Sure, and we had an internet before the world wide web (ARPANET). But that wasn’t hugely influential until it was expanded into what’s now the Internet. And that evolved into the world wide web after 20-ish years. Each step was a pretty monumental change, and built on concepts from before.
LLMs are no different. Yes they’re built on older tech, but that doesn’t change the fact that they’re a monumental shift from what we had before.
Let’s look at access to information and misinformation. The process was something like this:
- Physical encyclopedias, newspapers, etc
- Digital, offline encyclopedias and physical newspapers
- Online encyclopedias and news
- SEO and the rise of blog/news spam - misinformation is intentional or negligent
- Early AI tools - misinformation from hallucinations is largely also accidental
- Misinformation in AI tools becomes intentional
We’re in the transition from 5 to 6, which is similar to the transition from 3 to 4. I’m old enough to have seen each of these transitions.
The way people interact with the world is fundamentally different now than it was before LLMs came out, just like the transition from offline to online computing. And just like people navigated the transition to SEO nonsense, people need to navigate he transition to LLM nonsense. It’s quite literally a paradigm shift.
zbyte64@awful.systems 4 weeks ago
Enshittification is a paradigm shift, but not one we associate with the birth of the internet.
On to your list. Why does misinformation appear after the birth of the internet? Was yellow journalism just a historical outlier?
What you’re witnessing is the “Red Queen hypothesis”. LLMs have revolutionized the scam industry and step 7 is an AI arms race against and with misinformation.
sugar_in_your_tea@sh.itjust.works 4 weeks ago
It certainly existed before. Physical encyclopedias and newspapers weren’t perfect, as they frequently followed the propaganda line.
My point is that a lot of people seem to assume that “the internet” is somewhat trustworthy, which is a bit bizarre. I guess there’s the fallacy that if something is untrustworthy, it won’t get attention, but instead things are given attention if they’re popular, by some definition of “popular” (i.e. what a lot of users want to see, what the platform wants users to see, etc).
Well yeah, every technological innovation will be used for good and ill. The Internet gave a lot of people a voice who didn’t have it before, and sometimes that was good (really helpful communities) and sometimes that was bad (scam sites, misinformation, etc).
My point is that AI is a massive step. It can massively increase certain types of productivity, and it can also massively increase the effectiveness of scams and misinformation. Whichever way you look at it, it’s immensely impactful.