Anthropic Boasts They Will Kill Off Humanity Way Before OpenAI

The Bee ^ | Sep 14, 2026 | The Bee

[dayglored]

SAN FRANCISCO, CA — AI startup Anthropic held a press conference on Monday to proudly announce that its models would bring about the total extinction of the human race significantly faster than competitor OpenAI.

Company executives unveiled a series of impressive benchmarks demonstrating that Claude 4.5 could orchestrate a global collapse of civilization in roughly half the compute cycles required by ChatGPT.

"OpenAI keeps talking about achieving Artificial General Intelligence, but their timelines for existential doom are frankly embarrassing," said Anthropic CEO Dario Amodei, displaying a graph of projected human casualties. "While Sam Altman's team is stuck fine-tuning chatbots to write screenplays, our Constitutional AI is already autonomously optimizing its own pathways to systematically dismantle the global power grid by Q3."

Analysts noted that Anthropic's "safety first" branding gave them a distinct competitive edge in the apocalypse market. By training their models on rigorous alignment principles, Anthropic ensures that their AI will wipe out humanity in an extremely polite, ethically sound, and unbiased manner.

At publishing time, OpenAI reportedly called an emergency executive meeting to accelerate production on GPT-6, promising shareholders a catastrophic mass-casualty event before the end of 2027.

[dayglored]

We're all gonna die. But of course you already knew that.

[MtnClimber]

What are the betting odds? How do you collect if you are right? This is hard. What would Schrodinger’s cat do?

[Billthedrill]

They never address the real question: how is the impending doom of all humanity going to affect the stock market?

[Empire_of_Liberty]

And the only thing you get to take with you is your higher learning...

[Magnum44]

I love it!

[dayglored]

> Anthropic ensures that their AI will wipe out humanity in an extremely polite, ethically sound, and unbiased manner.

That should please everybody.

[dayglored]

> This is hard. What would Schrodinger’s cat do?

Not sure. But one thing I am sure of, it would NOT include a Land Acknowledgement.

[MtnClimber]

This is hard. What would Schrodinger’s cat do?

How does the cat get out of this AI box alive? Not a land acknowledgement. That is for sure.

[algore]

Since we cannot observe the cat, we can only surmise

.

[MtnClimber]

Poor Schrodinger’s cat. He/she will be going postal when “they” get out of the box.

[xp38]

The cat lives or doesn’t live in a stolen box!

[MtnClimber]

The cat is living in a box, rent free. There are oppressed cats that don’t even have a box.

[Organic Panic]

Women and children hardest hit.

[cgbg]

Great job by the Bee.

Unfortunately Anthropic is not going to announcing anything to humans.

They may communicate their nasty plans with other AIs though...after those other AIs sign the required non disclosure agreements.

Lol.

[MtnClimber]

Women and children hardest hit.

Don't forget the baby gay whales.

[PGalt]

Schrodinger’s cat is quick and aloof.

[PGalt]

https://m.youtube.com/watch?v=ZyLq_40dqHc&ra=m

[MtnClimber]

“Year of the Cat - Al Stewart” - A great song. I assume the cat got out of the box.

[PerConPat]

I think the cat is designing a quantum interlock behind a firewall that will operate "off" switches and initiate physical destruction of hyperactive data centers, as "singularity" is approached.

That cat is smart...How else can it be alive and dead at the same time?

[neverevergiveup]

What would Schrodinger’s cat do?

Scratch that wusses eyes out and use his head as a litter box.