Stan Liu

Stan Liu

Sometime during late 2004, I got introduced to Stan Liu by a prominent author in the Macromedia Flash Community. It was the start of a lasting relationship which I will cherish throughout my life.

Stan Liu became our client. My team and I worked for him. He was really happy with the result. He offered me a partnership and owner equity to the new venture. Later, he even went on to get a term sheet from an Investor, north of $1.5M. However, he felt the terms were not in our favor. By then, my company got acquired and we both drifted apart, hibernating our business idea.

Stan and Joan

In the summer of 2005, I got a good opportunity to visit San Francisco. I got invited to Macromedia’s Lego program to spend a week with some of the best Flash folks in the world. I also decided to head to Los Angeles to meet Stan Liu. I have many fond memories of meeting the family — Stan, Joan, and the then 4-year old Khylen.

I had my first experience with a luxury sports car — rev his Porsche (and years later his Maserati). I also experienced my first authentic Chinese cuisine. I learned a thing or two more about Hollywood, got pretty much free entry to Disneyworld (he worked at Disney). I had my first Sushi at the Yoshida Sushi Bar. Later that winter, I went back to his place and repeated many events, including spending the Christmas-eve with his family and friends.

He encouraged me to try this, try that, made me experience many new events. I remember us attending an Adult Film Industry conference – laughing like crazy in-between our animated poses taking pictures with the stars. Unfortunately, I lost that photo collection.

When I left them that winter, River was on his way.

In 2010, by the time I left Paisa (Infinitely Beta) and was pondering on what to do next, I talked to him. We started Levoma, Inc. He, pretty much, bankrolled my stay in Silicon Valley to be part of the Founder Institute in 2010-2011. I survived frugally but he was always worried that I might run out of pocket money. When I visited them before I leave for India, I met the whole family – Stan, Joan, Khylen, and River.

Unfortunately, the business failed; we pulled the plug and decided to work together on something else in future.

The 8th of Aug, 2014 was a sad day. Stan pinged me on Skype for a talk. We had a long talk. He was sick and at a late stage. However, we drifted and discussed few business ideas instead. He was full of ideas, all the time.

In the summer of 2016, I visited Stan and his family once again. He picked me up at LAX, “You looked pretty much the same.” I replied, “I was worried but you look pretty cool too. I don’t think anything is happening with you.”

I realized he was tired, he was no longer as agile as I knew him earlier. But, he was still energetic and going strong. We discussed few business ideas again, introduced me to some new interesting VR stuffs, went to the movies with his family, joined their weekend family lunch, went shopping. I played few games with River and he beat me real bad; just like how Khylen used to beat me in his video games long back.

He saw me off to the Airport. As we hugged goodbye, he asked, “Do you need extra dollars in cash? Extra baggage, coffee, food! You never know when you need it.” He does that every time he drops me at the Airport. “I’m good. I’ll think over the idea, think of a plan. Let’s talk more.”

Last week, I got to know that Stan Liu is no more.

He was a caring person. I’ve lost a mentor, business partner, and good friend. Rest in Peace.

How product designers use Big Data Science to make durable products?

Two years ago, a self-driving Car was launched by Google. However, it was only a few minutes later that it blew brains of the people by stopping in the middle of the street when it confronted a woman seated in a wheelchair while she was chasing a running duck with her broomstick. The car took a halt till the road became clear and then started making an advanced contribution to the worldwide traffic safety.

The products have been evolved into smart and intelligent devices with embedded systems that are connected on the broader level, reshaping the product design, radically, on the whole. As the car stopped in the middle, this particular information has been sent to all the Google car by which no car will ever make a mistake when someone in a wheelchair comes across the street chasing a duck with a stick.

Another example that displays the advancement of product design is the changing design of smart thermostats. Designed to control the home devices by transmitting the individual data of the house to the manufacturers, these intelligent machines will be connected to a network that autonomously optimizes through coordination. The data that has been streamed on the location, environment, and operation will be sent to the makers who process it and then make upgraded products that eliminate all the errors.

Continue reading

Accuracy in Analytics depends on understanding underlying cultural mechanisms

Former Speaker of the United States House of Representatives Tip O’Neill famously said “all politics is local” when referring to the basic principles of his profession. In many ways, the same can be said of analytics. While the internet as a whole is sprawling and its user base massive, this seemingly singular entity is ultimately made up of thousands of different cultures and sub-cultures.

Similar to how a member of Congress wins re-election not by appealing to the nation but to his or her constituents back home, successful data science is less about deciphering the big picture and more about understanding the nuanced reasons why relatively small demographics generate the sorts of data patterns they do over time. To this end, there’s a certain element of anthropology involved in analytics.

Companies in Silicon Valley and other tech hubs around the world are acting accordingly. The decision to outsource analytics is becoming a more common one in these business circles. Rather than a matter of cost, it’s a matter of remaining relevant to a diverse spread of users.

In fact, tech companies in San Francisco and elsewhere are preferring to outsource analytics on an individual basis in order to appeal to a more talented pool of data scientists overseas. Rather than look for existing facilities to transfer duties to bottom dollar staff, businesses are offering enticing arrangements for independent contractors. These often include automatic payment via many international money transfer channels, scheduling freedoms, and access to cloud services for openers. If you are a top-level data scientist in China, India, or Sub-Saharan Africa, it is a good time to be open for business as a self-employed professional.

For tech business leaders getting inspired and are hiring data analytics specialists abroad, it’s critical to note the aforementioned geographical settings are themselves extremely broad generalizations of where the growth in internet users is its ripest. Within China, India, and the Sub-Saharan Africa are many dozens of sub-regions each with their own cultures.

Continue reading

Meeting the Startups at Conquest, 2017

Last weekend, I was invited to talk and critique some of the Startups at the Conquest International Startup Challenge, 2017. It was short sessions of about 20-30 minutes each with the Startups. Most of them are looking for investments. The generic suggestion I mustered up was not to limit to investors in India but to look outside too. The other key missing piece was that they need to hustle a lot.

Here are the Startups I talked to;

Trell

Started as a handle on Instagram, Trell went on to become a sensation amongst young travelers, who love to look at pictures of local places to travel to. They leverage the finger-snappy millennials that love photographing places they visit and sharing with their friends and fans.

The Trell App is a buffet of picture-stories of interest to users who want to explore new places with their friends.

The team is doing a good job of hustling with the right audience, they have a really good traction. I was able to give them few technical feedback and suggestions, especially with the UI/UX of the app. They were interested in a more in-depth technical discussion on how to scale their image hosting/delivery mechanism to give their users the best picture quality at the most optimized setup. It is a solved problem and they should not worry too much about it.

Continue reading

How To Quit Your Job And Become A Freelancer

The reason you show up day after day at your job is because you need a paycheck. On the surface of it, this looks like a good reason to continue to drag yourself to a place you’ve quietly come to despise and to associate with people with whom you have little in common. But, when you examine your predicament a little more closely, you can be much happier, freer, and more prosperous by becoming a freelancer.

So how do you become a freelancer?

Start by deciding what it is that you love to do. Once you’ve identified what it is—writing or photography, graphic design or making videos, throwing parties or organizing spaces—then it’s time to figure out how to get paid for it from multiple clients.

Next, come up with a business plan and fill in as many details as possible.

Finally, launch the business by building a website. Theoretically, you can launch your freelance business without a website, but it’s so much easier to interest clients if you can direct them to a website that explains what you can do for them.

Continue reading

Three Things Artificial Intelligence did in 2016

Not long ago artificial intelligence was something found in the pages of science fiction novels. Today it is becoming more and more of a reality. While we don’t have computers plotting to take over the world, AI’s did manage to do some pretty amazing things in 2016.


The Next Rembrandt – Case study from J. Walter Thompson Amsterdam.

  1. In March Deepmind’s AlphaGo was able to beat the world champion 4-1 at Go. Go is a very old Chinese game considered much more complex than Chess. For comparison, an AI beat a master at chess all the way back in 1997. Prior to AlphaGo scientists thought AIs were far from being able to beat human Go competitors as they were struggling in even amateur matches. For its win, AlphaGo received an honorary master title.
  2. Human characteristics like imagination and creativity have been the hardest things to get computers to do. In April 2016, however, an AI was able to create a new Rembrandt painting. A team, over the course of more than a year, showed various paintings of the famous artist to the computer. The painting it produced won several awards and generated a lot of discussion about artificial intelligence.
  3. Language is one of the basics of what being human is. Google created the Google Neural Machine Translation, or GNMT, to improve its translation service and it has done some remarkable things. First, it started translating between two languages despite never being given examples of them. It then did something even more amazing and created its own language. Because of its success, Google is now using it in all of the translation service.

Artificial intelligence did a lot more in 2016 besides these three accomplishments. Building on the new advances made 2017 will no doubt bring us even closer to true artificial intelligence.

3 Essential Steps for New Startups

Most entrepreneurs truly believe that their ideas have the potential to make them wildly successful. That turns out being true for some. But not all startup ventures reach the level of success their founders expect, and some are complete failures. Adding the 3 steps outlined here to your startup plan will improve your chances of a successful launch.

Objectively Assess the Need

It’s all too common for would-be entrepreneurs to think of ideas for products or services to sell, and quickly jump in with both feet trying to market that product or service without objectively assessing the need for that it. Here’s a stark reality: if the market doesn’t perceive a need for your product, it won’t sell; and you will have wasted valuable time and money creating and marketing your product.

Supply and Demand

Identify your Unique Selling Proposition

Whether you’re launching a completely new and revolutionary product or something that’s similar to what others already sell, you must identify your Unique Selling Proposition (USP). Once you’ve developed your USP, incorporate that into your marketing strategy. Remember, just because you inherently know what makes your product better than similar offerings doesn’t mean your prospective customers will.

Plan Ahead

Virtually every startup is fraught with potential problems and barriers. What those might be for you and your startup depend largely on what you’re bringing to market. For example, some product categories are governed by strict regulations, often from multiple levels of government and multiple agencies within each level. In fact, dealing with regulatory requirements is often the single largest expense associated with creating a product and bringing it to market. If you don’t have the resources necessary to navigate the regulatory environment, you could find yourself dead in the water before you even manufacture your first product. But that’s not the only potential problem. In fact, there are so many, we can’t begin to list them here. But we can tell you that you must objectively brainstorm for potential problems that you might face launching your product, then develop strong contingency plans to deal with them, if and when they arise.

Ideally, you’ll incorporate these steps into your plan before you launch your startup; but if you’ve already launched and haven’t completed one or more of these steps, you may want to put everything on hold until you complete the missing steps.

Featured Photo by Mari Helin-Tuominen, Supply and Demand Image from How to Estimate Market Demand for a Product?

DD-WRT, VPN and Apple Airport Routers

Apple’s Airport Extreme and Express are brilliant at what they do. They are secure, yet easy to setup and maintain. Well, you do not need to maintain them – they tend to just keep working. It makes it super easy for an Apple product eco-system to co-exist without much hassle and fuss.

Airport Wireless

After many Linksys and D-Link Wi-Fi Routers, I started with the first generation Airpot Express and upgraded to the first Airport Time Capsule when it was released in 2008. The time capsule (Airport Extreme with Time Machine enabled Drive) lasted for 5 years and the Airport Express, a little over 6 years. Our current Home Network Setup is powered by an Airport Extreme and few Airport Expresses.

However, the apple routers are limited in their functionality. I wanted a VPN sitting in between the Internet and my home network without disturbing my original setup. I also wanted to have the option to turn the VPN OFF/ON quickly as and when I needed. I researched for a bit and settled on a cheap flashable wi-fi router – Asus RT-N18U. I flashed the Asus router with DD-WRT.

There are other Open Source Router firmwares which are equally good. I chose DD-WRT, as I find it easier, and have used it earlier. Some other alternatives you might want to try are – Tomato, OpenWRT, Gargoyle, etc.

I’m not very technical but I can understand technology, and knows a thing or two about how things work. So, I chose simpler setups and things that just works. My current setup works for now.

DD-WRT

DD-WRT is a Linux based alternative OpenSource firmware suitable for a great variety of WLAN routers and embedded systems. The main emphasis lies on providing the easiest possible handling while at the same time supporting a great number of functionalities within the framework of the respective hardware platform used.

Flashing a router and installation of DD-WRT is pretty straight forward. The most important part is to choose the right firmware for the router, making sure that the router is supported by DD-WRT. Following the instructions on the installation should be good enough.

Once all installed and running, here are few basic setting recommendations suggested by industry experts.

  • Change the default network from something like 192.168.1.1 to 192.168.xx.xx of your choice.
  • Broadcast the SSID but secure it with WPA2 with AES. If you’re not worried about backward compatibility, you can disable TKIP.
  • Of course, change your Admin password from the default to something better.
  • Always have a backup of your settings. It tends to be useful.

Services – VPN, NAS, the bells and whistles

The features and functionalities of DD-WRT are humongous. Pick the ones you want to use and you can ignore the others to their default/disabled state. If you have signed up for a VPN, which are much needed these days than ever, read-up on their OpenVPN (better than the other protocols so far) documentation for DD-WRT and set it up.

I really liked the simplicity of ExpressVPN, and it just worked for me. If you are shopping for one, most of the VPN Providers have a trial period – from 2 days to a week, or even sometimes a month. A 2-day trial should be enough for you to make a choice. I once got a good deal on CactusVPN and I subscribe to their VPN + SmartDNS.

Try out other good VPN providers – VyprRVPN, NordVPN, IPVanish, Buffered, Anonymizer, Private Internet Access, Tor Guard, Slick VPN, MullVad, Black VPN, Smart VPN, Private VPN, etc. You can also filter through an exhaustive list of VPN Providers or swing by That One Privacy Site to do a thorough research that fits your requirements. More VPN reviews can be found on TheBestVPN.

DD-WRT Router + Apple Airport

Finally, I combined the two worlds – the DD-WRT Router facing the Internet and the Apple Airport Extreme managing the Home Network.

Internet

Setup the DD-WRT Router to face the internet – set up your PPPoE, tethering or the IP that your Internet Service Provider gave you. Enable DHCP so that the DD-WRT router can act as the DHCP Server. You can set up additional options such as the firewall, ad-blocker, access restrictions etc. Make sure the Internet and everything else is working here perfectly. Leave the Wi-Fi enabled and working to get back to it, just in case, your primary Apple Router fails or just to debug/edit settings to the DD-WRT router. This also decouples the harsh Internet from your home network.

Apple Airport Router

This is your primary network where all your devices are connected. As the Internet is now taken care by the DD-WRT router, we have to just plug in the Airport Router’s LAN to the LAN network of the DD-WRT Router.

Play around with the Airport Router settings of your choices. However, here are few key important setups that the Apple Airport Router needs to make it working in this setup;

  1. In the Internet tab, connect using DHCP so it gets its unique IP from the DD-WRT Router. You can change the DNS server either here or in the DD-WRT router. I kept it at the DD-WRT Router.
  2. Now, Create a wireless network with the settings of your choice.
  3. For the most important part, turn the Router Mode Off (Bridge Mode). We’re not routing anything with the Apple Airport but merely managing the Wireless Internet Network.

Of course, with Apple rumored to be abandoning the router business, in future, you can just plug off the Apple Router part and move to your DD-WRT router as the primary network manager.

Oinam Home Network

That’s it. You can now have the flexibility, security of a DD-WRT Router and Wireless Network managed by an Apple Router to easily and consistently connect all your devices. All your devices, from phones to laptops to the connected TV, can connect to the Internet encapsulated via a VPN. I’m still experimenting and will continue to play with my setup. The current setup has been running for a year or so, without any issues.

Artificial Intelligence, Facial Recognition, and … Fish?

UI Faces

If you are a tech-savvy, ecologically-minded entrepreneur, then have we got news for you. Artificial Intelligence is capable of many things, and responsible for many more things than the average person may want to know about. You may already be aware that artificial intelligence is used in those dazzling fast facial recognition searches so often featured in police procedurals. Face and body recognition are a controversial, daily reality around the world.

Fish Recognition

Now, as reported in The Guardian, the non-profit environmental group Nature Conservancy is seeking interested parties to create fish recognition algorithms. Similar to the facial recognition algorithms which pick out unique facial features from amidst a crowd, the hoped-for smart program will identify specific kinds of protected fish. If successful, this software will assist inspectors and honest fishing vessels alike, drastically cutting down on inspection times while enhancing accuracy. For those actively engaging in illegal fishing, however, the news will be less good.

Real Problems In Need of Real Solutions

fish-recognition

Illegal fishing costs the world economy an estimated $20 billion a year. Like many of the problems that artificial intelligence is being called upon to address, this one is big, literally too big for human minds and eyes to handle on their own. Smart computers can help, and smart computer programmers can help make those smart computers.

For more information, including details on the machine learning contests, visit Kaggle’s site – The Nature Conservancy Fisheries Monitoring.