> "I appreciate your anger, but it’s misplaced here."

Weird mix of narcissism to think anyone cares if you appreciate their anger, and arrogance to think you know where it belongs.

You're definitely confused. We can leave it at that.

The cop comfortable with pulling over the funny looking car (that looks like a banana) is the same cop with unreasonable understanding of the responsibility they're given.

It wouldn't matter if the banana car man has a sign that says "I love cops free donuts in the back"

What's really crazy to me is that someone is working over time to try to police (literally) other people's negative feelings about cop misconduct.

What inadequacy leads someone to see others upset at an obvious misuse of power and think "I need to stand up for the guys misuing it!"

Are you under the impression all cops are known for is harassing banana cars?

And if that sounds hackneyed and like a ridiculous standard, you're damn right it is: we let them have outsized influence in our existence as otherwise free people. Their standard has to be a double standard.

If you mop floors and you have fun by twirling your broom and humming a tune, you're not affecting anyone.

If we give you a gun and the right to shoot people in the head and go home to sleep in your own bed, then we can ask you to lock in a little more than that and not pull over people because it's funny.

No it's not. Generating an SVG is asking the model to write text that can be rendered as an image.

Generating an image directly skips the intermediate step and directly outputs an image: so the accuracy is in a completely different league. The models people ask for SVGs typically cannot do this, only specially post-trained variants can.

(An LLM will do the svg of a pelican on a bike much more accurately btw.)

Even if you don't train on gore that's bad enough to trip an image classifier, the model learns the concept of "more [liquid/jam/syrup/chunks/etc.]" and that can generalize to creating gore that would trip the same classifier.

But Anthropic at least has openly admitted they try to detect that and interfere

The reality is a lot of interesting, trivially harmful to non harmful things are illegal and we still do them anyways.

You were always able to reserve a normal icloud email address just like you would a GMail account, so banning all icloud email addresses would be banning non-alias Apple customers

That being said, I'm not convinced anyone who wanted to ban aliases couldn't have already. The alias emails look weird enough I'm guessing you could ban them with few false positives.

I've had simple prompt engineering tasks that cause 4.8 to clamp down. In the past "browbeating" it (read: a sentence telling it not to read the task in bad faith) was enough.

Now it digs in and starts ranting about why it won't capitulate, I'm actually wrong, etc.

Extremely frustrating, and it became a problem with Opus 4.7 because they're trying to make up for the downgrade in parameter count with more RL, but RL does relatively poorly with non-trivially verified things like nuance in instructions.

But overall, not sure why the tone of these replies: then Venn diagram of "wants to rob people" and "cares Google's AV will record it" doesn't include as much overlap as you're implying.

A Waymo has even been used as a getaway vehicle a few times now, once even successfully

It's only the direction that has direct potential business impact they've decided to sabotage instead of reject.

gpt-5.5 isn't larger than gpt-5.4 but costs double.

60 days.