Exklusiv: Anthropic würdigt den Test eines neuen KI-Modells, das einen „schrittweisen Wandel“ der Fähigkeiten darstellt, nachdem ein versehentliches Datenleck dessen Existenz aufgedeckt hat

> We consider this model a step change and the most capable we’ve built to date.

Taking their cues from Apple. Releasing our most capable ~~iPhone~~ model yet. Since prior to this we actually released less capable models with each new version, we decided to try releasing a more capable model this time?

I’m noticing it’s not being tested for capabilities in materials science discoveries, childhood leukemia cures, insider trading detection, influence peddling, poverty prevention solutions, childhood hunger solutions, and elderly care robotics. When will AI pay off for society?

I was hoping that the CEO of the company would say that they’re working towards a better AI but just couldn’t do it. /s

> A draft blog post that was available in an unsecured and publicly-searchable data store prior to Thursday evening said the new model is called “Claude Mythos” and that the company believes it poses unprecedented cybersecurity risks.

It’s so hard to read this next to the incredible work they’re doing otherwise lol. TmI guess this is what happened when some of the smartest developers on Earth vibe code their CMS haha

As they say in showbiz: show don’t tell!

Death comes for us all

This is just a desperate attempt to generate buzz before the IPO. There is no more data, and if there is, it’s most likely generated by AI, which means the quality of the data is going down. We have already reached the peak performance of LLMs.

“The new model feeds off the gonads and intestines of young humans, but it provides an 11% efficiency to AI video creation efforts”