🍇🍷 When Old Compute and Data Get Better With Age 💾💻

Why better AI can make old hardware, data, and code more useful.

Here’s a strange and sometimes counterintuitive feature of the AI era: old data, old code, and even old hardware can become more useful as the models and algorithms around them improve.

I wrote about keeping AI logs in a ‘black box’, so that better models in the future can revisit them and possibly catch things that today’s models missed—clever cheating or other unwanted behavior.

This applies more generally because better AI models can extract more value from things you already have: hardware, data, and code.

An Nvidia H100 GPU is more useful today if it’s serving GPT-5.6 Terra or Sol than it was when it was first made and was inferencing GPT-4o or whatever. The same hardware can now produce more intelligence.

The H100 can be getting relatively worse and absolutely more useful at the same time. ↕ (put that in your depreciation spreadsheet!)

As intelligence gets better and the cost of a given level of capability falls, it changes the value of everything around it. Some existing assets become more valuable because AI can extract more from them, while others become less valuable because AI makes them easier to reproduce.

Similarly, the same pile of data can be more valuable when it’s analyzed by a smarter model. And when training an LLM, better algorithms and architectures can extract more capability from a fixed dataset. (Some data goes stale, of course, but unique, well-preserved data may actually appreciate over time.)

You can apply this in your own life.

If you had AI analyze important documents or work on a thorny problem a year ago, maybe have it analyze them again today with a smarter model. The older model may have missed connections that today’s models can see.

If you had AI build some software for you even six months ago, you should have today’s model look at the code and see what it can improve or what bugs it can fix. It could make a MAJOR difference.

That doesn’t mean all these old assets become more valuable in a financial sense. Better AI can make something more useful even as it makes similar things cheaper and easier to create. Old code may become much easier to improve even as writing new code gets dramatically cheaper.

That combination is unusual.

Generally in technology, old hardware rapidly becomes obsolete and worthless (or at least, worth less), especially during periods of fast improvement. I remember in the ‘90s, a PC just a few years old felt really ancient and was much worse than a new one.

Until recently, the analytical intelligence available to you didn’t become much better every six or twelve months. You could revisit the same material with a different person or a better question, but the intelligence applied to it wasn’t improving on a software-release cycle. 🕵️‍♂️👨‍🔬👨‍🔧…🤖

We’re in a new era right now. Some old work deserves to be revisited regularly. And that gives you another reason to preserve more of the raw material: the documents, logs, data, code, transcripts, notes, whatever. Today’s analysis may become obsolete long before the raw material it was built from does. 🗄️🗄️🗄️

🧭 This first appeared in Edition 652 of Liberty’s Highlights. New here? I made a page for that: Start Here.