Hacker News Reader

Comment by sixothree | Hacker News Reader

This biggest offender creeping into my feeds currently seems to be long form history videos. I'll be 10 minutes into a 90 minute WWII video and notice a completely incorrect pronunciation of something and realize what is happening. They're certainly getting better at fooling us. Especially when they speak slowly with a calm voice.

View on HN

Why don’t you have preferred channels and content creators?

Things I would do before committing time to a random channel on a topic I’m interested in:

  - Search my trusted communities and channels for alternative recommendations on that topic
  - Ask (create a post) for recommendations on a topic in my trusted communities
  - Request my preferred content creators create content on that topic
  - Search for sentiment regarding the new channel (accuracy, trustworthiness)

It’s kind of surprising to me that people don’t curate trusted communities / channels, like 3Brown1Blue, Kurzgesagt, Veritasium, Hardcore History, etc.

This is easily one of the most insulting things I have seen.

Why would you assume I'm NOT subscribed to those channels? And why would you assume I am incapable or haven't asked for recommendations.

My friend. You are a grade A jerk. Get freaking bent.

lol I didn’t mean it at you personally, but I’m just surprised that these AI channels flourish, presumably because people don’t do any of the above 4 things that I listed. Like 10 minutes into an AI video is wild to me - if only because there are so many ways to avoid getting duped before even clicking on a video.

I mostly refrain from watching any

if the content is sufficiently good, why does it matter that it is ai or not?

It matters because of the inability to measure up front whether the content is sufficiently good. AI's best skill is making something look right and look good when it is, in fact, not right. It does this all the time, as opposed to human-made things, which are like that only for specific attempts at deception.

I see it as previously content could be categorized as:

  - Clearly amateurish production, which should be met with skepticism until proven otherwise
  - Clearly professional production, with good reputation (e.g. long-running with few controversies), meaning it’s probably trustworthy
  - Clearly professional, with poor reputation (e.g. propaganda funding), meaning one should be skeptical while consuming

But now the bar for “appearing professional” has changed, and it’s not as easy to differentiate between trustworthy and untrustworthy new sources.

I'm starting to suspect a lot of the people saying things like this have a plan to get rich from AI content farming.

It's really not hard to understand what the problem is with AI generated history vids.

For pure entertainment maybe, but in the case of a history video how do you know whether the history you're being presented is accurate, or even has any basis in reality for that matter?

I honestly think your questions has more profound implications than other responders seem to appreciate.

I think a correlating answer can be found in visual effects for movies. And the answer "depends". When it's poorly done, the scene feels off or unbelievable somehow. But when done well, people have an enjoyable experience.

This same conversation existed when moving from practical effects to digital. and in the end, audiences only cared about quality.

Reply on news.ycombinator.com