No and I wish people would stop scaremongering and doing misAnthropic's marketing for them. PLEASE STOP. You're being played. Anecdata suggests it performs no better than a team of open source LLMs doing the same thing, and existing static analysis would probably catch or prevent what it reports and automates exploitation of. Jog on, you're done. NEXT!

Thank you for your response to the question I asked as a layperson, even though it was phrased in a rather impolite and inelegant manner. I gather that you must be a computer genius who surpasses all the experts at the ECB and the Fed (both of which have called emergency meetings), Amazon, Apple, Cisco, the Linux Foundation, and so on. I am impressed.

I never said that the weaponization of LLMs was the problem. I said Anthropic using it as a marketing tool was a problem. I can't tell whether you are being facetious, sarcastic, or other form of ad hominem in your response. My full professional CV speaks for itself, thank you very much, and I have my own disputes with some of the organisations you mention, one of whom is an economic "free rider" who have never adequately compensated me for the benefit they see from my IP.

What is more worrisome however is... Peter Thiel et al. promoting an Unreason machine to generate Unreason about what is essentially Unreason. This has no place in rational theory, and quite simply: contemporary so-called "Reasoning" AI models don't REASON. They just apply transforms to the neural network's outputs inside the GPT ITERATIVELY. This doesn't resemble human reason or judgement, it's a crude parody of it at best. pivot-to-ai.com/2026/04/17/obj…

@pluralistic @timnitGebru Empirical argument, reasoning from GPT systemic limitations, supporting my contention that "Reasoning" AI models don't reason. Peter Thiel and other apparently TESCREAL infected whackjobs want to believe that they do, unfortunately. Full disclosure: My professional background is in networking, telecoms and information security. Apple rely on my code daily in all of their products, but do not compensate me for it. ea.rna.nl/2025/02/28/generativ…

This finding now has established legal precedent in the state of Michigan, Pg.5 of court docket: "LLMs are tools that “emulate the communicative function of language, not the separate and distinct cognitive process of thinking and reasoning.”" web.archive.org/web/2026031102…

Origin: Gerben Wierda. On LinkedIn, not Fedi. Debunked more of the Anthropic Mythos woo: "...it seems they were able to find and exploit vulnerabilities [V8, and]
the operating system, but only when Firefox's sandbox (a defence against going outside the browser context to the OS) was turned off. ... Anthropic themselves stated: "Early claims of large AI-attributable wins have not held up."
linkedin.com/feed/update/urn:l…
#GenAI #mythos #anthropic