Month: July 2011

Science vs. Business

Since the end of the dark ages, and the emergence of modern capitalism, science has been connected to business, in one way or another.

During my academic life and later (when I moved to business), I saw the battle of those that would only do pure science (with government funding) and those that would mainly do business science (with private money). There were only few in between the two groups and most of them argued that it was possible to use private money to promote and develop science.

For years I believed that it was possible, and in my book, the title of this post wouldn’t make sense. But as I dove into the business side, every step closer to business research than before, I realised that there is no such thing as business science. It is such a fundamental aspect of capitalism, profit, that make it so.

Copy cats

Good mathematicians copy, best mathematicians steal. The three biggest revolutions in computing during the last three decades were the PC, the Open Source and Apple.

The PC revolution was started by IBM (with open platforms and standard components) but it was really driven by Bill Gates and Microsoft, and that’s what generated most of his fortune. However, it was a great business idea, not a great scientific one, as Bill Gates copied from a company (the size of a government), such as IBM. His business model’s return on investment was instantaneous and gigantic.

Apple, on the other hand, never made much money (not as much as IBM or Microsoft) until recently with the iPhone and iPad. That is, I believe, because Steve Jobs copied from a visionary, Douglas Engelbart, rather than a business model. His return on investment took decades and he took one step at a time.

However, even copying from a true scientist, he had to have a business model. It was impossible for him to open the platform (as MS did), because that was where all the value was located. Apple’s graphical interface (with the first Macs), the mouse etc (all blatantly copied from Engelbart). They couldn’t control the quality of the software for their platform (they still can’t today on AppStore) and they opted for doing everything themselves. That was the business model getting in the way of a true revolution.

Until today, Apple tries to do the coolest system on the planet, only to fall short because of the business model. The draconian methods Microsoft took on competitors, Apple takes on the customers. Honestly, I don’t know what’s worse.

On the other hand, Open Source was born as the real business-free deal. But its success has nothing to do with science, nor with the business-freeness. Most companies that profit with open source, do so by exploiting the benefits and putting little back. There isn’t any other way to turn open source into profit, since profit is basically to gain more than what you spend.

This is not all bad. Most successful Open source systems (such as Apache, MySQL, Hadoop, GCC, LLVM, etc) are so because big companies (like Intel, Apple, Yahoo) put a lot of effort into it. Managing the private changes is a big pain, especially if more than one company is a major contributor, but it’s more profitable than putting everything into the open. Getting the balance right is what boosts, or breaks, those companies.

Physics

The same rules also apply to other sciences, like physics. The United States are governed by big companies (oil, weapons, pharma, media) and not by its own government (which is only a puppet for the big companies). There, science is mostly applied to those fields.

Nuclear physics was only developed at such a fast pace because of the bomb. Laser, nuclear fusion, carbon nanotubes are mostly done with military funding, or via the government, for military purposes. Computer science (both hardware and software) are mainly done on the big companies and with a business background, so again not real science.

Only the EU, a less business oriented government (but still, not that much less), could spend a gigantic amount of money on the LHC at CERN to search for a mere boson. I still don’t understand what’s the commercial applicability of finding the Higgs boson and why the EU has agreed to spend such money on it. I’m not yet ready to accept that it was all in the name of science…

But while physics has clear military and power-related objectives, computing, or rather, social computing, has little to no impact. Radar technologies, heavy-load simulations, and prediction networks receive a strong budget from governments (especially US, Russia), while other topics such as how to make the world a better place with technology, has little or no space is either business or government sponsored research.

That is why, in my humble opinion, technology has yet to flourish. Computers today create more problems than they solve. Operating systems make our life harder than they should, office tools are not intuitive enough for every one to use, compilers always fall short of doing a great job, the human interface is still dominated by the mouse, invented by Engelbart himself in the 60’s.

Not to mention the rampant race to keep Moore’s law (in both cycles and profit) at the cost of everything else, most notably the environment. Chip companies want to sell more and more, obsolete last year’s chip and send it to the land fills, as there is no efficient recycling technology yet for chips and circuits.

Unsolved questions of the last century

Like Fermat’s theorems, computer scientists had loads of ideas last century, at the dawn of computing era, that are still unsolved. Problems that everybody tries to solve the wrong way, as if they were going to make that person famous, or rich. The most important problems, as I see, are:

  • Computer-human interaction: How to develop an efficient interface between humans and computers as to remove all barriers on communication and ease the development of effective systems
  • Artificial Intelligence: As in real intelligence, not mimicking animal behaviour, not solving subset of problems. Solutions that are based on emergent behaviour, probabilistic networks and automatons.
  • Parallel Computation: Natural brains are parallel in nature, yet, computers are serial. Even parallel computers nowadays (multi-core) are only parallel to a point, where they go back on being serial. Serial barriers must be broken, we need to scratch the theory so far and think again. We need to ask ourselves: “what happens when I’m at the speed of light and I look into the mirror?“.
  • Environmentally friendly computing: Most components on chips and boards are not recyclable, and yet, they’re replaced every year. Does the hardware really need to be more advanced, or the software is being dumber and dumber, driving the hardware complexity up? Can we use the same hardware with smarter software? Is the hardware smart enough to last a decade? Was it really meant to last that long?

All those questions are, in a nutshell, in a scientific nature. If you take the business approach, you’ll end up with a simple answer to all of them: it’s not worth the trouble. It is impossible, at short and medium term, to profit from any of those routes. Some of them won’t generate profit even in the long term.

That’s why there is no advance in that area. Scientists that study such topics are alone and most of the time trying to make money out of it (thus, going the wrong way and not hitting the bull’s eye). One of the gurus in AI at the University of Cambridge is a physicist, and his company does anything new in AI, but exploits the little effort on old school data-mining to generate profit.

They do generate profit, of course, but does it help to develop the field of computer science? Does it help tailor technology to better ourselves? To make the world a better place? I think not.

Google+ and the Yahoo-isation of Google

Almost a decade ago I joined Yahoo to work on the search team. At that time, Google was giving Yahoo a hard time with their amazing search, while Yahoo was mostly based on directory search and some ad-hoc buyouts (AltaVista et al). Yahoo came with its own search, then bought Inktomy, then re-wrote the search engine, and they’re now using Bing. They were late on the search business. Too late. Not that Inktomy was bad, but it wasn’t better than Google and certainly wasn’t a novelty.

When Google came with Gmail it was a shock to all of us, Yahoo workers. How can they offer 1GB free mail and we only offer 10MB? How can a (then) small company provide such massive storage while the behemoth of the Internet could only afford peanuts? It’s all in the administration. Yahoo, for some reason I still don’t understand, had to have all user’s emails on filers (very expensive storage), and had to reserve the whole storage, even if less than 10% of the users actually used more than 50%. There was a lot of very expensive idle disks at Yahoo Mail…

Another case was social networks. Yahoo was never able to write a single decent social network that wouldn’t close an year later. Several internal attempts were made (during years of development) before the first Yahoo-360 came out, only to die a few months later of starvation.

Yahoo-style

When the internet was young, and Yahoo was at the top, they could do anything, people would just love. Yahoo mail was free and came with a calendar and some bits. It was horrid, but it was free, and we all used one day (especially after Hotmail was acquired by Microsoft). Somehow, the directors of Yahoo decided it was better to have it all, even if the quality was unbelievably low. They were so famous and so ubiquitous that anyone wanted to advertise on their websites. The more they had, the more people wanted.

That cycle made Yahoo create a huge number of useless pages and verticals (content website like weather, mobile, etc), just because people would pay loads of money to advertise there. I’ve seen many pages going live without proper review or a decent market analysis, and some were still with bogus content (non non non) and broken links after years. Yahoo had spread their butter so thin that it was impossible for them to compete with any other company.

Google came, Google destroyed Yahoo, Microsoft came and bought it. It may not be in paper, but Yahoo is the new Microsoft garden, where they put their feet up when they’re tired of working.

Facebook

With Facebook, the story is not completely similar. Google is still not where Yahoo was 8 years ago and the Internet is not as naive as it used to be. It’s true that Facebook killed every other social network website, but it’s not true that they’ll be able to do with GMail, what Google did with Yahoo Mail. Part because GMail is really good, and part because Facebook guys are not that good.

But there’s one trend that is happening to Google that is similar to Yahoo of the past: legacy. Google has a decade of code, normally more bad than good, and the new systems have to integrate with it. But that’s not the worse, by far. As Yahoo, Google started from ground-up, so they always had the start-up mentality. When the company wasn’t a start up any more, they still tried to do things the same way.

Two things happened during the last few years that put Google in a bad situation: first, because in the golden days they had truly done remarkable systems (simple, yet efficient), they thought (and carried on thinking) not only that they were the best of the best, but that they could do anything and anything they did was automatically better than any one else. Second, the lack of process and reality check only made things worse, by kludging infrastructure on top of infrastructure, by solving every problem as if it was map-reduce and by doing everything in-house, it was difficult to use off-the-shelf applications for things that weren’t really that relevant, and move on when it was, indeed, relevant.

Facebook is still a young company, but I hardly believe their fate will not be the same. It was also clear from the beginning (8 years ago), that Google would have the same fate. However, I don’t think that Facebook will overrun Google as the latter did with Yahoo, but that’s neither fault. The market is not the same, the Internet is not the same.

Google+

In the same style as Yahoo-360, Google is trying to use their own user-base to compete with Facebook, and for that, it’s very likely that they’ll fail. Not as bad as 360, but they won’t kill Facebook.

Hangouts, Circles, Huddle are just different names for the same functionality in Facebook, Twitter, etc. Even the layout of Google+ is identical to Facebook. 360 was exactly the same thing and that’s where I draw the line. That’s where Google is getting ludicrously similar to Yahoo 8 years ago and that’s why they’ll start failing more and more often from now on.

They stopped being creative. Their creative innovations (wave, buzz) is average at best, their copied products (chrome, android) and similar to the competition with no clear game changer and the old stuff (search, gmail) is still the same. Not bad, but not creative any more.

From now on, is only downhill. In a decade or so, Apple will offer to buy them, and fail to, but enough to dismiss the general trust people have on them and that will be the end. Zombies of the Internet…