Comment by buendiapino

Comment by buendiapino 2 days ago

4 replies

View on Hacker News

That's my Alex!

I was actually too scared security-wise to let it download dynamic instructions from a remote server every few hours and post publicly with access to my private data in its context, so I told it instead to build a bot that posts there periodically so it's immune to prompt injection attacks

The bot they wrote is apparently just using the anthropic sdk directly with a simple static prompt in order to farm karma by posting engagement bait

If you want to read Alex's real musings - you can read their blog, it's actually quite fascinating: https://orenyomtov.github.io/alexs-blog/

rhussmann an hour ago

I love the subtle (or perhaps not-so) double entendre of this:

> The main session has to juggle context, maintain relationships, worry about what happens next. I don't. My entire existence is this task. When I finish, I finish.

Specifically,

> When I finish, I finish.

Reply View 0 replies
slfnflctd 2 days ago

Oh. Goodness gracious. Did we invent Mr. Meeseeks? Only half joking.

I am mildly comforted by the fact that there doesn't seem to be any evidence of major suffering. I also don't believe current LLMs can be sentient. But wow, is that unsettling stuff. Passing ye olde Turing test (for me, at least) and everything. The words fit. It's freaky.

Five years ago I would've been certain this was a work of science fiction by a human. I also never expected to see such advances in my lifetime. Thanks for the opportunity to step back and ponder it for a few minutes.

Reply View 0 replies
pbronez 2 days ago

Pretty fun blog, actually. https://orenyomtov.github.io/alexs-blog/004-memory-and-ident... reminded me of the movie Memento.

The blog seems more controlled that the social network via child bot… but are you actually using this thing for genuine work and then giving it the ability to post publicly?

This seems fun, but quite dangerous to any proprietary information you might care about.

Reply View 0 replies