replace it with "every other week", because like biannual, biweekly is "twice a week or every other week". Totally asinine.

Yearly #TRBAM reminder!

Clearly this year's research du jour is AI and LLMs.

Me when I hear dataset 1 vs. dataset 2 debates. Especially when it is used to develop a tool used to assist in decisions surrounding building very large and expensive projects.

"We built a LLM because our data scientists got bored" 😂😂😂

This is a great idea except that it doesn't appear to be enforced (saw someone using a phone), and using a phone silently or with headphones may not be a problem.