Press "Enter" to skip to content

Month: April 2019

AI in Retail

Kenshin Fujiwara Kenshin Fujiwara
April 26, 2019
3 Comments

This is my response to the latest TechCrunch article: Walmart unveils an AI-powered store of the future, now open to the public.

Every time I see this kind of AI-in-retail application, I somehow feel sad. It’s because of the introduction of retail AI is often designed to one particular goal, which is to encourage consumers to buy more foods and goods by analyzing the buying behavior and removing whatever obstacles between the seller and buyer.

To me, this is totally against a food crisis which we will face in the near future. Even worse, more food and good consumption lead to more logistics that will cause more gas consumption.  Electric trucks are coming, but it will take decades for every logistic company to adopt them.

In the ideal world, all foods and goods will be produced nearby a consumer neighborhood. Thanks to the innovation happening in plat factory and 3D printing, we are seeing some of them becoming reality. I’m a big fan of startups that are tackling a non-animal based protein problem.

AI should be used to foster the latter, not to drive the current consumption heavy economy. In the long run, the companies that use AI to create a sustainable economy will gain trust from consumers and eventually win.

What’s Wrong With Japanese AI Talent Education?

Kenshin Fujiwara Kenshin Fujiwara
April 26, 2019
2 Comments

The Japanese government finally announced its policy for AI talent education. To me, this is just another instant reaction without thinking too much that we’ve seen many times in the past.

In 1987, the government said we need to educate software engineers because there will be a shortage of 40,000 system engineers and programmers by 2000. In order to solve this problem, the government launched a plan to train 16,000 teachers who would be responsible for teaching programming at junior high school.

Did Japan become a top country in this domain? Nope. India, China, Vietnam, and the Philippines are doing a much better job right now.

In 2016, the government said we need to educate security experts who can prevent the country from getting cyber attacks because there will be a shortage of 200,000 security experts. Again, Israel and Chine are doing a much better job. Do you see the pattern here?

When the government says ‘educate’, it means they are trying to create more users, not inventors. Creating more AI users, particularly people who can use deep leaning, do not make Japan the leading country in AI. In fact, it’s forcing people to become consumers of AI, not producers. There is a huge gap between the two.

Look at Canada. Why does this country host so many scientists who contributed to fundamental research in AI? To name a few, Yoshua Bengio, Geoffrey Hinton, Yann LeCun, and Robert Tibshirani (a core contributor of LASSO which my startup uses it a lot), they are either born, lived, studied or worked in Canada.

If Japan were serious about being the leading country in AI, then all investments should be made toward creating more researchers in statistics, mathematics, machine learning, and AI fields. I strongly feel sorry for our children who are obligated to go through this ridiculous education policy driven by the government.

We must do something about it.

Narrowing Down To Two Slack Channels

Kenshin Fujiwara Kenshin Fujiwara
April 13, 2019
1 Comment

First off, I’m not exactly a big fan of Slack. I feel like I’m forced to find a tiny comment within a thousand lines of source code written by someone else. If you are a software developer, you know what I mean.

That feeling is coming from the fact that Slack is mainly designed for a software developer and it’s a very text-heavy product. Slack is an awesome communication tool when you use it properly in your organization. But at the same time, it can kill your time as CEO.

As a company gets bigger, you as a CEO get invited into so many Slack (or Facebook, WhatsApp, whatever messaging app) channels that you should not be part of. Even though I’m making my position crystal clear, I keep getting such an invitation from the people inside and outside the company all the time. When that happens, I simply share my above blog post and quit the channel with a little apology.

My ultimate goal is to narrow down my Slack channels into just two. They are not #general and #random channels, where you are a member by default. I’m talking about #whatceoisthinking and #troubleshooting channels.

The former is what the channel name says. It’s a place for the CEO to share his/her thoughts company-wide. The latter is the place for the employees where they can request a special assistance from the CEO in order to troubleshoot anything that’s preventing their job from getting it done.

Troubleshooting is the privilege of the CEO and not so many CEOs think that way. Let’s take a look at my typical day schedule. I will explain why.

Amongst the tasks listed above, everything except for the last item can be done by other people. You can delegate these tasks to your employees. However, there are types of troubles that can be solved by the CEO only and they range from fixing a relationship between employees to making an apology to a loyal customer.

You as CEO want to troubleshoot anything as early as possible because it gets really messy if you leave it for quite some time. In order to do so, you need a channel to watch out for any potential trouble within the company. This is the reason why you need a dedicated channel for it and you need to be open for feedback from the employees.

Hopefully, one day I can be the CEO who runs his company by dealing with these two channels only.

CEO Must Not Work More Than 8 Hours A Day

Kenshin Fujiwara Kenshin Fujiwara
April 11, 2019
1 Comment

This is my response to the latest episode from This Week in Startups podcast. The talk was done by Joel Spolsky, a founder of Stack Overflow and Trello. The latter was acquired by Atlassian for $425 million as we all know.

During this 50 minutes long conversation, I particularly liked the below statement.

“I rarely worked more than 8 hours a day in my entire career because I figured if I’m working more than 8 hours a day then I have failed to delegate something.”

A delegation, in other words, letting someone else do a job for you is one of the fundamental tasks that a CEO needs to do. If a CEO can leave the office before anyone, it is a good sign that the CEO is actually doing the right job.

Instead, the CEO should spend 100% of his time for hiring people smarter than him and making sure a company doesn’t run out of money. Telling a vision and defining a long-term goal is another thing only the CEO can do.

If you are a CEO working more than 8 hours a day or doing jobs other than ones stated above, then you are not doing the right job. You need to hire people to whom you can delegate a job. You need to finance in order to hire these people. You need to tell them a vision and a long-term goal so they don’t lose their way.