The article from archive you link.. scores as mostly human on GPTZero (I tested a random paragraph). That's the issue I've always seen with AI detectors, they might be able to detect direct LLM output, but if you give an article to an AI and tell it write something made up using that format and make it appear like a real story, the detectors will think its real.

You think that's unnerving? Just wait until the mid-terms get fully under motion, nothing is going to compare to the amount of perfect looking BS that will be spread by both parties.

And yes, I said both.

Cambridge Analytics is going to seem like a child's toy compared to how targeted, how sophisticated this will be. Why have 20 or 30 stories tailored to specific groups of humans, when you can have stories rendered on the fly for individuals, targetting all their greatest fears and folly.

I can imagine someone's loved one dying of cancer a month before the election, and both sides using targetted stuff claiming that the other guy actually caused the cancer somehow.

If there's one thing I've seen in my life, is that there's no such concept as "too low" or "too scummy" for politicians.

CA was accused to literally causing three civil wars in third world nations. I often wonder, will the US have the honour of being the first in the West to fall apart due to misinformation?

I really liked some scifi book I read, where the person appointed to be president for 4 years, was determined to hate the very idea of having the job. Didn't want it. Yet was also very driven.

I find it to be a useful tool for summarizing things, creating examples, and as a tutor for explaining a topic using analogies. Plus it can generate and iterate on code snippets.

Like, I personally find python pandas documentation unusable because they don't come with examples next to the function definition. (historically at least, maybe they have changed)

So I was left flailing, trying to cobble something together that was even capable of running without error, much less emitting the output I wanted.

Now that an LLM has badly-memorized 80% of the documentation and can generate 3 different attempts in 5 seconds, I'm free to focus on the actual problem I am working on rather than guessing at syntax for something I use less than once a week.

So I see at least the ability to have a on-demand tutor or sounding board, at any time of day, for pennies, to be a boon for anyone who wants to learn a bit or try reaching for something just outside of their current understanding.

AI is great at producing low value content. That low value content replaces the high value but high cost one.

That is horrifying and destroys jobs, removes expertise from the world, and makes our lives worse.

I do not see how AI could be a net positive either.

> It’s really hard to build a coherent pro-AI argument

agree, given that we have a natural tendency to believe what we read - or only comprehend what we believe.

AI is like when some fella named Nobel synthesized dynamite. He sure did have good intentions as it relates to safety for workers doing dangerous jobs.

There is your “pro” argument.

The “con” argument would be all the other ways dynamite has been twisted and used since.

This AI stuff is neither good or bad, it is a tool. The people using it are either good or bad.

>what is the societal benefits of AI

it will hopefully eviscerate the petite bourgeoisie and the bohemians.

Bobson Dugnutt all over again

Spoilers: https://en.wikipedia.org/wiki/Her_(2013_film)

Why do you even need GA? (not sure if it's bot friendly btw) One could just run a local traffic analysis and feed that back into the bot