[awskinda]

Sure, but it’s pretty amazing to me that ChatGPT didn’t just hallucinate a response to a generic request to decode a string. It recognized the string as base64, wrote a valid program to decode it, and returned the correct response.

Maybe I’m just old and amazed, but that seems pretty cool (terrifying?) to me.

[gtirloni]

What's terrifying about it? Serious question. Lately people are associating all sorts of doom to LLMs so I'm curious to understand.

[awskinda]

I’m not sure how I feel yet. Rapid rate of change can uproot systems pretty quickly. I guess I’m just holding judgement to see if this is a new Industrial Revolution, and fallout that might occur. Ideally this wouldn’t be a worry, but we don’t live in an ideal world.

I personally use AI as a tool, and feel more productive FWIW.

[0xDEADFED5]

If that's terrifying to you, definitely do not check out Sora by OpenAI