One of the only things more dangerous than ignorance is false confidence. An official USG chatbot potentially makes things much worse.
they can be taught where they went wrong.
if a pattern emerges they can be moved to a role more fitting for them. or removed from a project entirely.
we can judge their effectiveness from past performance.
we can put them in less important roles and gauge whether or not they should be moved up.
so, no, pretending that the correct route forward is to “remove humans from the equation” for a bot that doesn’t learn, is never held accountable, and not held to the same standard as humans is silly tier thinking.
As an individual, yes. As a population, no. They elected a conman twice and would rather burn down the country rather than changing their view.
Saying they didn’t learn is an incorrect analysis, since they learned well, based on the inputs they were provided.
I always recommend Network Propaganda for an empirical analysis of what the patterns actually are.
I used to believe Trump winning the 2016 election was a product of rightwing propaganda and fringe theory, but with the 2024 election, I had to revise my belief. What if this is the mainstream mindset of Americans all along, and Trump simply tried to capitalize on that?
The later chapters are where they go into the more recent details. There is a video of one of the authors (not either Robert Faris or Hal Roberts) is giving a talk at a university panel, and he gives a shorter synopsis of the main mechanism, Its about 11 to 15 minutes in.
The set up of the book puts the evidence for the case they make later, namely that the left and center media are trying to follow journalistic standards, while the machine on the right is entirely different.
I'm pulling from memmory here, bu as an example - a fringe theory will start circulating on the fringes of the internet, the most engaging narrative will be brought up on the podcast circuit. At some point a talking head (Guiliani back in the day) will reference the theory on the news. Which will in turn be referenced by the government who will say this is a topic being discussed on the news.
The other distinction they make is that people on the right who say the narrative is wrong, simply don't get platformed.
They go through this in exhaustive detail. If you know most of the history you can see if skipping to later sections works for you. Its the best description of the actual media mechanics at play.
The 2024 elections were the ones that forced me to really re-examine my priors, and I came to a similar conclusion that Benkler and co did. Their book simply brings all the receipts.
bad bots can be retrained or shut down far more easily
I don't see such a path with an llm.
particularly most of the commercial models.
again, i’m sure we can all come up with a thousand “hypotheticals”, and tbh, the hypothetical parade is not something i’m interested in having. but i’ll state again, “entirely removing humans” from the situation is silly-tier thinking.
particularly as even the ceos of sota frontier producing models will each and every single one tell you to never trust their model.
“our model is smartest thing in the world…”
…next breath..
“wait, you trusted our model? that was silly of you. always double check it”
This seems like a great source for domain-specific RL, but even without that, we can expect there to be higher-quality LLMs that can be trivially swapped in within a year, if not a week.
i didn’t say this at all. did you respond to the wrong comment?
i was responding to the comment which said:
> … remove humans from the equation, then.
> We shouldn't use an LLM because...LLMs never improve?
the context of my response was “removing humans from the equation” is sillytier thinking.
not “We should never use LLMs”
Maybe a human should be available as a backup or something, but not because LLMs don't "learn" -- the system seems to learn in all relevant senses.
(I'd also object to the no accountability, but another subthread is already on that.)
I was referring to the government side, as a mild quip.