Baidu Goes All In on AI
bloomberg.com
bloomberg.com
Baidu totally missed China's mobile Internet market compared with Tencent or Alibaba.
It's so called O2O (a term mostly used by Chinese) market share is burning money.
Baidu's dreadful bet is today's AI hype.
Baidu made the mistake of advertising itself as a search engine like Google, but failed to compete and catching up in 2000s when other competitors like Tencent was already taking up millions of active users. Yes, Baidu music exists, Baidu Bar (like Yahoo! Answers) also exists, but they are subpar.
Regarding to 360, basically back in 2000s, companies realized the best way to sell their product was (1) partnership with computer seller by bundling software just like McAcfee / Norton, and/or (2) provide toolbar so users can access content from the browsers. Tencents and others developed their own forked browsers. Then the era of virus/malware terror plague began as people learned either the software have backdoor (they all do actually), software doing weird things (e.g. system optimizer!), or people downloaded the software from an untrusted source because of piracy and poor Internet safety education (when people are pirating, they go to forums to look for software and if a thread is hot users would click on it and get convinced to try the software - little to their knowledge, most of them have backdoors).
360 was the most popular free anti-virus in China, and the company did a massive marketing campaign and everyone fell for it.
Be the Evil, serve the Evil
Should the wisdom of investing in ML depend on the probability of finding problems in that domain that can be solved by: regression, classification, clustering or prediction?
So, if Baidu has many problems that are likely (by whatever estimate) to be solved by those techniques, then this is wise.
The solutions would have to be valuable to Baidu (presumably, though not necessarily, by being valuable to the customers).