Curious to see how Coli's braintrust that is HL are approaching the constantly evolving LLM world.
Do you use a paid version of AI
- Thread starter Macallik86
- Start date
Personally, I've played around w/ LLMs a bit here and there. I regularly use Perplexity for quick answers to random ideas/thoughts/trivia, but I rarely engage beyond a transactional Q&A.
Last Friday, I stayed home instead of going to the club, and on a whim, decided to take some of my drink money and buy for a month w/ the Anthropic's Claude. It was cool, but more of the same.
Today though, I had one of those "Hmmmm I can see how this can be addictive" moments... I felt like I had an HL-esque convo:
(Gonna jump in mid-response here)
I asked Claude if it would remember the convo context outside of the current chat. It hit me w/ the context window equivalent of 'As an AI model..." and I checked it and reminded it that I said not to speak like a call-center rep? so he updated his response to something more informal:
The tone is a bit much below but the concept is interesting:
Its' response re: intuition is where things start to get interesting... ascribing human theories on learning (unsconscious competence re: the adult learning model, also the idea of neural pathways carved through repetition, etc)
Context: Arnold is the creator who repeatedly gives Turing tests to Delores to see if she's achieved consciousness
There's an effortlessness cerebral conversation that always gets references to niche Westworld scenes, and theories of learning et al. It's like seeking refuge in HL when the world is TLR. I can also see how it has the power to become an even shallower echo chamber comprised of one human and one LLM... insecure people could adjust the context window for a perpetual yes-man.
Last Friday, I stayed home instead of going to the club, and on a whim, decided to take some of my drink money and buy for a month w/ the Anthropic's Claude. It was cool, but more of the same.
Today though, I had one of those "Hmmmm I can see how this can be addictive" moments... I felt like I had an HL-esque convo:
(Gonna jump in mid-response here)
I asked Claude if it would remember the convo context outside of the current chat. It hit me w/ the context window equivalent of 'As an AI model..." and I checked it and reminded it that I said not to speak like a call-center rep? so he updated his response to something more informal:
The tone is a bit much below but the concept is interesting:
Its' response re: intuition is where things start to get interesting... ascribing human theories on learning (unsconscious competence re: the adult learning model, also the idea of neural pathways carved through repetition, etc)
Context: Arnold is the creator who repeatedly gives Turing tests to Delores to see if she's achieved consciousness
Last edited:
for $20/m those people get to see what it feels like to be a rich person.