←back to thread

674 points peterkshultz | 4 comments | | HN request time: 0.266s | source
1. constantcrying ◴[] No.45636092[source]
Some more advice:

Tests are all bullshit. They are just some arbitrary questions, trying to figure out whether you understood the material, which were made up by some guy who has much more important things to do.

If you want to spend your time well, either do networking or try to understand the material. If you are there trying to game the system (which hilariously Karpathy is suggesting you do, in a very mild way) then you should seriously consider why you are there in the first place.

Also consider that when you are tested outside of school you will always be tested to face to face.

replies(3): >>45636133 #>>45636250 #>>45636261 #
2. b33j0r ◴[] No.45636133[source]
My tests were always tricky variations on a problem, or the exact problem, which they completely solved in the TA sessions.

I couldn’t figure out why I got the first B’s and C’s of my life until I realized that.

3. Projectiboga ◴[] No.45636250[source]
Yep, and they are often scaled by the "Normal Curve". The catch with this flawed reasoning is that a Normal Distribution requires multiple independent imputs where none is deterministic. Having a professor and a curriculum are together goes against that proposition. There needs to be a better distribution to measure varation within an effective teaching method. Teaching towards a normal distributon result favors giving tests with flawed grading and other bad practices. In the late 1980s I didn't want to pursue a PHD due to the bleak future of acedemia, but if I had this would have been a major focus as it is causing a myriad of problems in our society abusing higher eduction students in this way.
4. aidenn0 ◴[] No.45636261[source]
The second half of this explains why he suggests gaming it, and seems to boil down to Grant Allen's maxim of "Don't let your schooling interfere with your education"