Roemmele: NSA Hack Exposes Security Flaws, Not Need For AI Gun Licenses

The answers are simple for Mythos the issue is Dario made it a reason to shut down AI research and make his company and a few others the only ones allowed to possess “dangerous” AI.

History is already laughing at how we are being played.

Our enemies will NOT abide and like the path of draconian regulation In the US.

You can’t make this stuff up.

We will pay a price listening to the Pied Pipers Of AI “Safety“.

This is the NSA pulling the same shit as they did in the 90s when they tried to get cryptography export controlled and banned. You were around for that, Brian. Total fesrmongering and misinformation.

Systems that can be hacked will be hacked even faster autonomously. Systems that are secure won’t suddenly become hackable just because of autonomous AI.

Sure, we are finding more vulns faster, but the sky is not going to fall. And the NSA’s story simply did not happen unless the most important infosec operators in the world just admitted to being a band of incompetent retards that shouldn’t be operating the most important infrastructure in the world.

@BrianRoemmele Marc Andreesen tells of how in the Biden administration, a room full of gov brass told them all, "don't bother making an AI company, there will only be a handful of AI (I believe he said "under our control) and

@BrianRoemmele people don't realize the implications of Mythos

@BrianRoemmele @grok how many zero days has Mythos found

@BrianRoemmele And even with all of this, Labour want to go full steam ahead with digital ID.

@BrianRoemmele Sounds like the six month AI pause that is needed is more for Mythos itself to harden governement systems before anything else

@BrianRoemmele Unless we know details of what kind of pentests they did, and we never will, consider this propaganda.

@BrianRoemmele Until otherwise proven I will believe the gov creates the narrative they need. In this case they need to implement ID pass. And ”national security” is the textbook example.

@BrianRoemmele So they then used it on China, right?

@BrianRoemmele “History is already laughing at how we are being played.” I’m so alarmed and aghast, I’m almost speechless. The significance of this..

The regulatory capture theory has one problem: the facts. Washington banned Anthropic from federal use in February, tagged it a supply chain risk in March, then forced the June export control that Anthropic is now fighting in court. Dario did not engineer exclusivity, he got locked out of the market he supposedly captured. The breach point still stands though: if Mythos walked through NSA systems in hours, that hole was already there.

@BrianRoemmele China?

@BrianRoemmele Models may become paid-only and accessible through ID verification. It would be like accessing a supercomputer only authorized people would be allowed access.

@BrianRoemmele It seems the loudest proponents of "AI safety" are actually saying, "wait a minute until we catch up so that we can overcome anything you do." Control and safety are already lost, just like privacy. Pandora's box has been opened. Buckle up.

@BrianRoemmele "Best way to plug a security hole is to kill whatever discovered the hole" is not a viable defense of a threat vector. If you stop going to Drs because they told you about your cancer does not make you cancer free!

@BrianRoemmele Still waiting on a comment Brian