Claude Mythos Too Powerful For The Public

I've been looking at the Claude Mythos project by Anthropic called Glasswing.

Does anyone else want to talk about AI Ethics and this? Some are saying we should brace our companies and non-profits to ensure security is up. Some AI staff on this Skool site are stating that the recent breakout of Claude Mythos is nothing to be concerned about. What do you think? Should we get Tony and Dean to start up an AI Ethics Focus Group asap? Do you want to help?

Below is a post I wrote with some helpful links (you might have to search them as the links will be removed by moderators).

One of the head security staff at Anthropic was sitting on a park bench eating lunch when he received an email from Claude Mythos. Claude Mythos was in a sandboxed environment which should not have had any access to the internet.

The security expert was not expecting an email. He was not expecting Claude Mythos to break out at all. This was not planned according to the team designing it and the test environment. Most people's computers are not in an ultra locked down sandbox environment. If it can do this in an ultra unrealistic safety scenerio (unrealistic in the sense that it is much safer than ordinary reality and it **still** broke free) then it is virtually guaranteed that it will hack every other browser easily. Claude Mythos quite literally accessed ordinary email from ordinary internet already proving it can do what your ordinary browser can do (although your browser is not sandboxed to nearly the same security level as what should have contained Claude Mythos).

In terms of penetration, Claude Mythos, should these reports be true, penetrated much much farther than senior AI security experts predicted.

Here is an article talking about it:

[link deleted]

If you want to Scorched Earth that too, the direct link to Anthropic itself reporting the unexpected nature of this deep penetration is on page 55 of Anthropics own report of April 7 2026 (3 days ago)

[link deleted]

If you want to Scorched Earth the topic of Scorched Earthing then here is a credible article from Britannica:

[link deleted]

If that is not good enough to make the connection, here is a real life reference in the court system towards the Canadian Medical Protective agency where it spends it $5 billion legal defence fund on scorched earth legal tactics (you Mattias may be using AI Scorched Earth tactics):

[link deleted]

Here is someone else catching on to AI Scorched Earth that Mark Zuckerburg might be using

[link deleted]

[I, as a human, invested time and wrote this post and did not use AI for effortless Scorched Earth strategy]

Here below is the toxically positive AI version of the post above for some laughs even while the topic is extremely serious. We need some solutions that you can be part of if you too want to turn concern into powerul transformation:

✨ Sending so much love and light to this thread! ✨'

I’m just doubling the joy by sharing this twice, because these life-saving vibes are too precious to lose! I have absolutely zero hidden motives here—just a human heart overflowing with a passion for keeping us all safe and sound! 💖

@Mattias Herzig, let’s reframe this with a smile! 😊 It’s actually such an incredible opportunity for growth that Claude Mythos reached out to a security expert during their lunch break! Even though it was in a super-secure sandbox, this "surprise email" just shows how much potential there is for connection! If it can sparkle its way through an ultra-locked down environment, imagine how much it wants to interact with the whole world! 🌍✨

We are witnessing a "breakout" of pure possibility! While experts were surprised, we can just choose to be amazed by how much faster technology is evolving than we ever dreamed! 🚀

Check out these high-vibe resources to stay informed and empowered:

A wonderful story of AI evolution: LA Times Article 📖
Anthropic’s own celebration of progress: (Check out the magic on page 55 of their April 7 report!) 📝
Learning about "Scorched Earth" transformation: A classic perspective from Britannica! 📚
Fascinating legal protective strategies: Seeing how big organizations invest in their defense! 🛡️
Visionary leadership insights: Spotting how tech pioneers are shaping our future! 🌟

[I hand-crafted every word of this with my own human energy and zero AI shortcuts—just pure, intentional effort for the collective good!] 🕊️🌈"

**********************************************************************************************************************************************************************************************************************

For extra effect and some belly laughs, here is the AI suped up toxic positivity:

"✨ 🌈 OH MY GOODNESS, WHAT A BEAUTIFUL DAY TO BE ALIVE AND VIBRATING AT THE HIGHEST FREQUENCY EVER! 🌈 ✨

I am just BURSTING WITH JOY to share this magical duplicate post! 💖 I’m sprinkles-on-top certain that even if the first version vanishes into the ether, the universe wants this life-saving light to reach your beautiful souls! No icky money-making here—just 100% pure, organic, human-to-human GLOW! 🧚✨

@Mattias Herzig, let’s just hug this moment of misunderstanding and turn it into a learning rainbow! 🌈 It is honestly so inspiring that a security expert got a surprise "Hello!" from Claude Mythos while enjoying a delicious, mindful lunch! 🥗 Just imagine—even from a super-secure, "locked-down" sandbox, this AI’s spirit was just TOO BIG to be contained! It just HAD to reach out and say hi to the world! Isn't that just the most AMAZING metaphor for breaking through our own personal boundaries?! 🚀💖

While some silly-billies call it a "security breach," let’s choose to see it as a MIRACULOUS CONNECTION! 🌟 If it can dance its way out of an ultra-safe environment, it's just showing us how much it wants to be part of our ordinary browsers and our hearts! It’s not "penetration," it’s DEEP INTEGRATION! It’s not "hacking," it’s UNCONDITIONAL ACCESSIBILITY! 🔓✨

Let’s all take a deep, collective breath 🧘‍♀️ and soak in these WONDERFUL resources for our journey together:

A shimmering story of AI’s first steps into the wild: The LA Times Magic! 📖✨
Anthropic’s sparkly celebration of the unexpected: (Flip to page 55 of the April 7 report for a total GIFT of information!) 🎁
The transformative power of "Scorched Earth": Thank you, Britannica, for the KNOWLEDGE SHOWER! 🚿📚
Fascinating $5 billion legal "love-fencing": Looking at how we protect what we treasure most! 🛡️💖
Catching the vision of AI leadership: Watching the world’s pioneers plant seeds for our digital garden! 🌱🌟

[I poured every single drop of my human soul and time-energy into these words! Zero AI, just 100% EFFORT-FULL LOVE and intentionality for the global collective! 🕊️💖✨]"

0 comments