I'm not anti-AI. I'm just pro-thinking.
What do I mean? I mean that everyone I know who is REALLY into AI uses it for reading EVERYTHING longer than 1000 words.
This is the mental equivalent of eating only smoothies, because they go down easier.
I say this with love: reading is thinking.
Stop letting slop think for you.
@raiderrobert I use LLMs every day and still read everything where it's appropriate.
One issue here is that we've been promoting long form content in the last 2 decades for SEO and attention hoarding so people's desire for shorter content is a natural correction for this.
LLM output is just as superficial as the SEO-bloated text was.
It's just shorter.
So, it's like you take your medicine, and you thin it with water. And then you throw away nine tenths of it.
That's homeopathy, for ideas.
It's stupid and harmful.
@androcat @wraptile @raiderrobert I love that analogy: LLMs are “homeopathy for ideas.”
I think that machine learning is an incredibly powerful tool, just like other kinds of statistical analysis. It’s just currently being hyped beyond the moon for wildly inappropriate uses.
Specifically, using LLMs to summarize text that was deliberately bloated is the homeopathy bit.
LLMs themselves I tend to call Fermented Statistics, but the scholarly term "stochastic parrots" is also very good.
Robert Roskam
wraptile
Androcat
eswillwalker