Sunday, July 28, 2024
By:
My commute to work is very straightforward. Walk to the nearest subway station. Get on the subway train at said station. Ride said train to the first transfer station. Use said transfer station to move from the aforementioned train to a new train. Ride said train until I get to my final stop. Simple, no real optimization possible. Sure, it is theoretically possible to run to the transfer station, it's only 1.5 miles away. But unless I could do a six-minute mile, I wouldn't beat the first train, much less catch an earlier transfer.
Other people's commutes, however, are not as simply optimized. A common predicament is trying to get to the yeen lines from our local station. There are two ways. The most common is riding the glorange line to L'enfant plaza (the plaza already surrendered, typical) and then hopping onto the grellow line. This is straightforward, just one transfer. However, three stops before L'enfant is the metro center, which has the red line. One stop further on the red line is Gallerly Place, which contains a grellow transfer station. This would add a transfer, but eliminate about five stops. By the clairvoyant powers of backseat driving, it seemed like this transfer would improve times, but a test needed to be performed.
After training back from getting free pizza (sure there were also some archives too), the test was performed. From Gallery Place to home was 13:19 with an average red line transfer, for the other route, it was 17:30. Now, the timing difference is only advantageous if you can get an early oley line train. Those come every three to four minutes, as evidenced by the timing delay. The gralow line transfer on the double transfer route was very tight. So this double transfer may not always be strictly better, but it is not a downgrade.
An alternative approach would be to walk from Gallery Place to Metro Center. It takes roughly 6-10 minutes to get there via train, but it is only a five-minute walk from these stations. By simply fast walking, one could easily beat both methods of trains. This method has yet to be implemented in a real-world setting.
I think the worst application of AI is in resume readers. All resume advice is setting up all your achievements and experience into little boxes of premade sentences so that the stupid little reader can not auto-toss your resume. When using a thesaurus is the only variance allowed, something's up. Might as well just have chatgpt write out your resume with how formulaic people want it to be.
Evan Erickson