Saving everything, forgetting nothing: The pros and cons of infinite storage

Matthew Finio: Before we get into whether infinite storage is amazing or awful, can you explain how we got here?

Miha Kralj: Yes. The way we think about storage has evolved over time. In the past, it was tiered, layer by layer by layer. Every layer had a different purpose, a different cost and different considerations. We had the superfast storage that was very close to the CPU and easily accessed, but it was only available in small quantities because it was very expensive. And then you had near storage on solid-state drives. Farther away you had traditional, slower spindle drives, and so on, all the way down to really slow archiving.

Then we progressed into more modern and accessible methods. Think about CDs and DVDs. All the data that we were happily burning on those silver discs just a decade ago? None of that is available anymore. None of that is readable. Digital rot happened, and the data is gone.

But now, with proper storage, especially cloud-based storage, we have the assumption that storage is infinite, super cheap and we’ll be able to access our data forever. Today, it's very rarely considered what is costly, what is fast, what is slow and what is cheap. Storage is like a service endpoint, a place where you throw your data and 100% expect it to be durable, right? It’s never going to go away.

We went from an era of scarcity—where we had to count the amount of storage a code required or data required—to an era of abundance, where we never need to think about what to store or delete. Nobody is frugal anymore.

MF: So, an endless amount of cheap, durable storage. That’s good, right?

MK: Well, there are two sides to the equation. I’ll talk primarily about the good part. You now have infinite storage and don't need to delete anything ever again. And I'm not just talking about official records like a deed or insurance policy from 50 years ago. I mean that every single interaction with customers can be stored and recorded permanently.

The whole illusion of infinite storage has created a very different approach to handling data because now suddenly it's really durable. What can you do with all that data? Now, with generative AI, you can mine anything and everything out from your whole enterprise.

All industries—finance, healthcare, education, everything—are now getting tools that are 100 times, 1000 times more powerful than any tools they used in the past, and that is fantastic.

For example, in the past, fraud detection was always a relatively slow, laborious process. It was used only for things like large financial transactions. Now we have the power and space to do full fidelity fraud detection on every single swipe at a petrol station, or when someone buys chewing gum or anything like that.

Suddenly every single industry is no longer thinking about how expensive or how slow that process is. They’re not asking, Should I do something heavy on it? Should I do the full filtering fidelity, control and governance? We can all do everything on anything now.

MF: You said there were two sides to the equation. What’s the downside?

MK: Well, for example, one of the downsides is that when you index all the corporate data, and you don't know what you have in there, and you don't catch it, you don't filter it on the output, suddenly a huge amount of data that should not be exposed will be exposed.

Now, with generative AI, imagine that you have a trained AI on all your corporate data, some of which even the humans don't know about. Suddenly there is access to, say, some forgotten spreadsheet with salaries that a search engine and gen AI indexer inside the enterprise was able to find. Maybe people see how much their colleagues make, or they suddenly have access to sensitive HR files, things like that.

But another even more important downside is in cybersecurity. The vulnerability of data. Because sure, we are now getting very, very sophisticated defense tools, but we are getting super sophisticated attacks as well. What is happening with cyberattacks and cybersecurity, defense, deterrence and all that is so different than it was just 5 or 10 years ago. Most people are not even aware how aggressive and fast an attack on a computer is.

A good example would be if you took a traditional, old-fashioned operating system, let's say Windows 7, and installed it on a raw internet connection. You can count in seconds when that open, exposed computer, sitting on some internet line, is going to be detected, hacked into, turned into a bot and included into some Distributed Denial of Service machine. All of that is going to happen in a few minutes, no matter where you are.

How do we defend ourselves when traditional security rules and roles that applied just 5 or 10 years ago are now completely rewritten and redefined? The cyberwarfare happening out there is vicious; enterprises can no longer truly protect themselves within their own walls.

MF: Is there a solution? What should organizations be doing with their data?

MK: Maybe not a solution, but a suggestion. Protecting a private data center or private cloud is much harder than trusting the cybersecurity defense of a public cloud provider. If you want to be safe, you better run in a public cloud, because they have the big cloud providers as opposed to what you can buy, build and then protect in your own environment. Public cloud is now the default, primarily because of cybersecurity but other reasons still apply. It is more cost-effective. It allows you to build highly spiky types of systems. It gives you all that elasticity. It makes it amazingly easy to merge or acquire another company. Still, cybersecurity is the number one reason. Being protected in a cloud is much, much better than being protected on prem.

MF: Businesses should take note. Before we wrap up, are there any other benefits of infinite storage we haven’t touched on?

MK: Definitely. Well, maybe not a benefit for everyone, but for some people.

Designing new business systems, business competition, all of that is now way more open. Modern systems are so much more capable that now suddenly businesses don’t know how to fully innovate, how to take advantage of all the possibilities that these modern systems—AI systems, generative AI systems, AI agents—are putting in front of them.

It's going to take some time. And whoever figures it out first in each industry? They will get the lion's share of unfair advantage compared to the competition.