The great majority of not terminally online people just don’t care about machine generated images, videos and logos.

With the free tier of ChatGPT you hit output limits pretty much immediately. And with the fact that money is running out for OpenAI, I am willing to say the time for free LLMs and image gen models is not going to last too long. There is a reason why Gemini fucking sucks in the free tier, and somehow Copilot sucks even when paying (I am not crazy enough to pay for Copilot, I have it with a work license).

Not default behaviour. KDE Dolphin doesn’t hide Sort By on its default settings. If you didn’t change it, this is the fault of your distribution.

But the loading error icon got mad cake

To be fair all medium to large news agencies would do this after getting treatened to be prosecuted and shut down for being “against him”. Smaller ones would need to stay anonymous, and be in an equally whistleblowing position.

I think this is more of a censurship problem than a specific news agencies one.

Now it is. Before removal it was empty. Likely seen as spam.

The Reddit post has no body

petercopter

It has the ChatGPT image gen grain. Also notice the sushi slice closer to the knife, the salmon makes no sense.

Indiana Woman Mississippi Man :(

Lemmy is way too niche, and probably hard to index too

So many VPN companies have zero interest in privacy or just produce malware. This is nothing new, but it is weird to target LLM prompts and conversations.

That’s not a kettle

Food tampering can absolutely kill someone, and it did. This is not the case of someone trying to pass petty items at a Walmart.

In that case they would not even receive the statistic

Misconfigured systems/browsers or misconfigured web crawlers.

Worms Armageddon and Teeworlds would have no players.

Pretty much everyone I find in those games are russians. And it’s not in low ping in Europe servers either, in high ping servers you just find Russians from the other side.

The robot has specific commands. The LLM prompt contains such commands, where the model is instructed to use by outputting them via text. The output text is then interpreted by the robot.