Aug 25, 2018
I really wish the journalist would have reached out to me and learned this is a REAL problem now, not 900+ years from now. Also, the journalist gives credit to a reddit poster, not to me. Finally, the journalist doesn't even know that I, too, have written articles published by Tom's Hardware in the past. Also, the author should consider his source. I think knowing a little more about me would help establish confidence:

Also, who wrote this article?,1126.html

Or this one?,1102.html

To fully understand this Y3K story, please read this article:

Sorry, feeling like there was zero due diligence here, though I am grateful it was considered newsworthy.
Last edited:


Luckily we have 980 years, 10 months and 17 days for Microsoft and Gigabyte to address this problem before it's actually cause for concern.
What about the 2038 bug? It might sound like a long way off, but we're more than half-way between y2k and 2038.

Though it's mostly a UNIX thing, don't pretend there aren't APIs and programs on Windows that use that date format, as well.
May 18, 2020
Hey Nathaniel Mott, are you planning on updating the article with what the Y3K bug is actually about? Clearly, you wrote this article based on very little understanding of the actual problem. Perhaps you didn't think much of researching at the time, but after others point out your mistakes, I would hope you would do something about it. I would hope any journalist would.

The problem is less that you can't set the date something crazy and more the hurdle that Windows makes to fix it. It's even spreadable to other motherboards because of how this bug functions. In the least, it's good for people to know this problem exists so they can deal with it.

On top of that, it shows there's a possible vulnerability that could be exploited if Windows has that much control over your motherboard settings.
Last edited: