284 - Legacy Systems

As the use case of "Port and Verify" was recently brought back to my attention by Lloyd Watts, I did a bit of digging for some recent facts and figures in a place I knew I'd be sure to find it, US Government spending. As expected, legacy languages like COBOL and the critical systems that rely on them were costing the US over $330 million per year according to one major report.

Even 4 years after that initial report 2 out of 10 of the agencies noted in it still hadn't come up with so much as a "plan" for modernization of their systems, let alone starting the work, or for that matter actually completing said work. If they attempt to apply "Generative AI Coding Assistants" to this it will likely end very badly, and the risk of that is substantial, as the US has been known to swallow the bullshit that Microsoft feeds them with fair regularity.

Remember when the US military signed a $21.88 billion deal with Microsoft for their VR headsets?

The safe bet is that these numbers are substantially and systematically underestimated, and they certainly aren't limited to the US. Divisions of the British and Dutch governments, among many others, have attempted modernization of technology and methods to little or no avail on and off for many years. All of the funds blown on those failures to modernize can be considered losses on top of the losses incurred by legacy systems themselves.

These legacy systems are of course inherently more vulnerable, with fewer people skilled in maintaining them, but even while systems like LLMs can't be expected to modernize them, they can assist in creating more effective cyberattacks against this critical (and often military) infrastructure. Just how many of these systems have already been quietly compromised should become rapidly and clearly evident if any country with them gets into a war with another technologically skilled adversary.

As my team continues our work on preparing to demonstrate the capacities of ICOM-based systems in an increasing variety of ways, some benchmarks and some use cases, this will be among them. We had already planned on the use cases of organizational management and governance, so the updating of legacy technical systems pairs well with the updating of legacy management and governance systems. Our systems have also demonstrated an antifragile aptitude for countering bad actors, so the modernization of systems could further integrate with them to provide unique security layers that grow increasingly resilient and adaptive over time.

Some ".mil" domains don't even require a login to begin accessing data, and they look like they haven't been updated since 1995, with security certificates 5+ years expired. While the US may lead the world in military spending, it is obvious where that money isn't going. (GAO Report)

Earlier this month, a report on FAA Legacy Systems was released.