HackerTube 中文?

I tried using AI. It scared me.

Tom ScottYouTube ↗

Further reading 1

Why do tests pass when the program is wrong? →Gary Bernhardt’s isolated tests and Andrej Karpathy’s neural network both deliver good news when the code is wrong.This video helps answer · Did the implementations agree because they read the same documentation?

Caption notes

Tom Scott and the generated script both misunderstood what a Gmail label attached to. Running the code exposed the difference between a thread and a message. Explaining that observation to the model changed its fix.

Archived captions (transcription method unspecified) ↗
Original passages · 2
made the same mistake I did with my backup.
I tried using AI. It scared me. · ▶ 7:59–8:03
treats labels as attaching to threads not messages.
I tried using AI. It scared me. · ▶ 8:05–8:09

Hacker News comments

View original on HN ↗