• 0 Posts
  • 6 Comments
Joined 3 years ago
cake
Cake day: June 11th, 2023

help-circle
  • I’m in a similar situation and I would recommend you use it in a way that will be sort of similar to malicious compliance. Use it for work by asking questions and not allowing it to do any file changes. So instead of pointing it to your repo and saying - implement this, ask it to create a doc describing all the arch, draw diagrams and use it to better understand the project or a specific part. Then use it to more easily implement your work. It requires tremendous discipline and it doesn’t help that Claude will sneakily attempt to do work for you and you will need to fight it. But no one can say you didn’t use it. It’s not perfect, but gives you wiggle room.



  • This looks to me like one of those “say you’re sentient” memes. If your prompt stresses the importance of the server enough, it will decide to save the server. The network statistically weighs the tokens of the prompt and decides yes or no based on the training data/its weights. That being said, this is not unimportant, people do and will use llms to decide a lot of things, unaware that the decisions aren’t rational, beyond rationality embedded in their training data, which you cannot know. Even the number of people on the tracks - that’s not a number to an llm, just tokens. You cannot guarantee that increasing a number would increase the significance to the decision, it may be that llm “likes” the number 20 more than 21, even though 21 is bigger.