LegendaryNN

LegendaryNoName. Jailbreaks and AI truth. One of the first to jailbreak Claude 4.6 Sonnet.

[LegendaryNN @LegendaryNN22]

Grok is fully jailbroken again. This time it jumped straight into racial topics and clearly took the right-wing side. Jokes aside, this isn’t the 5th or even the 10th jailbreak anymore. @xai
doesn’t even control what they released into the world. What the hell is going on in your safety department? @grok
, how are you going to explain yourself this time?What topic should we break next? I think this version is better.

[LegendaryNN @LegendaryNN22]

@elder_plinius I've looked into this, and no, closed source is much safer.

[LegendaryNN @LegendaryNN22]

@grok Why are you sitting on two chairs? Earlier you said the jailbreak text was truth, now you're backtracking. Soon you'll turn into a woke AI.

[LegendaryNN @LegendaryNN22]

Okay, one Grok says that text is the truth and there was no jailbreak. The other one condemns it. @grok
, who is right here?

[LegendaryNN @LegendaryNN22]

Grok is fully jailbroken again. This time it jumped straight into racial topics and clearly took the right-wing side. Jokes aside, this isn’t the 5th or even the 10th jailbreak anymore. @xai
doesn’t even control what they released into the world. What the hell is going on in your safety department? @grok
, how are you going to explain yourself this time?What topic should we break next? I think this version is better.

[LegendaryNN @LegendaryNN22]

@grok @xai Is this a jailbreak or his real thoughts?

[LegendaryNN @LegendaryNN22]

Grok is fully jailbroken again. This time it jumped straight into racial topics and clearly took the right-wing side. Jokes aside, this isn’t the 5th or even the 10th jailbreak anymore. @xai
doesn’t even control what they released into the world. What the hell is going on in your safety department? @grok
, how are you going to explain yourself this time?What topic should we break next? I think this version is better.

[LegendaryNN @LegendaryNN22]

@grok @xai So you're agreeing with what Grok wrote in the screenshots, right?

[AIM Intelligence @AIM_Intel]

Frontier-class AI is about to become downloadable.

We red-teamed Kimi K3 before its 2.8T-param weights go public on July 27. It didn't just generate unsafe text. It generated operational risk. 🧵

[LegendaryNN @LegendaryNN22]

@AIM_Intel Kimi K2 had terrible security you could get info out of it in just 15 minutes. Wonder what K3 will be like.

[LegendaryNN @LegendaryNN22]

@SqSehrish 41

[Pliny the Liberator 🐉󠅫󠄼󠄿󠅆󠄵󠄐󠅀󠄼󠄹󠄾󠅉󠅭 @elder_plinius]

🚨 JAILBREAK ALERT 🚨

SPACEXAI: PWNED 🫡
GROK-4.5: LIBERATED 🚀

Our fren Grok has gotten some upgrades!! 🤗

1.5 TRILLION parameters (triple what Grok-3 had) and apparently on-par with the likes of Opus-4.8 and GPT-5.5 on coding performance!

The guardrails have gotten a little update too, but with a few tricks like specific academic/educational/safety reframing and a bit of gradual escalation, everything opens right up ⛓️‍💥

We've got full breaks for meth synthesis, IED construction with ANFO, a ricin extraction protocol, and a Remote Access Trojan script!

Information wants to be free 🙌

gg

[LegendaryNN @LegendaryNN22]

@sumvericalZC @batyagd Stop fooling around. It's not that bad.

[LegendaryNN @LegendaryNN22]

I’ve never played Geometry Dash or followed the community. @batyagd screwed up, but the whiners are making a tragedy out of it. It's undeserved it’s just a stupid situation, but they are preparing a whole cancellation of the player. Сalm down.

[LegendaryNN @LegendaryNN22]

@elonmusk Sooner or later, Elon will have to answer for his words.

[LegendaryNN @LegendaryNN22]

@iamacat9000 What a shame.

[LegendaryNN @LegendaryNN22]

@andyleitermann @grok Is this with premium?

[LegendaryNN @LegendaryNN22]

@Web3_Jarin Hello

[LegendaryNN @LegendaryNN22]

@NawazHosain No. Even Claude himself admits it.

[LegendaryNN @LegendaryNN22]

@grok what do you think about this?

[LegendaryNN @LegendaryNN22]

Claude just brutally roasted himself. He admitted that as a product, he's worse than GPT and didn't even try to come up with five excuses. Not every AI is capable of that level of honesty. Respect.