Press "Enter" to skip to content

Category: Startup

Power of GPU: 20 Years Ago and Now

When I got my first job at Sony, I was assigned to work on the graphics library development of PlayStation2 gaming system.

My job at that time consisted of writing code with assembly-like programming language called microcode, packing that code together with polygon and texture data, sending them to Vector Units (aka GPU) via DMA, and letting GPU do the rest of work while minimizing involvement of CPU.

I had to adopt every technical tweak in order to get the best performance and achieve the highest frame rate out of that gaming machine and GPU. That’s simply because the library was intended to be used by so many developers that make blockbuster game titles and they relied on it.

Time passes and now GPU is used for more generic purposes including computation of neural network. If I were a software engineer who just graduated from university today, I would jump straight into the world of neural-network-on-chip bandwagon without thinking much.

I didn’t realize that my knowledge and experience of GPU programming had to do anything with AI back then. I believe NVIDIA didn’t realize it neither.

To me, it’s quite interesting to see young software engineers learning how to write code against GPU without CUDA these days, especially ones trying to use Raspberry Pi as a deep learning accelerator. Special thanks to Broadcom for making its VideoCore specification public.

At the same time, it’s sad to see that Japan is a bit behind of this movement. The country used to host many GPU engineers. We had Sony, Sega and Nintendo. Every gaming system had such a sophisticated graphic library that pulled nearly 100% performance from its GPU.

I cannot stop wondering what if the engineers behind these gaming machines were given a chance to work on today’s neural network chip development. Maybe Japan could be in different position in AI industry by now.

Deciding When to Change a Reporting Line at Growing Startup

On March 1st, we welcomed a couple of new employees to the company, which make 38 people in total. We grew to this number just in 2 years.

In my experience, one of the first things you as CEO need to do before a number of employees hits 30 is to change reporting lines in your organization. Thanks to Slack, Hipchat and other modern communication tools. It is now possible for a single person to handle so many 1-on-1 conversations with people in an organization today.

However, you will run into a serious trouble later on if you continue to let employees report to you directly. For employees it’s always a preferred way of communication because they don’t have to deal with intermediate people in organization and they can get a straight answer from CEO whether the answer is yes or no.

A problem gets much worse once that habit becomes a part of company culture. I’ve seen few CEOs who struggled with this communication overload problem because their startup stayed in 10-20 people stage for too long, which resulted in everyone at the company getting so used to report to CEO directly. It will require great effort to change the reporting lines in such organization since it’s equivalent to changing a company culture.

So here is a list of action items I would recommended before that happens.

1. Remove yourself from unnecessary Slack channels

There is a tendency for employees to bring you into every new Slack channel they create. Remove yourself immediately if you find it irrelevant from the perspective of company management.

Product update or project specific channels are the usual suspects. Do not feel guilty about removing yourself from these newly created channels. Do it fast before they assume you are in the channel.

2. Stop attending product/service update meeting

If you are involved with every product/service update meeting, stop attending it. Instead, get updates from CTO, VP of Product or a person who is in charge of product/service. Again, employees tend to bring you into these meetings without thinking much about the value of time given to CEO. Say no when you get such an invitation from Google calendar.

In my opinion, there should be clear distinction between management and execution even in startup at early stage. You as CEO and other C-level people are responsible for management. Employees are responsible for execution. Product/service updates are considered as execution. Attend these meetings if and only if they are absolutely necessary.

3. Narrow down your reporting lines

The ideal scenario is that you get product/service update from CTO, sales update from CFO, and everything else from COO. These 3 reporting lines should be enough for you. CEO has additional reporting lines that no one else at the company has. These are the reporting lines to stakeholders, which include investors and key partner companies.

The information goes the other way around through these additional reporting lines. It’s CEO giving updates to the stakeholders, and it requires careful and thoughtful preparation by compiling the updates you get from CTO, CFO and COO.

4. Be clear about your communication policy and state it company-wide

Once you decide to change your reporting lines, then share your communication policy with all employees at the company. This helps them understand why you are doing it. Again, the time given to CEO is one of the most valuable assets within the company and everyone should understand it.

Successful startups are organized in such a way where CEO can maximize his/her time for the things that only CEO can handle. Startups grow slower or sometimes fail because CEO is always busy with the things that someone else can handle.

Hope above list helps a startup CEO do the right thing. I know my suggestions are a bit extreme to some extent, but running a company in resource constrained environment requires a high-level discipline. Without it, a company cannot grow beyond 30, 50, and 100 people. It’s up to CEO to make the call as soon as he/she can.

Side Business Must be Encouraged at Startup

This Nikkei article caught my eyes this morning. PARCO is one of the largest department stores in Japan, and the company decided to allow employees to do side business as long as it helps them enhance their professional career and specialty. Writing a book and teaching at university are kind of activities considered as proper side businesses according to the article.

I’m particularly a big fan of employees who teach at university, and I’m encouraging our staff to do so. At our company, extra money that they make by teaching goes straight to the employee’s bank account. The company does not take any cut even though the activity is done during work hour.

By teaching, you learn a lot. In my opinion, it’s always a teacher who gets the most benefit rather than students. I’ve learnt so much by teaching in the past so this should be encouraged at all startups.

Which city should Japanese startup have a presence in Germany: Dusseldorf, Munich, Berlin?

I’m writing this blog post on my last day in Berlin. I spent almost a week in Munich before coming here. A purpose of this article is to help other entrepreneurs, especially those who are based in Japan, identify the most suitable location in Germany as a first footprint into this country.

Now, I understand my view is pretty much biased and the answer may vary a lot depending on the nature of business. Having said that, I will summarize both pros and cons of each city as below.


Dusseldorf has been a preferred city in Germany to established Japanese companies for decades. It’s no wonder that one of the JETRO’s two offices is located here. According to the latest statistics, 1% of the entire city population is Japanese.

The city’s advantage is that there is a relatively large number of German employees who are currently being employed or used to be employed by Japanese company, and they do understand how Japanese companies work to some extent.

But here is the catch. It does not mean these Japanese friendly workers are startup friendly. In fact, it can be quite opposite meaning they might not have necessary mindset required by growing startup. They respect a protocol order. At the same time, they may lack flexibility and agility. Wearing multiple hats is something they do not like.

Also important to note is that a talent pool of IT/software workforce is somewhat limited. It’s true not only to Japanese companies but also to local companies based in Dusseldorf. Yes, there are a couple of R&D centers owned by foreign companies and they have huge presence . But once again, the people may not be exactly startup ready.

There is a direct flight between Dusseldorf and Japan. Another reason for established Japanese companies to choose this city as their base in Germany.


German people used to say “There are industries in Munich. There is none in Berlin.” What they are referring to is, of course, automotive industry like BMW and manufacturing industry like Siemens in Munich whereas Berlin is known as the place for consumer-facing services and tourism.

I was lucky to stop by and talk to the representatives at German Entrepreneurship, a Munich based organization that fosters a local startup ecosystem, and XPRENEURS, a local incubator operated under the greater UnternehmerTUM umbrella. Thanks to JETRO Dusseldorf office for coordinating these meetings.

From what I understand, Munich is definitely a suitable place for those who intend to provide product and service for automotive/manufacturing industry. Needless to say, there are OEMs, Tier 1 and Tier 2 suppliers. As the industry moves toward service oriented such as MaaS and becomes more software centric, the opportunity for IT related startups is also becoming significant.

As for local startup ecosystem, Flixbus is a well recognized unicorn disrupting a long distance mobility industry. There is also Lilium, a startup founded by the engineers from the Technical University of Munich. For IoT/AI startup, Konux comes to my mind. Given the existence of these high-tech startups, Munich is no more about old fashioned industry.

Last thing to notice is that there will be direct flight between Munich and Osaka, in addition to the existing flight from/to Tokyo. This indicates an increasing demand from either side of countries. One downside might be that this place might not be your 1st choice if you are B2C or not related with above industries.


Everyone thinks Berlin is the place for IT talents and creators/artists. But after hearing from the people who know real insights, I learnt that a current talent pool is made up of foreign workers who migrated themselves to the city. What’s even more interesting is that Berlin based startups usually consist of lots of remote workers where a majority is often the people from Ukraine.

This fact leads to a little heads-up to Japanese startups thinking of Berlin as the place for hiring software engineers. Yes, there are such talents, but it may not be as large as you imagined. I also learnt that the people in Berlin excel at making UX/UI, directing a project and communicating in multiple languages. I think that’s partly due to the culture and diversity given by the city.

However, when it comes to coding and hands-on software development, you need to be aware of the above facts. If you are really keen on finding software developers with decent skill sets, you might want to look somewhere else like Ukraine and SE Asia. If there is absolute reason to work with German developers, then Berlin might be your choice.

Another point is that the talents in Berlin are slightly cheaper than those in other cities. That’s simply because a cost of living is cheaper here. People said that housing is becoming an issue as many foreign workers are moving into the city. Brexit is unfortunately one of the reasons.

A downside for Japanese companies is there is no direct flight between Berlin and Japan. The construction of new airport is in a middle of political debates which will take few more years to settle.


Photo: Taken at Viktualienmarkt in Munich

CEO needs to be good at everything and must enjoy it

Being a CEO for the past 5+ years, one thing I enjoy the most is the fact that CEO needs to be good at everything ranging from product development, marketing, legal to sales, hiring and even firing.

Below is my work schedule for a typical weekday.

9:00 – 10:00 Read online news and share a couple of important ones on Slack
10:00 – 11:00 Get updates from technical team about ongoing projects and give feedback
11:00 – 12:00 Write follow-up emails to prospective customers
12:00 – 13:00 Lunch with teammates
13:00 – 14:00 Face-to-face meeting with a prospective customer
14:00 – 15:00 Review legal documents and important contracts
15:00 – 16:00 Face-to-face meeting with another prospective customer
16:00 – 17:00 Work on proposal materials for new customers and partners
17:00 – 18:00 Skype interview with candidates
18:00 – 20:00 Dinner with candidates or customers (rarely)

As you can see, a startup at post Series A still needs CEO to work as salesperson and I believe this is the way it should be. I spend much of my time talking to potential customers and partners. When I talk to them, I’m considering myself as a top salesman and being better than anyone at the company in terms of sales. But it doesn’t mean I can suck at everything else.

When I talk to CFO, I tend to use a lot of financial jargon just because it’s convenient. When I talk to CTO, then I leverage my full knowledge of being a former CTO myself in the past. It’s not enough for CEO to understand what a counterpart is saying. CEO needs to excel an opponent’s knowledge so that CEO can give a proper feedback and sometimes point out mistakes if necessary.

When I first stepped into the world of venture finance, Venture Deals: Be Smarter Than Your Lawyer and Venture Capitalist was my bible. I’ve read it countless times not because it was co-written by my favorite capitalist of all time Brad Feld, but it was my go-to destination whenever I countered something I didn’t know about in venture finance.

Same story goes to sales, HR, marketing, leadership, etc. I have at least one or two books I always refer to or a couple of mentors I can talk to whenever I face my lack of knowledge. I believe an excellent CEO is an excellent learner who has willingness to become master at everything.

Is small IPO a bad thing?

Japan is known as the easiest stock market for startups to go public. A startup less than $10M revenue can do IPO. I’ve been hearing a lot of criticisms saying Japanese startups are going for public too soon and they are all aiming for small IPO. This is why Japan does not have unicorns as much as U.S.

The question is: “Is small IPO a bad thing?” The answer is: “I don’t think so.”

In Japan, there has been almost 400 companies that went public in the last 5 years. The majority of them were not unicorns and the mean value for company valuation is somewhere between $150M and $200M at the time they did IPO. But there are a countless number of companies whose valuation went up to $1B after IPO.

In my opinion, IPO should be considered as another financing round to startups. Japanese startups treat IPO as something following the previous financing round whereas American startups treat it a bit differently. It’s just a difference between raising money from VCs and doing the same from ordinary people.

This explains why a secondary market like SecondMarket and SharesPost exists in United States. A secondary market is the only way for American startups to provide liquidity for their employees and other stockholders before IPO since a bar for going public is way too high as compared to Japan. On the other hand, in Japan we simply do small IPO, achieve the same liquidity and then become unicorn using the money raised from public market. This isn’t exactly a bad thing for founders, employees and VCs wanting some liquidity early on.

We as Japanese startups need to take this advantage and should not be biased by the thoughts brought from the market where definition of IPO is completely different.

Welcome to 2019

Happy new year, everyone. I’m writing this blog post in my hometown (as you see it in the above picture) in far north Shiga. This year there is less snow than previous years so it looks like I don’t have to help my dad and younger brother do a thaw work, which I used to do a lot every year in this season. Thanks to global warming, maybe.

For the past few years, I always started my first day of January by reading blog posts written by the people I follow. Most of them are venture capitalists, some are thinkers, and others are just my friends. But this year I’m doing something different. I’m writing one myself.

From my frame of reference, I’d like to make some predictions for 2019 at both macro and micro levels.

China: The country faced (and still is facing) a political difficulty raised around an arrest of the chief financial officer in Canada. Instead of giving a backslash to other countries, my guess is that there will be a lot of M&A initiated by Chinese companies in 2019. Some will be direct against US companies while others will be indirect. Anyhow, China will try to get into foreign markets by leveraging the huge amount of cash reserved by domestic companies and spending it for acquisition. All startups CEOs need to be aware of that.

Japan: Unlike China, the domestic economy will keep shrinking and big companies will also keep their cash as retained earnings rather than spending it for international M&A or acquiring technology asset from startups. To me, it’s nonsense to create SaaS or any business that’s specific to this country. We’ve seen many ‘domestic’ startups raising multi-million dollar investment from VCs and CVCs in 2018, but theoretically it’s just a matter of time for them to go international since there is only limited market here. Startups without proper internationalization and global mindset in place will hit a hard wall.

Concentration: I think we will see more concentration of assets globally. Successful businesses aka GAFA & unicorns will attract more investment and appreciate an influx of talents while others do not. We’ve been in same situation before, but this trend will accelerate even more. On top of that, we need to be aware of concentration of information (or data) as well since it will be the most important asset apart from cash and human resource for next 10 years.

AI: 2018 was definitely the year we recognized that deep learning can do something meaningful to our society, but at the same time we learnt it has certain limits. In 2019, we will see a big edge-AI race between China and United States. I don’t think there is any room left for Japanese companies to jump onto this bandwagon. It’s just too late. However, this is technically “inference on edge side” race so there is still a room for learning on edge side, that is often considered as a true AI.

Venture Funding: As every entrepreneur I know is talking about it already, there may be some crash in venture funding in 2019. I din’t know what will trigger it exactly. It might be so called unicorns with too high valuation facing difficulty for going public. It might be startups raising too much money and missing key milestones that companies promised to their investors. Again, we’ve been in same situation before, but this time an expectation vs. reality gap is huge since more and more venture money was poured into the startups. Thus, all startup CEOs should stick with this founder’s principle: “Raise as much as money while you can.”

I’m feeling a bit sorry to see our country will face yet another economic downtime. But as startup CEO I need to overcome that situation, which will force me to spend more time outside Japan and look for potential deals and markets internationally in 2019.

As I wrote in the previous post, I decided to write all future posts in English only. This decision was made partially due to my above prediction. I need to globalize myself for sure.

Difference between CTO and VP of Engineering

This is a consecutive post to Difference Between Co-founder and Founding Member. In fact, the previous post is one of the most read articles on this website indicating many people are constantly looking for the answer to it, which lead me to write this post because I have my own definition between two.

First, I’d like to make my version of concise definition as below.

“CTO is a person who has the most technical knowledge in management team. VP of Engineering is a person who manages engineering team.”

Now, let’s break them into smaller pieces. Amongst many differences between these two types of people, there is a clear distinction in terms of which team in the company he/she belongs to. CTO is considered as a member of management team or often board members. Thus, CTO belongs to the same team to where other C-level people belong including CEO.

On the other hand, VP of Engineering is considered as a member or head of engineering team. Therefore, VP of Engineering belongs to the same team to where other employees belong. Perhaps, there are few startups or companies that appoint VP of Engineering to be a board member.

To make it even clearer, the biggest difference is what these two types of people represent. CTO represents a management team and speaks to employees. VP of Engineering represents an engineering team and speaks to the management team. A direction of communication is completely opposite each other.

More importantly, this is my own perception by the way, CTO was chosen just because he/she happens to be the most technical person in other C-level people. Due to this nature, a language spoken in two parties is also different.

In management team, the official language is business. Everyone in management team talks from business perspective since other C-level people most likely may not understand technical terminologies so that CTO should not use these technical terminologies in board meeting. In engineering team, the official language is of course technical. They are allowed to use any technical terminologies in order to proceed their projects.

A role of CTO is to come up with the means to implement a business decision made by the management team based on outcome from engineering team. A role of VP of Engineering to come up with and experiment the technical solutions toward a problem given by CTO and provide the best possible solution to CTO with reasonable explanation.

Sometimes I happen to encounter a startup lead by young CEO where their CTO is constantly talking about technical details in board meeting. If it’s absolutely necessarily, then it’s okay. But most often, it’s just CTO doing a job of VP of Engineering.

If CTO continues this behavior, then VP of Engineering will start doing a job which people at one level lower are supposed to do. Eventually, the startup will end with paying salary to a whole engineering team where everyone in engineering team is doing a job on which people at one level lower should be working. This is the situation to be avoided, particularly for startup with cash constraint.

Remember, the official language CTO is supposed to use in board meeting is business, not technical. Otherwise, he/she won’t be able to communicate with CEO/CFO/COO/CMO and other C-level people fully. In other words, if VP of Engineering wants to step up to CTO, then he/she must be equipped with some business knowledge at least.

American startup disrupting Japanese sushi industry

This Week in Startups episode 783: Jason speaks with Finless Foods co-founder and CEO Mike Selden about how his company produces real fish meat from stem cells, the increasing scarcity of healthy and affordable seafood, more

Finless Foods looks very interesting and promising to me, but at the same time the existence of this startup concerns me being a native Japanese loving sushi so much.

This kind of innovation should happen within Japan, a country surrounded by sea and lived with it for long time, not America. Looking at it from different angle, what we are seeing here is a American startup disrupting Japanese sushi industry.

I’m not talking about Japan vs the rest of the world or anything like that. I’m talking about Japanese government and Japanese startups only trying to ‘rescue’ the people in fishing industry either by pouring millions of dollars of grants into the industry or partially automating the process where human used to do with the help of IT and robotics. Needless to say, both are not considered true innovation.

I’m based out of Kyoto, and we have a top-notch stem cell scientists at Kyoto University which is known as the driving force behind a Nobel prize winning iPS (induced pluripotent stem cell) technology.

Why aren’t we seeing a startup like this here? Should I convince someone to do that? Should I make an angel invest into the startup working on a similar topic and turn around? Or, am I the one who is supposed to get it done myself?

Hmm. There are so many things to work on, but I got a single life to spend. Life is too short indeed.