Qentis Business Plan Doesn’t Add Up

qentis claims

There’s a claim by a Russian company called Qentis about creating ‘all images’ to try and monetise copyright by basically ‘creating all copyright’. It builds on an idea I helped explore at TorrentFreak back in 2011 on Quantum Computers. At the heart of both, is the idea of generating all content, mathematically.

The problem is, the maths just doesn’t work.

Qentis makes several claims on their site, which are about generating music, text and images. Let’s focus on the images, as it’s the easiest. The same things apply to music and text of course, but the mathematics are much easier for images.

They claim to be

  • Generating “all possible images in the 1000×800 pixel format”
  • Have done 3.23% of these possible images since 2007
  • Is generating at 1 million images/second.
  • Will have 10% done by the end of 2015

It may sound plausible at first, but anyone with a maths background (like me), or distributed computing projects (me again) immediately sees problems. There’s also additional problems legally, so let’s deal with them first.

Legal Issues

The biggest problem is that as a mechanical creation, in most places nothing they generate will be eligible for copyright, as it’s not a creative expression of a legal person. It’s the same reason the “Monkey Selfie” is ineligible. Brute force creation is by design, not creative. So, there’s a problem.

Of course, if they pick a jurisdiction without that requirement, they’re going to have another problem. If they’re going to create all works, then that means they’re going to also end up copying pretty much every image currently under copyright. Oops. Their plan is to sue anyone for copyright they that creates a work they’ve already done, so doesn’t that mean that anyone who’s got a registered picture can thus sue them for infringement in return?

The answer is, of course, no. There’s only a copyright infringement when the Right to Copy is violated, independent creation is not infringement. Which is lucky for Qentis, as they won’t be sued out of existence. That all assumes they are creating the images, which is questionable, leading to:

Mathematical Issues

The biggest thing that jumps out is just the scale of the task. 1000×800 images contain 800,000 pixels. Doesn’t sound a lot (especially if I write that as 0.8MegaPixels) but how many images can you create from that?

If we assume a straight monochrome (black and white only) image, that means there’s 800,000 pixels, each of which can be black or white. There’s an easy way to do the maths there, and it’s using powers. And by limiting to two colours, we’ve reduced things to a binary state, which is easy for most computer-literate people to deal with.

The maths is simple though and works out to 2800,000. Again, a number that won’t mean much to people, until you turn it from a binary number, to a decimal.

2800,000 = 9.920 x 10240823

If you aren’t sure what that means, that’s a 9 with 240,823 zeros at the end of it. to give you an idea how big, just writing the number out would give you a book bigger than Moby Dick.

It’s a stupendously huge number that it makes no sense. So, let’s take it smaller. Instead of 800,000 pixels, let’s take it down to just an 8×9 image- 72 pixels. And there’s a reason I’ve picked just 72.

The distributed.net RC5-72 project has been going on since 3rd December 2002, crunching RC5-72 keys. As the number indicates, the problem is 72 bits long, and so there’s 4,722,366,482,869,645,213,696 (or more simply 4 x 1021) keys available. A much smaller number than the number Qentis is trying to process, and they’ve been doing it for almost 12 years, so how far have they gotten?

Well, if you check their stats page, you’ll see they’re at 3.323% (at time of writing) complete. But cracking keys is slower than making pictures, right? Well, no. See Qentis is boasting about 1 Million Pictures per second. In 2011, distributed.net was doing roughly 300 BILLION keys per second; right now it’s average rate is 420 Billion, significantly faster than Qentis claims (420,000x faster in fact).

So, to compare

  • Qentis claims to have created 3% of all images of 800,000 pixels, at a million images a second, in 6 years.
  • Distributed.net has crunched 3.3% of keys just 72-bits long, at 420 billion keys a second, in 12years.

The maths just doesn’t add up for Qentis. Not even close. How long would it take for Qentis to create all those Black and white images, at a million per second? Well, the maths is east, just take off the last six zeros, and you’re done. That means 9.920 x 10240,817 seconds.

Problem.

The universe currently, is only 4.3 x 1017 seconds (1.4 billion years) old. So, to see how that compare with the time, just take off the last 17 digits.

It’s going to take 3.13 x 102,408,000 universe lengths to do their plan. I think 2020 is an optimistic goal.

And bear in mind, this is just for black and white only. If we take it to the early 80s and allow 8 colours, it becomes 8^800,000 or 9.7x 10722,471, and when you get into numbers that are three quarters of a million digits long, you’re way beyond reality.

And that doesn’t even address storage problems, because there’s only around 10^80 atoms in the universe. The Library of Congress would have to store 10100,000 (that’s 1 with a hundred thousand zeros after it) records in every single atom in the universe, to cope.

And imagine how long it’s going to take to process!

Either way, Qentis is talking absolute crap. Of course, their plan is not to actually follow these claims (as I’ve demonstrated, it’s impossible) but to get funding.

If you are interested in this business model and are an accredited investor, we invite you to speak with us about investment opportunities. We accept accredited investors who are willing to lock up at least $5 million for up to five years.

and

In 2005 Davies Guttmann became CEO of Qentis, 2006 Qentis issued a private equity placement which was bought up by institutions and private individuals. This funding has allowed Qentis to triple the size of its Research Department and is yielding exciting ladvances on a monthly basis. The management team around Davies Guttmann was joined by Michael Marcovici as CTO in 2011.

“ladvances” – that’s one way to describe pissing away people’s money. Regardless, Qentis is a scam, a joke, and a testament to people not able to do [fairly] simple maths and estimation.

Funnily enough though, Michael Marcovici is a name some may have heard of before. He used to run Europe’s biggest power-selling company on eBay, before it went bankrupt in 2005; the name of that company… Wiener Qentis Holding GmbH. Sound familiar?