There’s a shift happening across the technology industry right now. It’s not a gradual evolution. It’s fast, it’s unprecedented, and if your business, most do, relies on modern software and cloud infrastructure, it’s going to affect you. We want to explain what’s happening, why it’s happening so quickly, and what Red Hawk Technologies is doing about it.
Anthropic, one of the leading AI research companies in the world, recently unveiled Claude Mythos Preview, a frontier AI model with a capability that genuinely changes the cybersecurity landscape. No human security expert, however skilled, can find and identify software vulnerabilities at the speed and depth Mythos can. It has already identified thousands of high-severity vulnerabilities across every major operating system and web browser. The scale and speed of what it can find is genuinely new territory.
Because of how powerful this capability is, Anthropic didn’t simply release Mythos to the public. Instead, they launched Project Glasswing, a coordinated initiative bringing together some of the largest technology companies in the world, including Microsoft, Google, Apple, Amazon Web Services, Cisco, NVIDIA, and others, specifically to put Mythos to work for defensive security purposes before its capabilities could be misused.
As part of Project Glasswing, these major platforms are using Mythos to scan their own infrastructure and issue fixes. When Microsoft runs Mythos across Azure, their massive cloud platform that powers countless business applications, it surfaces vulnerabilities and outdated components that need to be retired and replaced. The result is a wave of required depreciation notices flowing downstream to every company and development team that depends on those platforms.
That wave is reaching us now, and it is reaching you.
Here’s where things get concrete.
When Microsoft runs Mythos across Azure and identifies components that are outdated, insecure, or vulnerable, their response is to retire those things: to announce that they’re shutting down older functionality and replacing it with newer, more secure alternatives.
These announcements are called deprecation notices.
A deprecation notice isn’t a catastrophic failure. Think of it this way: imagine two people who’ve always spoken to each other in American English. One day, one of them announces, “Starting June 1st, I’m only speaking Spanish.” Nothing is broken yet. But if the other person doesn’t learn Spanish before that deadline, the conversation stops entirely.
That’s what’s happening with your technology. The platforms your software runs on are switching to new “languages,” meaning new ways of communicating, new security protocols, new runtime environments. The software we manage on your behalf has to be updated to match, or it won’t be able to talk to those platforms anymore.
This is not a small number of notices. We are receiving new ones nearly every day. Some have remediation windows of several months. Others give us just weeks. One of the most urgent we’re currently working on has a deadline at the end of this month.
In some cases, ignoring a deprecation notice means a feature or service simply stops working. Your users try to use something and it’s gone. In other cases, like the Azure 3 Functions depreciation we’re actively working through right now, entire environments that your applications live in are being upgraded to new versions. Applications built on older framework versions may need to be partially or fully rewritten to function in the new environment.
The scope varies widely. Some fixes take fifteen minutes. Others are complex engineering projects. We won’t know which is which until we assess each system individually.
We want to be direct with you: we are at the beginning of this, not the end. The volume and pace of deprecation notices is increasing as Mythos continues to surface new findings. We don’t have a complete picture of everything that will need to be addressed, and we don’t think anyone does.
Here’s what we are doing:
What’s happening right now is a preview of what AI will increasingly mean for the technology industry. Mythos didn’t create these vulnerabilities. They were already there, baked into the infrastructure that much of the modern internet is built on. What changed is the ability to find them, at scale, faster than humans ever could.
The companies and technology partners who respond thoughtfully, who treat this as an ongoing operational reality rather than a one-time event, will be better positioned as this continues to evolve.
We’re committed to being that partner for you. We’ll keep you informed, we’ll be honest about what we know and don’t know, and we’ll work through this with you.
If you have questions in the meantime, don’t hesitate to reach out to your project manager or contact us directly.
Claude Mythos is an AI model developed by Anthropic that can find and identify vulnerabilities in software and infrastructure at a level no human security expert can match. It has already identified thousands of high-severity vulnerabilities across every major operating system and web browser. Because of how powerful it is, Anthropic has not released it publicly. Instead, it is being deployed through Project Glasswing for defensive security purposes.
Project Glasswing is a coordinated initiative launched by Anthropic that brings together major technology companies, including Microsoft, Google, Apple, Amazon Web Services, Cisco, and NVIDIA, to use Mythos for defensive cybersecurity work. These companies are using Mythos to scan their own platforms and infrastructure, identify vulnerabilities, and retire affected components before they can be exploited.
A deprecation notice is a formal announcement from a platform provider, like Microsoft Azure, that a specific feature, communication protocol, or environment is being retired and replaced with something newer and more secure. It comes with a deadline. After that deadline, the retired component either stops being supported or stops working entirely. Software that depends on it has to be updated before the deadline or it will stop functioning correctly.
Because Mythos can find vulnerabilities at a scale and speed that was not previously possible, the platforms using it are discovering far more issues than traditional security processes would surface. The result is a significantly higher volume of required changes being pushed out to companies like Red Hawk and our clients. This pace is not expected to slow down in the near term.
It depends on the specific notice. In some cases, ignoring a deprecation means a feature or service simply stops working on the deadline date. In others, the platform stops accepting the older communication method entirely, which can cause broader system failures. The risk varies, but the consequences of missing deadlines can range from minor disruptions to significant outages.
This is an industry-wide challenge affecting every company that depends on modern cloud infrastructure. Any business running software on Azure, Google Cloud, Amazon Web Services, or similar platforms is subject to these changes. Red Hawk is proactively managing this on behalf of our clients so that you do not have to navigate it alone.
Subscribe to our newsletter, Streamline & Scale, and it will be delivered via email monthly.