受動的/能動的メディアの違い

最近アメリカではtwitter経由のトラフィックがGoogleを超えたとか、もう検索の時代じゃないとか、そういうメディア記事やブログ記事を見かけて思うのは、そもそも比べること自体がおかしくない？ってこと。

twitterやfacebookでは確かにソーシャルストリームで共有されるurlから各サイトに飛ぶような使い方が一般化していて、そのトラフィックが増えてるのは間違い無い。滞在時間や訪問頻度も高い。けど、検索とは別の使われ方だし使う時間も全く異なる。

これはメディアへの接し方の違いであり、ソーシャルメディアの使われ方がテレビなどと同じ受動的だから。
どちらかというとweb創成期を支えた旧ポータルサイト、例えばヤフーのカテゴリサービスに近くて、何もしなくてもpushされる情報を流し見ている。無論、マスメディアと違ってソーシャルレコメンドされた情報なので、質や量は全く異なるが。

一方の検索サービスは待っていても何も教えてくれない。聞きたいことを伝えないといけない能動的メディアだ。その行為は「人に尋ねる」ことに近い。人類の叡智を蓄えた図書館に調べに行くのも同じで、機能がweb上の全文検索に変わっただけだ。

従ってテレビが登場したときに本がなくならなかったのと同じように、ソーシャルサービスと検索サービスは並存し、まだまだ成長する。一方でマスメディアやポータルサイトは可処分時間をまだまだ奪われるに違いない。

#初めて全文iPhoneで書いてみた。以外といけるもんだね。フリック入力サイコー！

人間とコンピュータの可能性

現在最高のバイオコンピューターであろうヒトの脳でも、まだ数%しか使われていないらしい。
だとしたら今「頭が良い」かどうかなんて、そもそも脳のポテンシャルから考えたら誤差の範疇でしかない。ナンセンス。

ちなみに今のペースでスーパーコンピューターが進化すると2025年に脳の演算が可能になるらしいが、これはあくまで演算能力の話であって脳の動き（=プログラム）を再現できる補償は全く無い。

プログラミングとは人間が脳と同等の動作を一部模写することであるが故に、そのものは脳を超えることはできない。（機械学習を除く。）模写されたことを忠実に、高速に処理するのがコンピューターの得意なところだが、脳が処理する膨大な条件分岐を再現し高速に処理するのは至難の業である。

スポーツでも学問でも人間自体がまだ進化していることを考えると、コンピューターの進化はまだまだ天文学的に伸びる可能性があるということだ。今Googleがどうとか、Appleがどうとか、そんなことはどうでも良くなる時代がかならず来る。50年後には今は存在しない企業がきっと世界を席巻しているはず。諦めるのは文字通り100年早い、ということ。

こんなに心跳ることはない。それを思うと人生で初めて長生きしたいと思った。

IT業界に興味を持つきっかけを与えてくれた亡き祖父に感謝。
そしてこの終わりの無いフロンティア開拓に全力でコミットし続けることにします。

"cloud" vs "P2P"

As I was talking with my friend this weekend, I've got a mind like "is cloud going to be the only way?". Though it's going to be an ideal world that every data exists on the other side, such as on the web, but it costs so many hardware and bursty network transition.

My concern is that despite Microsoft and Apple have big OS market share, they had started working on implementing cloud services. I think that's a Google's field. Since Google's mission is that "make everything searchable", they do have reason to do that. But for Microsoft and Apple, the way of using P2P technology would be the counterpart. They can integrate P2P into the file system on their OS, and let the user uses their application and documents from anywhere. Using P2P lets the system to be not only scalable but also reasonable. Hardware makers can keep their earnings from the users.

I understand there are several problems using P2P technology (such as security, fail over, etc..), I believe it will be the new way after burst of cloud bubble.

Google is marketing company rather than technolgy company

Though there are many news about the Google's A new approach to China, I realized that Google really is a marketing company.

That is because I think the actual aim of this activity is to open the internet in China to the foreign company. Not only Google but also Yahoo! and Microsoft are joining in to this clamor, so government of China should decide which to take. Get all the restricts off from censoring internet contents and let it be in the free competition, or bounce out all the foreign companies. But the internet works well with in the global network, so the latter means China would never win the game globally.

Google did this very well using mass media to gain support from all over the world.
It is still interesting to see how the government of China will respond to this issue.

is double byte character objective?

When I sent an english tweet to twitter, I found it difficult to express my feelings in 140 characters because it's too short. In japanese, I've rarely used all of the characters on twitter.

And my consideration is that this means double byte language like japanese is objective because they have much information in one character than sigle byte language. Chinese characters even have multiple meanings in it, it's just like an objective programming language, isn't it?

similarity between translation and web search

I came to a conclusion that the action of translation is very similar to the logic of web search.

When you try to translate, you have to change the words into intents first, and look for the same words in the different language to fit that intents.

As for the web search, you also start with analyzing the intents of the keyword requested, and search for the documents that might be match to that intents.

The difference within these two is that mostly we translate a sentence, not a word, but web search works with only few words yet. This is because of the difficulty in analyzing the whole sentence. So that might be the reason both translation and web search are currently not to be done well. To put it the other way around, some of the translation technology might be adopted to search service.

forecast direction of 2010's internet?

In 1990's after Windows95 had released, the internet became familiar and thousands of people started to publish their contents on the web.

At that time, the only path to access their contents was typing the url to the browser. So, you could only know the new web sites from other media such as magazines or words of mouth.

And then some people started to make their own link sites, just like a bookmarks in the browser so that other people can get to the appealing contents. Some of them gained popularity such as Yahoo! earned big scale of traffic from their directory service edited by contracted web surfers.

In the end of 1990's, web had fattened so much that human based sorting could not satisfy the needs of internet users anymore. No other robot crawled search service was suitable for practical use than Google. Google served full-text search of almost all of the web sites and ordered them by relevancy.

In 2000's, though Google-type search service was the major traffic hub of web, the new wave of internet innovation occurred mostly in social sector. SNS, social bookmark, mini blogging service made a new path to get the web contents, “feeds from your friends”. These are pushed information very similar to mass media, but the big difference is that social media is personalized as it is.

So, what about 2010's? The keywords would be personalization and semantic web.
Personalization is needed because web service need to know more about the users. Classical web service returns same results to all the people, but users are individual. There are several information on the SNS-type websites, so Google Friend Connect or Facebook Connect will let the web to be more personalized.
Semantic web is much more tough to be achieved. Web develop environment and the browsers should let the webmasters implement there websites more structured. HTML5 might be a key to do this.