Oliver Thylmann’s Thoughts

I am wondering if I am not slowly but surely getting into problems. I like trying out things and there is a lot of great stuff out there. Aggregators of your personal life are starting to come up but this actually poses a problem.

I recently made a change in how I blog, twitter, tumble, the like. Real thoughts go in here, little notes go into twitter, which is more of a conversation. All the stuff in between now actually ends up on my Tumblr Blog. But the Tumblr Blog actually aggregates del.icio.us tags tagged linkblog, as well as flickr posts with tag moblog and my kyte.tv posts as well as qik live streams. It kind of aggregates my short thoughts.

Above that I also installed Noserub on my own server, for a large part because Dirk works at Ormigo and because I think it is one of the most interesting implementations in the open social graph movement. It aggregates all my accounts into one view. I can’t really have it aggregate my Tumblr blog because that would double aggregate a lot of stuff. It is actually more likely that I will remove some stuff and only have it take the Tumblr blog. That is actually more what goes on in my life that might interest other people than all the other stuff.

This is also one of the reasons I kept twitter out of the aggregation via Tumblr, because it’s kind of a different level that people have to be able to choose distinctly to follow.

What we really need is something that will aggregate all my feeds, and somehow allows my friends to selectively subscribe to that stream of my ideas, somehow only giving them what might interest them, maybe based on what they read. This would actually be a very fun project. There you go, another start-up idea for somebody out there. Even better would be if somebody would write something that extends Noserub to that extend. I am sure Dirk would not mind.

With that off my chest, on to some more work

In 2006 we had Project Blackbox from Sun, who have now bought MySQL and are starting to have some really cool servers. Google then patented something like that, a data center in a shipping container. Now we have something new.

Ars Technica posted about a new start-up called International Data Security. They are launching something described in this PDF.

In short: they are buying old container ships and making them into hosting centers. First one goes online in April in San Francisco. They want to deploy 50 ships world-wide. The cool thing is that having so much water around, makes cooling easier, and having an entire ships gasoline tank, makes for 1 month of power outage. Of course the ships are set-up to handle containers, hence blackbox … . Above that there will be offices, food and quarters on the ships.

All sounds to cool indeed. Sadly they can’t move efficiently due to the stupid bandwidth that has to go in there… damn

Congratulations to OpenAds! I included OpenAds in two recent posts and now they are getting more money to really develop their vision further. With their announcement of a hosted service I think they might be further along to bringing real change to the advertising market in general, even though that might not be their initial focus. The market is big enough as it is. One thing they really bring to the plate is that they are a real alternative for big publishers as a banner server now. With a $15 million investment, they have the money behind them to explain to a big publisher that they are not going away and especially the big ones don’t want WPP, Microsoft or Google running all of their ads.

As Saul Klein puts it, OpenAds, the server, already serves billions of ad impressions (I’ll add “a day” to that quote) and with the hosted version, access to OpenAds will even get easier. I hope they do have a good idea on how they will handle the infrastructure, especially for statistics delivery.

I have been running OpenAds here on the blog from time to time, and have introduced it in different places where something more banner server like was needed, in relation to the stuff I have built in the last few years (being more AdSense like Systems, which add the complexity of ads within ads). OpenAds is really powerful and is getting better with each release, meaning that if you are running ads on your site, I’d look into it.

I started testing the German AdScale some time ago, and sadly they do not filter out non-German traffic. So if I run their system on my blog as default, 90% of the ads served make no sense whatsoever. That is where OpenAds is an easy solution and where a hosted version would be even better, because setting up OpenAds on my Server is actually a bit overkill for my little bit of geotargeting. At the moment AdScale is off again while I set up OpenAds (or get access to the hosted version due to this post ;)) and then I will probably benchmark it against our system at Ormigo (which we haven’t opened up yet, doing specific deals with specific publishers at the moment). I might try out Proximic, which surely has an interesting system (they are using the Vector Space Model for content match, possibly a learning system ala Support Vector Machines, which would be a lot nicer). The problem I see with Proximic, is that they announced having secured access to eBay’s Shopping.com and Yahoo! Shopping’s product listing, among others. That means that they have access to a lot of potential “ads”. The problem is that they don’t really pay. Google isn’t so good because it does content match, but because it has high prices. Sure, content match is nice, but YPN made more money for publishers when they optimized for revenue without content match as far as I know (this was in the private beta phase). Google does the content match for traffic quality. What CPC might you get for a book buy. A presume a few cents. I just searched for a few books on google and the only one that gives an ad is “The Innovator’s Dilemma” … for Clayton Christensen’s consulting company (he wrote the book ;)). So I would even presume that if 3 cents minimum bid would be possibly earned in these searches, somebody would bid on them. I do see a lot of power in their matching technology though, but I do know that there are other players out there that do content match (even matching the content to keywords you give them). All the power to a startup from Germany that has somehow gotten international attention though!

I know of a trial by one of the big affiliate system to get us out of a publisher deal and their performance was well below ours (previous company). This stuff is very complicated, with strategy (if it exists), options, brand and performance being just a few of the influencing factors on the publisher side.

Sorry that this turned out to be a bit of a rant … but the Ad market _is_ interesting. And bringing the local ad market really online is even more fun

Update: OpenAds running again on blog (via old server, problems with lighttpd here it seems, don’t have invocation codes or banner types selectable and preference settings don’t save for those). Hence AdScale enabled for a second time, now with more ads, and hopefully limited to Germany. If you are in the US and see AdScale Ads … tell me

I’ve been wanting to write about this for some time and now that Louis Gray asked, I really need to listen and post about it. First of all, at the moment I put my account private again. I am still torn whether I should keep it public or private but for now, keeping it private makes it just this little bit more free to say stuff. It’s probably just because I am chicken

Above that, I have to say that I only subscribe to @olivert posts on my mobile, nothing else. I would not want to get the gazillions of SMS from all the posts from friends, especially because there are some people that are syndicating their blog posts on there, which I personally think is a clear don’t. You can post your blog post if you think your friends should know, but don’t make it a default.

I am moving further away from answering the question though… ok. I learn stuff. It’s trivial things like two of my friends getting a MacBook Air, or being live with Jason Calacanis forgot his passport at the hotel on his trip back from Paris, or just now David Sifry wondering why all major travel sites have downward trending traffic … I learned about the Sun-MySQL deal first via Twitter; I learned about Qik.com and got my Seesmic invite from Loic. I just answered a friends question about how Google does the Geolocation on the iPod Touch update. It’s weird, but stuff happens on there and as it is a bit like a very very slow IRC channel, you can handle the load and be part of the communication.

It does take a little bit of getting used too, but with something like Twitteriffic, it gets really easy to post and react. And as you have to be short, it is not really something that takes up a lot of time. This post probably took more time now than all the twitter posts I will do today.

So yes Louis, I think you should sign up. It’s not a drag on your performance like IRC (I entered a channel with 150 people yesterday and wondered how I ever managed to keep up :)), so that worry should not be there. Just be selective with your friends. It’s like reading RSS feeds. You might have too many at one point and then you learn to scale down. Slowly you can’t live without it anymore.

Now that would be a killer MacWorld announcement wouldn’t it? Just imagine Apple owning Quicktime and Flash. Just imagine them owning Photoshop. Just imagine them being content creation and distribution. That’s Cringely’s idea in his latest article entitled: End Game: Why Apple will buy Adobe.

The thing is that Apple has $15 billion in cash (as Cringely pointed out). That is not enough for Adobe’s $22 billion valuation, but they could throw in some cash. Let’s say they need to pay a 20-30% premium, then they need something like $30 billion to get Adobe. A few billion in cash and the rest in stock. With Apple’s valuation at $152 billion that might be doable.

The thing is that I do not currently know what else they could do that would really rock our world. Sure an ultra-thin laptop with  touchscreen might be nice, but without some added things it will not change the world. IT will not lead to soo much more profit, manely because the market isn’t that huge. They might do a package system with a desktop and laptop and things. I am pretty sure that they will do something great. Buying Adobe would be one possible thing.