Press "Enter" to skip to content

Category: Uncategorized

Goals for 2019

2018 has been yet another significant milestone for business and personal life. As for business, we just released our first AI-on-chip called HACARUS-X Edge, enabling both learning and inference to be executed on FPGA where a typical deep learning based chip can do only inference.

For those who know a little bit history about ourselves, we started as IoT company initially. We even did a crowdfunding campaign which did not go well, but learnt a lot from it. Thus, the release of above AI-on-chip feels like going back to square one to us.

As for personal life, I and my wife were lucky enough to welcome new family member this year. Nothing makes us happier than inviting new one to our home and live together. We also moved to new place after living in the previous area for more than 5 years, giving us a totally fresh feeling to everyday life.

That being said, the following is a list of goals that I’d like to achieve by the end of 2019.

  • Produce more techno/EDM tracks
    This has been my hobby for the past 30 years. I bought my first synthesizer when I was in high school.
  • Do at least one gig
    Running a startup as full-time CEO is tough business, but it’s not excuse for me that I cannot do a gig. I’m looking forward to do live performance using my own tracks.
  • Write article in English only
    I used to write articles in both Japanese and English. This is my statement of intent going global.
  • Spend more time with family
    Work-life balance is a challenging thing at growing startup. However, a great CEO is always top business man as well as family man.
  • Spend more time outside Japan
    Now, this goal somewhat contradicts above goal. This is required for me to think big and look at the long term.

What’s your goal for 2019?

New Rockmelt is Almost What I Need

So it’s been more than two months I started working on Google Reader replacement.

I got a some sort prototype and it’s working okay for personal use. Thanks to AWS CloudSearch, I got even search functionality that other RSS reader apps don’t have yet.

I tried a couple of new comers like Digg Reader and AOL Reader. They serve same basic functionality as Google Reader. However, none of them really suited my needs.

One day I came across Rockmelt, which I remember it as a social web browser for long time. It appears the company pivoted from browser making business to news aggregation service. It’s got Pinterest like UI and ability to import RSS feeds from Google Reader.

Well, I have to say this is THE reader app I’m looking for. Indeed, it lacks ability to access feed from social services like Facebook and Twitter, but that’s okay for now.

Unfortunately, there is less incentive left for me to create yet another RSS reader app. Maybe I will find new direction and push development further, or not.

Google Reader Replacement

I started using Google Reader ever since I started surfing blogosphere many years ago. Now that a long-favored RSS reader is being shut down. Question is which service should I switch to?

I’ve tried other RSS reader apps like Feedly, The Old Reader, NetVives, Pulse, etc. After spending good amount of time evaluating each app, I found myself liking none of them.

It seems current trend amongst feed aggregation apps is either being good old, meaning app is really a RSS reader but nothing more, or being picture-heavy magazine style like Flipboard. I totally understand why people are leaving so many comments on this blog post from Digg because they are also looking for a replacement. And, I’m one of those refugees too.

This was the moment I realized that maybe I should create one for myself. A feed aggregation app that just works for me. Good news is building a RSS reader itself isn’t that difficult, and necessary components are already available as open source or provided by cloud computing companies.

For instance, building search engine was a bit tough work back in 2007, the year Google Reader was officially released. Thanks to Amazon CloudSearch, with this service available at my hand I can outsource the entire search engine to AWS and forget about scalability and maintenance issues. This allows me to focus on building features that make the app standout from the crowd.

Nowadays a difficult part is deciding what to do with the app, not how you are going build it. Below is a tentative list of what I’m going to do with the app.

  • Manage feeds from both social network and RSS.
  • Search by keyword in my feeds as well as feeds subscribed by other people.
  • Minimalistic design. No magazine style.
  • Mobile app with offline cache.
  • Fast response.

And, here is a quick wireframe for prototype. I wanted UI to look like Google Reader so that people like myself don’t have to go through a whole learning process again.


A challenging part is universal search. As I mentioned above, a search engine itself can be outsourced, but getting desired results from various data sources still requires a fair amount of engineering work. In particular, I would like to search for a specific keyword from three data sources: feeds from my social network, RSS feeds subscribed by me, and RSS feeds subscribed by other users.

If everything goes well, I might be able to release first prototype in a short period of time. If I do, I will post it at this blog website.


Running Python on iOS and Android

Usually, I don’t do blog post about technical topics, but this time it’s different.

Disclaimer: I have Computer Science degree from US university so in fact I love talking about technical stuff. I was pretending not to. ;)

These days I’m spending a lot of time playing with Kivy, an open source Python library for multi-touch app. The app developed with Kivy can be deployed on many platforms including iOS and Android.

If you are a seasoned mobile app developer, your first response might be “Wait. Can I run Python app on my iPhone?” The answer is yes.

Kivy packages your app along with Python interpreter and statically links it to the app itself. No dynamic linking. Hence, you can submit your app to Apple (and Google) without problem.

(For those who are unfamiliar with Apple rules, dynamic library isn’t supported by iOS and your app will be rejected if the app uses one.)

Long story short, one thing that makes me exciting is a possibility of creating Unity like IDE (Integrated Development Environment) on top of Kivy, where developers can directly generate iOS and Android apps while entire code is written in Python.

Why do we need yet another tool when we have Unity? Well, there are a few reasons.

First, competition is always good not only to the end-user but also to the industry. Second, developers might be able to benefit from numerous amount of existing Python libraries that are battle-tested and proven to work. Third, I’m a Python guy.

Disclaimer: I have written Coworkify using Python. The app is built with Pyramid framework. For the record, I’m a Python guy.

I’m not saying Unity isn’t good enough. In fact, it’s one of the best tools out there for cross-platform development, especially for iOS and Android gaming apps. However, developers are only allowed to use C# and JavaScript since Unity’s script engine is built on top of Mono, an open source .NET framework.

What if we could combine the best of two worlds: Python being as powerful language and Unity being as easy-to-use IDE that can instantly generate mobile apps on the fly?

It seems my next project will be somewhere around this topic, and I’m sure I cannot do it alone. Please shoot me an email or write a comment below if you are interested in working with me.

There is no guarantee this project will be successful or it will be any sort of startup business, but at least it is an interesting topic to investigate, I believe.

Building Lean Startups workshop

I got this info from my folks at Pollenizer Singapore so I re-post it here. If you are into the startup scene in Singapore, this event might be interesting for you.

Building Lean Startups is the introductory course in our Academy training program. The course covers lean startup theories like minimum viable product and pivoting using a practice-oriented oriented approach.

Our goal is for participants to leave the course with specific tools for implementing lean principles in their startup or workplace.

The course materials have just been updated, with a stronger emphasis on lean analytics and on making the materials work for corporate participants.

Of course startups are also welcome and will receive our 70% startup discount. The course is open for registration here [].

Below is the event details from EventBrite website.

This course is how Pollenizer employees learn their craft. We have been building startups for the past four years and have refined a practical approach to building “lean startups.” This is a science for transforming a feeling in the gut to a functioning new businesses as efficiently as possible. This course will re-wire your brain to the art of new business discovery. If you want to start a business, or innovate from within an existing company, you will value this journey.

Course Outline

  • Disrupt: find an opportunity
  • Focus: have impact
  • Discover: test your assumptions
  • Loop: iterate to success
  • Pitch: the story of value

Who should attend?

  • Entrepreneurs
  • Product Managers
  • Executives and managers with a mandate for innovation

What can you expect to get out of this course?

  • Practical knowledge in developing new products using lean startup approaches
  • Creative approach to unlocking new innovative ideas
  • Professional course content that can be applied to your business

For more information please visit Pollenizer Academy.

Every Startup Needs to Have Wow Factor

There is a favorite quote of mine borrowed from Zappos.

WOW is such a short, simple word, but it really encompasses a lot of things. To WOW, you must differentiate yourself, which means doing something a little unconventional and innovative.

You must do something that’s above and beyond what’s expected. And whatever you do must have an emotional impact on the receiver. We are not an average company, our service is not average, and we don’t want our people to be average. We expect every employee to deliver WOW.

In a today’s highly competitive market, companies without much differentiation are forced to go out of business no matter how well the companies are funded or how experienced their founders are. It’s simply because market isn’t big enough to sustain all of the incumbents and customers tend to choose the best one or two of them only.

So what’s a WOW factor anyway? For web startups, it might be a product that truly disrupts existing industry. For content startups, it might be something original and unique that makes people say “Wow, I’ve never seen this before.” For startups in general, it might be a superior customer service like one offered by Zappos. It’s easy to say, but difficult to execute.

From time to time I happen to face a meeting where all attendees are discussing topics that are completely irrelevant to their customer’s interest and have nothing to do with differentiation of their company. If it’s an operation level meeting, perhaps that’s okay. If it’s a company board meeting, that’s disaster.

Above statement applies not only to startups, but also to big players. Sony used to WOW people for long time. However, does 4K TV really WOW customers today? Well maybe, but so does its competitor Samsung.

Startup needs to understand WOW is a critical factor and treat it respectfully or startup will have a tough time down the line without it.

Being Hyper Local vs. Chasing Big Market

One thing I’ve once again recognized from yesterday’s event was strength of being hyper local.

Amongst 9 startups that pitched on stage, Hong Kong-based ButterBoom caught my eyes. They said it started as a fashion blog site covering celebrity photos, luxury shops and related stories in the region. The service offered by the startup can be particularly useful to those who are traveling and shopping in Hong Kong. At the end of the pitch they talked about their intent to expand business to other cities in Asia.

Being hyper local may sound against fundamental strategy that startup has to always think bigger because the bigger market size startup is aiming, the better chance there is for startup to become a billion dollar company.

To realize this, initial set of questions any investor asks you always includes “What is your market size?” Therefore, you are so tempted to say “Sir, we are looking at XXX million dollar market.” You say this because you want to attract investors by giving impression that your startup will grow like anything and become a billion dollar company in the future.

Unfortunately, reality is completely opposite. Startups aiming at such big market right from the beginning never succeed. If you design your product just for everyone in favor of big market, probably you will face difficulty with attracting people for long time.

Instead, design your product to very specific group of people. Say communication app for people who have their parents needing daycare service but those who are unable to speak or write. Market size may not look big enough initially. However, this kind of service is not easily replaceable by others and people will gladly pay for product as long as it removes their core pain.

While you design your product this way, important thing is that you need to look at something ‘repeatable’ and ‘scalable’ at later stage of startup cycle. By repeatable it means product can be replicated at different location to a slightly different group of people. By scalable it means company can increase size of business by pouring more money into cost factors (e.g. marketing, sales, development.) You can always start your business in small area or with small group of people, copy it to somewhere else and then scale it once your business model is verified.

So the better answer to above question is “Sir, we will become a billion dollar company but we need to start with this niche market because…” Facebook wasn’t designed for everyone initially. It was a website exclusive to students at Harvard. Telsa did not aim at wider audience of business person with model S initially. Instead, it captured enthusiastic fans of electric sport car. These companies repeated what’s already working and scaled it until no other competitor can catch up. Same thing can be applied to any product or service.

Hence, there is nothing wrong with being hyper local. It’s just a matter of how you draw a path from the point where you started to the point where you want to reach.

A Word About Government Grants

I’m writing this post while rehearsing my pitch for Myojo Waraku, Japanese SXSW taking place Sep 8 & 9 in Fukuoka. I’ll be on stage demoing latest product there.

A couple of months ago I asked one of my prospective investors/advisors who live outside Japan whether it’s good or not for early stage startup to participate pitch contest. His answer was basically yes as long as it helps startup gain visibility and attract audiences that it has to talk to: investors and customers.

From time to time we see same kind of contests initiated by government, and in most cases startups are given some sort of grants by government in reward for winning the prize. Only difference is that the former is just one-off event, but the latter is long-lasting event that may span across many months if not years.

This type of government-driven pitch contest or sometimes called ‘collaboration’ can be seen often in countries and cities where they promote themselves as next Israel or next Silicon Valley. People at government believe pouring chunk of public money into startups may foster foundation of whatever next thing they are trying to be.

While it is true that government plays important role as substitute for angels/VCs where there is shortage of venture money (aka risk money) in industry, there is a couple of things you need to be aware when taking money from government.

Among them, one pitfall that is particularly cautious to early stage startup is that everything has to be spec’ed out up front and you need to produce hard evidence at the end of funding period. For those who exercise Lean Startup Methodology it’s obvious that product that’s being spec’ed out prior to hitting market is almost guaranteed to fail.

There is even greater risk for early stage startups taking money from government such that it has to compromise something important, speed. And, speed is sometimes the only advantage over incumbents that startup is trying to compete.

While you are busy talking to other parties participating same pitch contest and preparing whatever materials you need to submit to government, some smart startups may come in and just get market before you actually even start. Not to mention it becomes harder for startup to pivot once you start working with government and it may be too late to pivot if you wait until funding period is over.

At the end of the day, government is not a customer and startups facing customers on a daily basis always win. Facebook, Instagram, AirBnB weren’t born out of government grants (as far as I know.) They might not be successful today if they scarified one of the most important assets at the time, speed.

I’m not completely against a whole idea of government grants helping startup kick-off business. It does help startup up to certain point. But you need to be 100% sure about what you are receiving from it and what you are giving away to it before committing great amount of your time to work with government and other parties associated with it.

Creating Momentum

One common mistake I see amongst first-time entrepreneurs is being trapped with feeling of “not ready” for too long.

This is the state of mindset where you feel like your product isn’t good enough to attract your audience or think you only have one-time shot for major media coverage or press release.

My advice: Forget all that. Just ship your product and start talking to your customers ASAP.

Successful entrepreneur is good at creating momentum around himself that somehow makes people believe this product is a game changer regardless of whether the product is ready or not. He just puts out whatever he has at the moment in front of people and start talking to people to understand what they think about it.

During this process he knowingly creates sets of initial fan base aka. early adaptors who bring another thousands of users in following weeks or months. Perhaps, I must say acquiring early adaptors isn’t good enough these days any more. You will soon realize by talking to prospective investors that you will need to acquire ‘paying’ customers early on. Otherwise, you are not on table to discuss with them.

Unfortunately, founders with technical background aren’t good at this kind of things. It’s still pretty rare that they understand importance of customer development at early stage unless they read books like The Lean Startup or The Startup Owner’s Manual, and exercise what’s written there by themselves.

Particularly, CTO type of person tends to spend too much time talking about things like architecture, framework, scalability, security and so on that don’t matter with acquiring ‘paying’ customers at all. If you are that kind of person, stop doing so. Start focusing on shipping whatever you have right now. I know this because I was once that kind of person.

Second advice to students: Don’t try to learn new way to produce things. If you are engineering student and already know how to code with PHP, don’t spend time to learn Ruby on Rails or Django or Express. Just ship Minimum Viable Product with PHP and verify your idea can turn into actual business. Then, start worrying about technical stuff.

Lastly, Startup Weekend is the best place to learn what I described above from forming team to keeping momentum after the event is over. We plan to have a couple of more Startup Weekend events in our local region and will announce them as date nears.






残念ながら、技術系バックグラウンドのある創業者というのは、この手のことがあまり得意ではなかったりします。今でも、この手のことの重要性をしっかりと理解している人は稀で、それこそ The Lean Startup や The Startup Owner’s Manual のような本を読んで、そこに書かれていることを「自分の手で」実践していないとダメです。

特に CTO 型の人は、アーキテクチャー・フレームワーク・スケーラビリティ・セキュリティといったことに時間を使いすぎる傾向がありますが、これらのことは大切ではあるものの、「お金を払ってくれる」顧客と獲得することには、ほとんど関係ありません。



もし、あたながエンジニア系の学生ですでに PHP でコードが書けるなら、Ruby on Rail や Django や Express を学ぶために時間を使う必要はありません。それよりも、Minimum Viable Product (必要最低限の機能をもった製品)を出荷して、自分のアイデアが実際にビジネスに成り得るかどうか、いちはやく検証しましょう。技術のあれこれを考えるのは、その後からでも十分です。

最後に、上記のことを実践する最適な場所として Startup Weekend を紹介させてください。チームの作り方から、イベント終了後の「イケイケ感」を維持する方法まで、おそらく全てを経験できます。


New Coworkify, New Vision

We released new web app as public beta yesterday.

This is the result of our effort listening to feedback from our own users for months. In this web app you’ll that notice lots of features used to be there are no longer available. Part of this change is due to our re-focus to a bigger issue that businesses and governments are facing with excess of unused workspaces they own.

Coworkify is now an infrastructure to help these businesses and governments get the most out of these existing assets. Hence, we are more than just a coworking space listing service.

To talk about implication of this new web app and how it complements existing mobile apps, we must also talk about upcoming “wish list” feature that we plan to release in the near future.

A “wish list” is the place where users can vote for workspaces they wish to unlock. By having enough number of votes we believe it will be easier for businesses and governments to think about opening up their spaces since there is known demand up-front.

We understand this is a whole new thinking and requires some amount of education to venue owners. However, it will give huge impact to industry if it works.

We welcome any comments and feedback from people. Thank you.



Coworkify は、企業や行政が既存資産を十分に有効活用できるようにするためのインフラを目指します。従って、これまでのコワーキングスペース検索サービス以上の存在になりたいと考えています。

今回の新アプリの影響と、既存のモバイルアプリへの関連性を説明するために、今後リリース予定の “wish list” 機能について触れておかねばなりません。

“wish list” は、ユーザーが unlock (解放) したいスペースを登録・共有し、投票する場所になります。ユーザーから十分な投票を集めることで、企業や行政は、自身が所有するスペースの解放に対するニーズを事前に知ることができます。

これは、全く新しい考え方で、スペースの所有者の方々に、この新しい考え方をご理解頂くのに少し時間がかかるかもしれません。しかし、この考え方が受け入れられれば、Coworkify は世の中に対してとても大きなインパクトを与えることができると考えています。