Gizmodo and Kotaku Staff Furious After Owner Announces Move to AI Content | Hacker News Reader

Gizmodo and Kotaku Staff Furious After Owner Announces Move to AI Content

futurism.com

39 pointsA4ET8a8uTh031 comments

I used to write short articles for a PC hardware tech news outlet as a side thing.

It went out of business, mostly through no fault of its own, because proper writing and investigative journalism made no money anymore. Basically the site owner had two choices: "sell out" and focus on producing a flood of SEO garbage, or "honorably" shutter the site. The owner went with the later.

Google basically buried us. PC hardware was on the rise, but most of the audience went to YouTube, and the leftovers were presented with a flood of junk sitting on top of us (and our excellent peers, may they rest in peace) in Google search.

...And this was years ago. The problem has gotten exponentially worse. Honestly, I don't know how modern written news sites with any integrity are even surviving.

I do have a prediction, but only time will tell if it actually pans out. Things will get a lot worse the same way things got after introduction of printing press. It will, however, eventually result in equivalent of 'usda organic' content for news.

The printing press and organic produce markets never had global social media/search/advertising monopolies.

Hmm. Yes mostly. Capital P press had access monopoly, which I would say qualified in pre-internet world. I can agree that organic produce is a little different, but even that has been somewhat incorporated into offering by various corps.

I feel for the writers out there, but at the same time most articles I read are terrible. I hate them. It makes me emotional how much I hate them. I don't frequent these two sites, but that's partially because I've given up on reading articles by most publications. Truthfully, in many cases, I thought a lot of news articles were already ai-generated using much lower quality algorithms

That’s a pretty defeatist attitude.

Gizmodo has maintained their quality reasonably well over the years.

I agree that most (in a numerical sense) articles on the internet these days are blogspam, but that’s no reason to accelerate the elimination of journalists.

> Gizmodo has maintained their quality reasonably well over the years.

I disagree. I haven’t read an article on Gizmodo that hasn’t made me angry about how bad it is or how full of horse apples in maybe 10 years.

These are bullshit sites that people must only work for out if desperation. It’s unfortunate for writers but AI articles will probably be less bullshit than the current site.

agreed

> AI articles will probably be less bullshit than the current site.

I'm not so optimistic. Its all in the training. I'm expecting BSGpt that will be trained to produce exactly the kind of nonsense these sites produce and their consumers like.

I think AI content is terrible. But I think Gizmodo is worse. They’ve basically been doing manual AI content for years. A mix of blargh-speak, advertiser commissioned content, and just really bad and lazy authors who don’t have creative freedom to write.

I also don’t think consumers like the current content, but just end up there due to dark patterns and circular advertising. So the content doesn’t matter as all they do is trick people into “reading” enough to watch a pre roll or click a link or whatever they are monetizing.

Agreed with everything you said. But with AI, they will be able to remove the scaling problem - instead of 50 people writing 100 bs articles, there will be an LLM generating 1000 bs articles and the degrees of required bs + dark patterns + commissioned content injection etc can done at a much larger scale. Their core model is bs; they can't/won't change that - they will now just be able to do that at scale.

Good point, so it will be more profitable for Gizmodo as they’ll have much higher content levels.

But for me I can ignore a billion crap results the same as a thousand.

I’m expecting there will be some version of an “AI Long Tail” article soon that explains how the tail will get really long now. Although it would be bad if AI causes a fat tail where everything is the same and there’s no more long tail.

If — as Nick Bostrom and Elon Musk contend — we are more likely than not living in a simulation then ALL content is AI.

I don't know much about Gizmodo, but replacing Kotaku writers with AI would be a net gain for everyone.

There is plenty of good journalism these days, but it is increasingly difficult to find.

This was a problem long before AI.

These kind of articles are what we get when writers are given impossible deadlines and targets to hit (multiple submissions per day in many cases), and incentives that favour clicks and engagement over actual imparting of information.

Even if you’re a good writer, you’re not doing your best work under those conditions.

Gizmodo and Kotaku write poorly constructed outrage-bait articles…AI will honestly do a better work than their perpetually underpaid writers.

I don’t even understand the logic of writing outrage bait all day if you get paid so low for it…sounds like horrible living.

You realize _thats the business model_.

Why would their switch to AI change the business model?

Because it might be better written at least.

It will not.

It would be even more rate inducing if the AI was trained right, as the previous replier is right: it doesn't change the business model.

If your website has such awful offering that you think an automated system can do better then maybe you shouldn't be in business anyways. But then again normal human hands are just to slow to react to the latest gossip

I guess we'll be seeing a substantial improvement in the content coming from Kotaku and Gizmodo in the near future. In all seriousness, it sucks for the writers that might lose their jobs in the future, but at the same time these two companies are some of the worst in the field.

Not to add Jalopnik, either -- they've been some of the worst offenders lately of just running slideshows nonstop, or slideshows of user comments.

I kind of don't see what the editors or writers are actually... doing.

As usual the headline misrepresents the article. They are starting to use it for modest things like articles that are mostly lists. Not actually replacing any journalists yet.

I only hope the AI will be less racist than the humans currently writing for Kotaku.

https://boundingintocomics.com/2023/05/03/kotaku-journalist-...

I read the article. I even followed the link. I am not sure I understand the racist part. Final Fantasy ( and JRPGs ) are being accused -- wrongly -- of being racist, but that is all I could find there.

Is it one of those 'even a hint of someone being upset somewhere makes it racist'?

The corporate blog collectives were always going to bust. Blogs from a single person or laser focused on a specific topic with very low overhead will always have a greater chance to succeed. Collecting groups of podcasts and blogs and gaming them to generate revenue creates this layer of corporate slime that slowly turns off the reader/listener. Each article or show devolves into a funnel to click troughs. Never mind a group of writers who while in this corporate structure boil the frog and their careers.

So essentially requires less journalists in the end, or at most 3 for reviewing.

All it needs is a human to check for the links needed to cite and we can already see how LLMs and the killer use-case (summarization) will be very disruptive in journalism.

They know about other attempts [0] but the reality is, it won't stop and the owners will push on anyway.

Perhaps these journalists should learn to adapt.

[0] https://futurism.com/bankrate-ai-generated-article-errors

These issues are also present for the current television writers’ strike.

I think for most TV shows even GPT-4 would struggle with things like having enough context.

But I did run into one TV show that had good reviews, but seemed like it may actually have been written by an LLM that had been trained on fan-fiction or something. "A Discovery of Witches".

The writer’s strike concern is more hypothetical, but still existential and real, given that studio heads were on earnings calls saying they were looking into AI to cut costs.

Reply on news.ycombinator.com