Posted on : 20-10-2014 | By : Bozho | In : Opinions
I've had several announcements about pet-projects of mine that got to the front page of HackerNews and some technology-related subreddits. And although that doesn't mean that the projects were promising, game-changing, or investor-attracting, it at least means they are somewhat noteworthy in the technology world. My pet projects are not a startup, but nevertheless I pitch them to popular tech news sites, for the sake of the experiment. And, well, some could have very well been "startups", like the social network aggregator tool that I once made. So I read all the "tips and tricks" for pitching, I kept things concise and to the point, but still giving a good picture of what the product is and how it fits with the competition. With absolutely zero response. At some point several people (including here on HN) pointed out that I should not write to the generic, catch-all firstname.lastname@example.org, but I should target individual authors that seem to be interested in similar topics. And so I did, I started following authors on twitter, quoted relevant articles of theirs in the pitch; with no response whatsoever, again. Two concerns: 1. Why do they have a news@ / tips@ emails, if nobody reads that? 2. Are technology people, and especially startup founders at the mercy of tech journalists, trying to get personal connections to them in order to get something published? While my projects are not my startups, and I've not thrown money at them, I imagine the confusion that a startup may have, when it hits a brick wall when trying to get publicity. It seems to methat tech journalists (not all of them, of course) are not quite competent, but are also over-confident (just like any other journalist, actually). And they not only report "news" from the technology field (though many of the big tech sites are now tech tabloids; just look at the front page of Mashable), but they are also often judges of whether a startup is good or bad. And I question their competence in doing that, even though they have experience doing it. Or it may be just that a lot of crap gets poured into their inboxes and sometimes noteworthy entries are missed because of the volume. On the other hand, does it actually matter? If a piece of technology is really good, it probably won't need TechCrunch coverage to be successful. Maybe only the readers suffer from the inability of tech journalists to curate interesting content. But on the other hand - that's why we have hacker news.
We read. We watch. We learn. We gain knowledge. What's the point? The point of knowledge is to allow our imagination to make unexpected connections between more and more elements. In other words, we feed our serendipitous imagination. Then many things may come out of it - ideas, inventions, art. Anything, in fact. Einstein wouldn't have come up with the theory of relativity if he didn't know so much about physics. Shakespeare wouldn't be that great if he didn't know the rules and techniques of poetry. Great minds amass knowledge and then, out of nowhere, ideas appear. Why it works like that, I can't tell. But it does. If the amounts of knowledge is so important, then why computers can't come up with new ideas? Google and wikipedia have all the knowledge in the world. They even make it harder for humans to remember information that is easily verified, because, hey, you can look it up (a.k.a. The Google Effect). The Semantic Web is an attempt to make sense of the interconnected information on the web. But even with all these technologies in place, a computer wouldn't have been able to generate the string theory, the light bulb, Midsummer's Night Dream or Beethoven's 9th. Knowledge is one prerequisite. But you can't always have all the knowledge, therefore you may need someone else to steer you into the right direction, to complete the missing puzzle of both your knowledge and your pre-hatched ideas. That's what Steven Johnson is telling us in his RSA talk about ideas. Chance favors the connected mind, he says. I'd say that chance also favors the knowledgeable mind. But that's just another prerequisite. Then you need to go into a state of mind that accelerates the process. John Cleese calls it "the open mode" in his wonderful presentation about creativity. The process of generating ideas goes through obtaining knowledge, connecting with other people's ideas, and then going into the open mode to fuse it all into a new theory, invention or masterpiece. One tiny personal illustrative story here - I'm a computer programmer and I had some rudimentary knowledge in music. I have tried using mathematical functions to make computers compose music, and it failed. Then one day on a 8-hour train in the heat, having nothing meaningful to do, I got the idea a computer composer that would work. I made it, and that's the result of having some knowledge in both fields, having the ability to connect with the ideas and knowledge of others (through the Internet) and being in the open mode. But let's not fool ourselves - the ideas are rarely great. Not every chemist out there is Mendeleev and not every composer is Mozart. Sometimes ideas are small or only of local importance. Sometimes they are not about science or art, but about social processes, economics or personal relationships. Sometimes they feel important, but turn out not to be useful. But regardless of the field or the importance of the idea, it's there to fuel other ideas, to allow other inventions to be built ontop of a failed invention, to allow new poetic forms to be developed out of an unpleasant attempt for a poem. Even if your political ideas have failed so far, they may be tweaked after 10 years and work for a the well-being of more people. Even if my implementation of a computer composer fails, someone else will pick it up later and make it work. Just as Steven Johnson explains. But in order to go that far, we need knowledge. Not encyclopedic, computerized knowledge, but one that makes us understand the concepts behind different aspects of reality. So go and read. Books, articles, whatever. Listen. Teachers, presentations. So far only we are capable of handling the whole process, machines can't. Train your brain to perceive information from the world in order to allow your imagination to connect scattered pieces and generate new ideas. That's the point of knowledge.
When I first released Computoser (my service for generating music, without any human input), I didn't think of a business model - it was just something interesting to play with. And the lack of business value was one of the factors affecting the success of previous attempts at algorithmic composition (as I mentioned in my first article on the matter). But then I realized there can be a business model. Not one where I make millions, but one that can support the infrastructure and make the service progress - both in terms of the algorithm and in terms of the UI. The business model is called "stock music". Yes, it's an existing one, where a given piece of music is sold for a relatively low price and is then royalty-free - that is, you can play it anywhere and as many times as you like, and you don't owe a dime to anyone. Currently stock music composed and performed by real people who get paid. The music is then used in supermarkets, elevators, promotional videos, ads, customer service phone lines, online games, etc. Computer-generated (algorithmic) music can enter that market by lowering the cost quite significantly. There are no composers and performers to be paid, and the product is sometimes comparable. So, instead of paying 15 dollars for a piece of music for you online game, you can get it for 50 cents. The main question here is - is it good enough? Not yet. Even if the composition is nice, the performance is MIDI, which means it sounds a bit artificial. But in some cases and for some purposes that might be OK. For now I decided that the the price is 50 cents per piece (you can pay a minimum of 1 dollar, because the payment providers have a flat fee of at least 35 cents, and tiny transactions mostly go to the payment provider). Bitcoin payment is supported, without that limitation (there are no fees with bitcoin). Of course, in case you don't need the music for commercial purposes, it is licensed under Creative Commons, and you can freely use it. Will algorithmic music become good enough to eventually drive human composers out of the stock music market? Unlikely to happen soon. But it can certainly get some share. I'll do my best.
Posted on : 23-04-2013 | By : Bozho | In : Opinions
Comedy series are a great thing. But how do you get to choose a comedy series to watch? You can either listen to a friend's advice, or dedicate 2 hours for watching a couple of episodes, before you know you like it. Well, there's another option - that was the case of how I got to watch Family Guy, for example. Someone sends you a short sketch/joke on YouTube. Then you click through a couple more. That way you "test" the series by watching some preview jokes. It's like a perfume tester - a small dose that helps you decide you want to watch it. But there's a bonus - it can get viral-ish. You tweet a joke, someone retweets it, and immediately you get this to the attention of hundreds of people. Great, right? Well, it seems producers don't see it that way. I've seen a lot of these sketches removed from YouTube. I want to send a particular joke relevant to some real-life situation, and I can't find it anymore. It's deleted by the "owners". It's their content, they have the right to do so. But who, in their right mind, would reduce the virality of their own jokes? When will a perfume manufacturer forbid testers. Not only that - do you think a perfume tester would ban users from making their own testers and giving them to friends? Sounds ridiculous. Somehow, though, it sounds OK when it's online. And when the product is not something you just "buy", but something you watch, thus increasing its rating. The "war" on piracy now has a lot of clueless decision makers and that damages both content creators and consumers.
Posted on : 28-03-2013 | By : Bozho | In : Opinions
Yesterday I had to part with my company. I was a senior software engineer and team lead, but it turned out we are no longer a good match. I was opposing too much of management decisions on one hand, and many of these decisions were rather irrational at times. It appeared, though, that the person in charge is also reading my tweets. And although I don't usually tweet workplace-related stuff, I reported to my twitter followers about the last instance of an irrational management decision which our office (at a remote location) didn't want to follow (I'm keeping it a bit vague, because I wouldn't like to either expose internal information or defame the company - that's not the point of this post). The point is twitter (and other public resources, like this blog). I'm a relatively prominent twitter user in Bulgaria and I tweet a lot - around 30 tweets / day on average (you can tell I'm using social media a lot since I've built a tool to use all social networks simultaneously), and I tweet what is currently on my mind. Without any thought of censoring myself. It appears, though, that some bosses are too eager and use google translate (I'm tweeting in my native language) to get an idea of what employees are tweeting about, even if it that costs time (going through hundreds of tweets in a foreign language per week is an odd endeavor). So they found a work-related tweet, and a tweet about me having an interview with another company, and considered these a viable reason to not have me onboard. The feeling was mutual at that point, so leaving the company was the only logical and possible thing - and for the record, it was "on mutual agreement". The interview with the other company is a usual practice of mine - that way I can be up to date with what's being looked for in the market, but these details aren't visible out of the short tweet context. So, you can't actually rely on the tweets for adequate information about the person's attitude, even if you read them for some reason. (In fact, I've been advertising our open positions using my twitter account, but probably they didn't dig that far back in my timeline). So, is it OK to think twice before you tweet? Absolutely not. Apart from contractual (and logical) obligations not to reveal company secrets, you shouldn't have any other concern about your tweets. Because that's what freedom of speech is - not to worry that something bad might happen to you if you say something in public. Revealing company secrets is something you should not do even in an offline context, and that's a separate topic of discussion, but apart from that, expressing what you think, even if it is somehow related to workplace events, must be OK. Some companies have rules about not showing anything work-related on social media. First, that's a very broad definition. Is a tweet such as "damn, I'm late for work again" or "I really like debugging concurrent applications" breaking these rules? Even if we assume we should limit the freedom of people to express themselves online, it's impractical to do that with such rules. "But what if you say your company sucks and thus drive off some potential recruits". Well, that's part of the picture - there are tons of forums and platforms for expressing employee opinion about a company, and if you really think your company sucks (but for some reason haven't left it yet), there's no way to stop you from defaming it anonymously. My twitter, at least, is not anonymous - I use my real name and a photo. And I will continue disregarding the option that some boss out there might be looking for a reason not to like me. If a company reads your tweets "just in case", you probably shouldn't work there anyway. It's as if a company representative would always sit at the next table in a pub and write down your conversations. Sounds ridiculous, but somehow when things are brought to internet-scale, they stop sounding so ridiculous. The internet scale doesn't matter in principal - you are free to say what you think and what you do online. I'm not sure we can have regulations for that - e.g. "a company cannot fire you over a tweet", because it will officially fire you for another vague reason, or won't fire you at all, but make you leave. But, as society, we can make that practice "uncool". Not everyone has the hundreds of options to join a better company in no time, like I do, and watching out for the contents of tweets is something a lot of people are probably doing in order to keep their job. The practice is limiting expression without an actual benefit for companies or employees. Can we make it "uncool"? I will do my share - from now on, on interviews I'll ask about the company policy about stalking employees in social media and I'll state clearly that I don't intend to censor myself.
Posted on : 13-03-2013 | By : Bozho | In : Opinions
I just made my algorithmic music composer (Computoser) accept Bitcoin (only if you need tracks for commercial purposes, they are free otherwise). First, why Bitcoin?
- there are no fees, there is no business entity verification process - you just set it up and start accepting payments
- it is practically the only payment solution that works in my case. I need it to be international, to allow merchants from my country (Bulgaria, an EU member), not to have too high fees (the price is $0.99 for two tracks, so PayPal's $0.40+ fee is not an option), not to have a monthly fee (I don't think I'll sell anything in the beginning, and even if I do, the volumes will be too low). Stripe is a good option that meets all but one of the above criteria, but I hope it will land on the European market soon.
- computer-generated music is entirely "digital" - it doesn't require any human input or physical materials and so is Bitcoin. It made sense for a digital currency to be the means of paying for generated music.
I've been following social networks and social network news and analysis for a while (because of welshare), I've even written a couple of articles. But recently I stopped doing that. And not only because welshare became an "on hold" project. Social media is boring. Nothing interesting is happening there. Neither from a technical perspective, nor from an end-user perspective. What is the situation in three sentences? Facebook and twitter have tons of users, Google+ tries to catch up and LinkedIn have occupied their niche. Neither of these make any meaningful innovations and simply refine their UI. Apart from that, tons of new startups ("a social network for X") try to get off the ground, and either fail immediately, or appeal to a limited number of people and usually lack a sustainable business model. And yes, I know facebook introduced the graph search. Reality might prove me wrong, but I think it's a useless feature. I'm sure it has been great fun implementing and allowed engineers to dive into quite complex tasks, but the premise that you'd like to search like that is odd to me. The front-end of LinkedIn would probably be considered good back in 2006. Now sowing a big green "Your request was successful" when you click "Like" is just lame. Twitter is simple by definition and apart form a couple of bugs that stop you from tweeting and viewing replies quite too often, there's nothing interesting in their development. Google+ is still a mystery for me - nothing appears to happen there, and they continue claiming their user base increases. I've stopped using foursquare for the total lack of point, and few other application makes actual sense - I've tried Path, Pinterest, Instagram, and they are boring. Don't get me wrong - all of these applications may appeal to some people. But what are they? CRUD (created/retrieve/update/delete - a technical term for simple data input and retrieval software component). People think of some scenario and make a beautiful CRUD for it. If it's beautiful enough, lots of people that have nothing more interesting to do jump in. 2 years later there's still no revenue and the service is still pointless. But I'll not digress in the topic of the internet startup bubble. Basically, nothing happens in social media. Marketers may tell otherwise, because they still find it interesting to understand how facebook page administration works (and it doesn't work well - tons of bugs and seemingly incorrect statistics), but for most of the people social networks have become something like a street or a park - it's just there for you to use, you don't pay much attention to it and don't expect anything interesting to happen with it. By that I don't mean to devalue social networks. On the contrary - they have become a part of our lives. What I'm ranting against is the enormous interest that is still invested in them. Tech journalists, analysts, entrepreneurs, venture investors and business angels - all of these seem to have some odd obsession with "social media". I think no innovation can happen in that field. Let's focus our efforts elsewhere. Creating more interesting software, reviewing more innovative applications, investing in more thoughtful startups.
The top three social networks (even though Google+ is getting up to speed) are considered to be Facebook, Twitter and LinkedIn. Reddit is a huge site as well. And it won't be news to you if I say that there are a lot of bugs there. Including basic functionality:
- On twitter every third tweet ends up in "Oops, we did something wrong. Please try again".
- On facebook, if you are trying to post something on your page wall, it hangs. If you press "Post" multiple times, nothing happens again, but all the submissions are saved. So when you refresh the page you have to delete.
- On LinkedIn the same as above happens - if you try to post something, nothing happens. With the tiny exception that when you refresh and see the multiple posts you can't delete them, because the "delete" functionality is also not working.
- On reddit you can obviously get shadow-banned by mistake (I was) and nobody can/is willing to do anything about it. You lose your account due to a random factor.
Last week I released my latest project - computoser - an online service that generates music algorithmically. Feedback has been very positive and this certainly means I'll continue improving it. But what could that service actually become? First, as I noted in my technical description, the idea is not new at all. Many people, including scientists, have attempted to that the same thing - make the computer generate music. And some of these attempts (you can see some links in the discussions on HN and my blog) actually generate nice music, sometimes better than that of my algorithm. But why is it that this hasn't got any traction? Why there isn't an industry and a business model around these things?
- it's mainly research. Researches don't and don't have to think of business applications of their findings. Researchers are interested in the essence of the music composition and rarely in serving it to a large amount of people.
- the algorithms are not that good yet. Something can't get popular if it generates dissonant, boring or non-varying music. Mine also can still be classified as boring, but at least it's not dissonant and tracks may differ from each other significantly. Most of what I've seen is just recombination of a set of samples, which produces a limited set of results, or attempts to employ math principles, which I think is not necessary (mainly because existing, human-composed music doesn't seem to exhibit such patterns, apart from some low-level details, e.g. note pitch frequencies). Mine is far from perfect, but I've tried to address many deficiencies.
- existing software is not marketable - even if someone with a business plan took the software, he can't make it popular - the UI is in many cases horrible and unusable, and it expects a lot of human input. Something which is not actually needed.
- although it has been around, it is not an extremely popular idea. Not because it's bad, but because it requires a person or a team to have a good grasp on both computing and music theory.