3427: Can Smarter Compression Help Save the Planet? CompressionX Think so

[00:00:04] What if the way that we store data is quietly shaping the planet's future, with data creation exploding towards 510 zettabytes by 2030, and data centers already consuming more power than the entire UK? I think the question is no longer whether we need to change, but how. Well, my guest today, he's the co-founder of CompressionX, and he believes that lossless

[00:00:34] compression could be one of the most overlooked tools in tackling both rising storage costs and digital carbon footprints. So, quick question for you all, can smarter compression really help slow the tide of global emissions while giving businesses faster, cheaper access to their own information? Well, there's only one way to find out. But before I get today's

[00:00:59] guest on, I want to give a quick shout out to Careerist, who are the sponsor of Tech Talks Network this month and all eight shows that we have here. And it's a message that I'm passionate about. So, if you've been thinking, I want to work in tech, but I don't know how to code, then this is for you, because Careerist has built up one of the most accessible paths into tech. Their software QA Engineering Bootcamp is a great example, because in just four months,

[00:01:27] you'll get live training from experts at Google and Meta, not to mention hands-on practice, and even a real remote internship to put on your resume. Plus, you'll work one-on-one with a career coach who will make sure that you're job ready. And here's the best part. QA is in huge demand right now. There are 37,000 plus open roles in the US alone, and average starting salaries go up to over $105,000.

[00:01:58] Now, Careerist graduates are already working at companies all across the country, and if you don't land a job within a year, you can get your money back. That's how confident they are. But seats for the next cohort close very soon. So, please check the link in the episode description and see if Careerist is your launchpad into a tech career. But now, let me introduce you to my guest right now.

[00:02:23] So, a massive warm welcome to the show. Can you tell everyone listening a little about who you are and what you do? Yeah, thanks Neil for having me. My name is Nicolas Stavrilou, and together with my co-founder, we have created a product called CompressionX, which is a lossless compression utility with high compression, but it's also very fast compression. So, hopefully as useful as possible

[00:02:49] to the users. And it is quite a unique problem to go out there and solve. So, if I were to take you back in time, tell me more about that personal journey, the origin story that led to CompressionX, from that first idea to turning it into a working product. There's got to be a story there. Well, it all started back in 2007. I was studying in California, and it was a very simple class. We were

[00:03:17] looking at limits approaching the top of a parabola, the maximum of a parabola. And I just saw that they could never get there, and I had this experience, and it sparked off this paradox that I just couldn't shape. I was like, there's got to be a point somewhere, but it's there, but it's not there. And it just kept going over and over in my mind, and I kept seeing it in dimensional space all around me. And it was very interesting, because that Christmas, I came back to England, and all I

[00:03:47] wanted for my brother was an empty book. And he got me this very cool leather-bound parchment paper book. And I started writing this thing down. And as I wrote it down, I kind of got a bit of relief from that noise, that paradox thing. And I started to think that I had something that was perhaps quite important. And in the subsequent years, I was always like, I've got to finish my theory. I've got to write my theory. And I was back in London working at my father's restaurant in the office

[00:04:16] with my good friend Stuart, who wrote all the till systems for the restaurant. He'd been coding since the age of 15. And once I'd finished the theory, I just went into the alcove one day during lunch and just said, got a napkin out and just did some doodles and just literally a couple of circles on the line. And I explained it. And he just looked at him and, oh, that's data compression. Because that wasn't my skill set.

[00:04:43] And it was this very interesting time where I could see the pure maths theory that everything could be compressed into nothing and brought back again. And Stuart would be saying, no, that's not, if you apply that in the real world, it can't happen. You need a header. And there are limitations because of there are, there's eight bits in a, in a byte and there are 64 bit processes. You can never get to the thing, but I, but I always bring it back to my, my maths and we have that

[00:05:13] kind of rapport. And we went to my father at the time, who was the boss and said, look, we've got this idea and we incorporated the company in 2012. And he gave us one day a week to work on it. And it's been a long journey of 13 years, but we finally got our first product on the market as of July. So we made a desktop downloadable version. We made it, we've got a free version for like

[00:05:40] occasional users to, for anyone to kind of try it out. And then a more professional version for those that would be the heavy, heavy, heavy usage. And we're hoping to get it out there to as many people as possible because we think we've got something special, very useful. And I love the story behind it there and how the universe almost put the two of you in that room together. And both of your skill sets compliment one another. And it's kind of, you've created

[00:06:06] something that neither one of you would be able to do in your own right, but together all this comes possible. And I've got to ask that napkin, that leather journal there, do you have them somewhere? They could be worth a fortune in the future. I do. Yeah. I've got, I've got everything that I've ever written and it starts off as exactly that. It starts off with squiggles and lines and pencil written thing. And it ends up in a very neat USB

[00:06:31] stick with the formula, as I think it should be presented and all the versions in between. And again, with Stuart, he's got all the iterations that we had to go through, through all these years. Each one of them has been just as important as the last, even some of them that led us down kind of dead ends and we had to come back on ourselves. The kind of two steps forward, one step back versions, every, every bit of it was crucial to get us to this point where we,

[00:07:02] where we've got something that's actually out there that people are using, which is really exciting for us. It is exciting. And it is such an important topic because I think if we look down our news feeds, we often hear about AI models and cloud infrastructure, but what we don't talk about enough is the sheer volume of data that we're creating now every day. And how do you describe the scale of storage and the scale of energy that businesses are challenged with today?

[00:07:31] Yeah, it's incredible how it's, how it, how it's growing. There were some figures that I saw in, was it 2023? Might have been 2024. It's 120 zettabytes of data created in the world. In 2025, that is apparently going to be 181 zettabytes. And by 2030, they forecast it to be 510 zettabytes.

[00:07:57] So, and they're all on top of the existing zettabytes that already exist. So if you think, if you think about that extrapolated across, across data centers, across different districts in different countries, the impact, the possible impact that can have on energy infrastructure, water supply, because I mean, this is a fantastic, amazing inventions like, like AI should be applauded.

[00:08:24] We should be really grateful that these things exist. Some people liken the AI revolution to that of first discovering electricity and what it could do for humanity. And that's great. But how are we going to actually provide all the infrastructure and make it sustainable and knit it into our way of life without having a big ecological impact? And for example, there's other stats like 90% of total

[00:08:52] global global data stored was created in the preceding two years. So it really is, it can become a problem. But it's up to us to try and create technologies to stem the tide and help make it less of an impact. And it really is an increasing problem. Because although we're recording this episode in the autumn

[00:09:18] here in the UK, or the fall for our friends in the US, in the summer, we had somewhat of a heat wave over here, and there was a water shortage. And one of the things that caught my attention was the UK government asked citizens to maybe think about deleting old emails and photos to save water. And one of the reasons I bring this up is you've highlighted that data storage alone could account

[00:09:42] for up to 8% of the global carbon emissions by 2030. So why do you think this issue isn't getting enough attention compared to, I don't know, maybe other sustainable topics? Because the more we lean on AI and heavy computing, etc, this is only going to increase, isn't it, right? Absolutely. I mean, apparently data centers already consume more than the entire UK.

[00:10:05] That's about 3% of global electricity. And if you think about the, there's been a lot of press, for example, about the aviation industry. But the aviation industry actually emits less CO2 emissions than data centers, which is extraordinary. It is going out there. I think the initiative by the government was going along the right track. If we can delete unnecessary data, of course,

[00:10:34] the, and we have that kind of data responsibility. Like we do, for example, if everyone, when they brush their teeth, they turn the tap off between brushes, then that contributes to a lessening of the amount of water supply being wasted down the plug. In the same way, I can see the government saying, well, delete your emails. And, but the problem with just getting rid of duplicates is

[00:10:59] our data is so valuable to us. We want to keep all the data. We need the data. We want, we want AI models of the future to be able to work on the, all the data that's been harvested. We don't want to get rid of it all. Yes, we want to trim it down and optimize it, which of course would be great practice. But what if we could use a very good compression tool to compress the data down as far

[00:11:24] as possible while it's been created and get in the habit of storing everything compressed, then we can get more out of the data center. We might not need to build the data center as quickly because we're getting more, more out of less. And it's about adding that data mindfulness, contributing person by person and company on company to not only reduce the costs, but also the cost of the environment of

[00:11:50] having it all just sitting there. There was a stat that I saw yesterday where 80% of global data is enterprise data. And up to 90% of that is cold data. It's data that's just stored forever. It's for compliance or whatever it's used for, which means 70% of everything is just sitting there just in case, just because, because of laws and rules and everything. And so if we all compress everything

[00:12:20] as much as we can and we can reduce that, the load, then we're heading, we're heading in the right direction, I think. A hundred percent. And as you said, we need the data. We need to access that data from any device, any location on any network at any time. So when we're looking at compression here and more specifically compression X, what is it that makes it different from some of the legacy tools like Zip,

[00:12:45] for example, and, and tell me a little bit more about your adaptive algorithm and how that achieves up to something like I was reading here, 65% to 90% reduction without data loss feels phenomenal. What you've achieved here, but tell me more about that. Well, thank you. Yeah. Well, the best way to describe it is legacy losses, compression tools, they kind of operate on a one size fits all. So they do the same operation to the data,

[00:13:10] no matter what it is. Whereas our, our algorithm adapts itself to the data. A good example of that. Say you take a picture of a room, a room, it's got, it's got people and chairs and desks and things like that. Well, the legacy compression will just have a look at it and go, I'm just going to compress that all the same way. And we use, we use kind of entropy calculations, all of our special things

[00:13:36] that we do. And we, and, and, and we chop the, the picture up into compressible parts, the more compressible part, the more compressible the, the, the, the parts are, the more we can get out of the compression. And you can take that down to a really granular level if you've got the time, but our, our job is to do it quickly. So we're, we're, we're prefix encoding, which is great for that.

[00:14:04] So it's about how, how much compression can we, can we get while still being very quick and very usable? Because in this world, time is of the essence and people, people want things done efficiently. So if we can beat the, all the legacy compression algorithms and do it quickly, then we've, we've achieved something. We've given people back that time that they've lost. Another interesting thing, and this is where the difference really is, is marked between us and others. Well,

[00:14:32] I mean, first of all, I'll just go one step back. There are two types of compression, lossless compression and lossy compression, as we know. So with lossless compression, you can compress an MP3 or a JPEG, and you might not get as high compression as if you use a lossy technique that would say, take away a frame or it didn't, didn't care what the output, whether the output was,

[00:14:58] was, was, was perfect or not. So we do that competitively compared to other lossless compression algorithms. But where the real magic happens is on structured data. So think of fixed width data. So data from sensors, logs, CSV files, anything where we, where we have a mostly fixed, fixed output.

[00:15:23] Our algorithm is really, really a lot better than other lossless compression algorithms. Like I say, getting CSV files down to 93% compression, things like MRIs at 20% better, x-rays 7% better. And there are other file types that we, that we excel on. And that's really interesting because that kind of business data is where, where we think the biggest gain is.

[00:15:48] And for the bean counters listening, they're going to be concerned with the ROI of any new technology. And I would imagine that smart and secure compression could save space, cut costs, and also help businesses reduce its carbon, digital carbon footprint, and which again, is incredibly important right now with ESG scores, et cetera. But anything you'd point out there around

[00:16:14] the ROI of a solution like this, because you're not only fixing a problem, but also saving right too. Yeah, sure. Well, every time, every gigabyte of data that's not stored on the cloud is, it's savings. It's savings to, it all adds up to not needing to buy another server, not needing to spend quite as much in the cloud, not needing to maintain the cost of,

[00:16:39] of having these data centers with all their cooling and their energy uses. So there's the obvious cost saving, but it's also, we believe that you should just have these, have the best possible data practices. So optimize your workflows, find out which data should be stored cold, maybe, maybe automate it so that you have say every, every files that haven't been used for 30 days

[00:17:05] automatically compressed to cold storage, like get rid of duplicate files, that kind of responsibility. Because once the data is stored small, it can be transferred quicker, obviously. It can be, so it's more readily usable by people. It saves time. There's less for the kind of the IT tech teams

[00:17:32] to, to, to jump in and kind of help people things with. I think everyone's been there trying to email a file that's, that's too big and phoning up the tech support on a Friday afternoon, you know, he's like, Oh, I have to do, I have to do this. Well, if we can help to alleviate any of that pain, kind of streamline people's businesses, then, then, then that'd be fantastic. Yeah. And as someone that creates a lot of podcast interviews every week and

[00:17:56] sharing out two gig video files, et cetera, it makes a big difference. And you've argued here, I mean, beyond enterprise IT, you've spoken a lot around the responsibility that each of us share as individuals in reducing our digital footprint. So what role do you see both individuals and organizations alike playing in? And how can small actions like compression make a real measurable

[00:18:22] difference as the old IT belts and braces approach? If you can only improve what you measure it, how could compression make a real measurable difference in your eyes? Well, if people have the same attitude to, to their data, as they do, for example, to their air miles in the aviation industry, a lot of, a lot, a lot of thoughts gone behind that. And companies are wary about meeting their, their obligations when it comes to the carbon and carbon emissions in regards to,

[00:18:51] for example, flights or, or the amounts of carbon emitted by the cars they drive or whatever. If we have the same mindfulness when it comes to data, data storage and data transfer, so it becomes second nature to compress everything when it's created and to, to streamline your, your, your, your, your data that can really make a big difference. The effect on, of everyone doing it globally,

[00:19:19] of course, will, will dramatically slow down this, this massive need for extra infrastructure. I was thinking about that the other day. It's not like we don't want to replace infrastructure. We just want to give infrastructure that kind of boost. Yeah. Get more of that data center, get more out of those servers before you have to buy a new one, reduce your costs, reduce the efficient, increase your efficiency, give back your workers the time they need to do other jobs. And, and we all kind of work

[00:19:48] together as a community to, to kind of offset the, the growing ecological problems that could come about from creating so much energy usage around data storage. And if we look at everything from aerospace to AI training data and healthcare, you, you've mentioned some striking examples of compression X technologies that are in use, but what industries do you think are, are most urgently adopting technology

[00:20:15] like, like this? And what surprises you about how customers are using the product? Because very often you and your founder there, you will sit down, you create a solution for a particular problem. This is how you think it will be used, but then you'll have a customer or client that will think of a completely different use case that you've never even thought about. So how are they using this? Anything surprise you? Yeah, we were, we were approached by a company that was flying airplanes over energy

[00:20:42] infrastructure and taking LIDAR of the trees that were encroaching upon the, upon the pylons. So the heavy fines for having energy outages in, in, in the industry. So to mitigate that, if you can get the data off the plane as fast as possible and action it and get those, get, get to work and making sure that you, you work on those trees before one falls down, then there's a real gain. So it was very interesting

[00:21:11] because we had this huge sensor was creating a million records per second. And Stuart, my co-founder drove around Kent in his car, streaming the 1 million, the 1 million records per second on 3G to prove it was possible. So that, you know, that, that's something that, because having data accessible, having all data accessible almost instantaneously is a fantastic thought, but we have, we have problems.

[00:21:37] Everyone has that, the upload speed might not be high enough. They might not have service in certain areas. So there's a, and, and with, with all these new technologies coming along, creating so much data, autonomous vehicles, for example, smart cities, imagine being able to get that data, huge amounts of that data anywhere in the world very, very quickly. That's really, really exciting. And that's,

[00:22:03] and that's made us want to put, put in our roadmap a bit further down line, a streaming version. So we, we help, we are, our customers are helping us to build our roadmap going forward. The, the, the other, the other demand is for a software development kit, which, which we plan to do next, because integrating our technology into people stacks is such an exciting opportunity. They know that they know that their tech better than we do. We want to provide them with the, with the means to

[00:22:32] be able to get the best compression they can. It's very interesting. Actually, you can, you can, ramp up the compression to get more. We can dial it down and make it super fast. We can adapt our, our algorithm for each need in each part of each person's stack. That's very, very exciting because, like I say, partnering with, with people is what we really enjoy doing. And of course, as we look to the future, there are a lot of climate goals aimed at 2030,

[00:23:02] a date that's getting closer and closer. So if we were to fast forward, five years and look at your future vision here, what role do you hope compression X and smarter compression as a whole will, will broadly play in, in making digital infrastructure more sustainable? What, how do you see this panning out? What do you expect in the next five years? Well, I mean, already there are concerns, you know, there's over about a hundred new data centers

[00:23:28] plans right here in the UK. And also apparently 8% of all emissions will be coming from data centers by the year 2030. So that exponential rise, if we, if we, if we can position ourselves, if, if, if we can help people on a personal level, reduce the impact they have and on a company level, but also be with it, within people's stacks, we can, we can mitigate that, that, that, that, that tide, that avalanche of

[00:23:57] the new data creation that's coming and kind of one bite at a time help in any way we can to reduce that, that, that impact. So it's very exciting. It's a big job and many, many clever people are inventing many, many clever technologies to do that, different cooling systems for data centers and all this really cool tech is, is coming out. We want to be a part of that. We want to, we want to

[00:24:22] do our bit, um, to, to, to help, to help that as much as possible. And I think that is a powerful moment to end on now. We did meet at the IT press tour in Amsterdam recently. I believe you're at another event in London this week. So tell me where people can find you on, uh, on the road. Maybe it's the London event, maybe there's future events as well, and also website and where people can find

[00:24:49] out more information about everything we talked about and continue the conversation with you. Where should they go? Yeah, it was a great event in Amsterdam. It was great to see you there. Yes, we're going to be at the Olympia for the big data London convention conference, which is Wednesday, the 24th and Thursday, 25th of September. So just next week. And we're going to be presenting

[00:25:12] in the data for good lecture theater, two 40 or Wednesday, the 24th. And we have a stand G 58. So please come and see us if, if, if, if you're around. And like I say, our first step is to try to promote our downloadable desktop version of compression X, but we're also creating an SDK wait list for partners to join. And we've also got some fun things on the stand as well. We've

[00:25:42] created a fun, a fun compression game with a, with a prize draw. And we're looking forward to meeting lots of people in the industry and yeah, just having, having a good time sharing and collaborating with other people. Cause that's where the magic happens. You know, like when we sat in Amsterdam with all the other journalists, so many ideas were coming out and people are feeding off each other. So we hope to do more of that at this event coming up this week.

[00:26:10] Awesome. Well, I'll add links to everything, including where you will be at the event, your website, social channels, all that stuff. And again, anybody listening that's in the London area, big day at London, if you're dropping by, please go and say hello. I mean, senior presentations, et cetera. I do urge people to check that out. And today we've covered so much from the problem

[00:26:34] of data growth, the environmental impact of it, but also the solution here, how compression can be used as a sustainability tool. And especially with your adaptive algorithm and our personal responsibilities around digital footprints, not to mention the financial implications and increasing cloud costs out there. There are so many different reasons to take this technology seriously, but just thank you for shining a light on it, bringing it all to life today. Really

[00:27:01] appreciate your time. Yeah, thank you. I really appreciate it. Thank you. So what do we take from Nicholas's journey from a mathematical paradox in a leather bound journal to a working tool already cutting data footprints by more than 90%. And I think another message is that technology doesn't always need to be flashy to make a difference. Sometimes it's just about

[00:27:24] efficiency, responsibility, and asking better questions of the data that we insist on keeping. And if data is really growing faster than we can manage, perhaps the bigger question is, how will we choose to store it responsibly before the weight of our digital lives overwhelms the physical world? Food for thought indeed. Love to hear your thoughts on this one.

[00:27:50] TechTalksNetwork.com. You'll find all of our podcasts and many topics to explore and how you can leave me an audio message. If you want to go old school, techblogwriteroutlook.com and social media. I'm just at Neil C. Hughes, everything. So give me a shout there. But that is it for today. So huge thank you to Nick. Huge thank you to Philippe and Celine for inviting me to the IT press tour and

[00:28:14] introducing me to Nicholas. But more than anything, thank you to each and every one of you for tuning into this show. Bye for now.