alt Hacker NewsI was surprised to see so many top comments here pointing fingers at Ars Technica. Their article is really beside the point (and the author of this post says as much).
Am I coming across as alarmist to suggest that, due to agents, perhaps the internet as we know it (IAWKI) may be unrecognizable (if it exists at all) in a year's time?
Phishing emails, Nigerian princes, all that other spam, now done at scale I would say has relegated email to second-class. (Text messages trying to catching up!)
Now imagine what agents can do on the entire internet… at scale.
Replies
The Internet is dead, long live the Internet.
LLMs are just revealing the weaknesses inherent in unsecured online communications - you have never met me (that we know of) and you have no idea if I'm an LLM, a dog, a human, or an alien.
We're going to have to go back to our roots and build up a web of trust again; all the old shibboleths and methods don't work.
Analogously to the surface of last scattering in cosmology, the dawn of the LLM era may define a surface of first scattering for our descendants.
I don't think it's besides the point at all. The Ars Technica article is an exact example of what you go on to talk about for the rest of the comment: the public internet as we knew it is dead and gone. Not in the future, it is already gone. When so-called journalists are outsourcing their job to LLM spam, that's a pretty clear indicator that the death knell has been tolled. The LLMs have taken over everything. HN is basically dead, too. I've gotten some accounts banned by pointing it out, but the majority of users here are unable to recognise spam and upvote LLM-generated comments routinely. Since people can't be bothered to learn the signs, we're surrendering the entirety of the internet to being LLM output that outnumbers and buries human content by 100:1.