Artificial Intelligence, Machine Learning, NLP and Speech Processing.

I find it disturbing that a number of companies use terms like these to describe themselves to journalists. The problem is not with these words or verticals. As a company if you’re working on things like these, you’ve earned yourself the right to be called engineers. The problem is that most companies talk about these without really doing anything.

If you’re a journalist who’s reading this, please take a little time to understand what each of these terms mean, and ask the company more specifics about what exactly do they do. If their CEO can’t, please label them as bogus, irrespective of the fact that they’ve been funded or not.

If you’re the founder of a tech company, please take a little time to explain in more detail the cool algos that you’re writing. When you don’t do it, it all looks really fake.

If you’re a investor here in India, please invest in companies which are really disruptive. Mobile phone recharge is not disruptive. Local deals is not disruptive. E-Commerce is not disruptive anymore. You’re really killing all us techies out here.

Image by x-ray delta one

The Early Signs

Initially, it was all good. But then I started running into a few strange issues. It’s possible that some of them came up because I use email in a way which is different than most users. Each of my emails (yes each and every one of them, for the last 6 years) have been tagged with at least 3 tags. One tag is for the name of the client, one is for the project, and one is for the priority or the importance. This means that I have more tags than you can possibly imagine. The Gmail web app handles them better than I could ever hope. However, this became a problem with the initial version of Sparrow. I used to allocate my tag trio to each email and some time later I would search for the tag and the email would not show up. It took me about a week to realize that the tags were getting lost!

If you’ve used a software like your life depends on it, you know bugs can creep you out a little at a time. My wife would even go on to say that I became angry and irritated. That was when I knew I had to give these guys some more time to fix it.

And of course, they did. But we’ll get back to it later.

Back in love with Chrome

I spend the next month or so using the Gmail web client, and all was good, except when it’s in chrome, it’s easy for the tab to get lost. So I would have multiple windows open at any time, and I couldn’t Cmd+Tab to the email client. This made me a little jealous of all those people happily using Apple Mail and Outlook, even though I knew I could find any email on my system faster than they could on their systems, I still could not reach my email client fast enough.

A Fling with Fluid

It was around this time that I discovered the Fluid App. It kind of reminded me of the old Windows 7 days when I could open up my Gmail Web app is a new browser window and reach it faster than a speeding bullet. So I downloaded Fluid, and quickly installed it. One of it’s coolest features was that when you closed the window, and then opened it again, it opened from the exact same place. For a week, life was beautiful. I cracked more jokes and made friends easily.

Sadly, Fluid had a short life as the app running my email. It was just a tad slower and downloads were not as smooth as chrome. I had very quickly migrated both my WordPress Admins, Basecamp, Gmail and Google Calendar to their own windows on Fluid. However, the more things I added, the slower it kept getting. After some time it reminded me of the sad old browser called Firefox (some of your kids might not even know that it existed, like we never knew until much later than Netscape existed).

So, I was back to Gmail on Chrome.

Falling in love again

One fine day, a few months later, I opened up my email to find a really cool deal for all SumoLings. The AppStore sells Sparrow for $10, these guys were giving it for $6. The difference isn’t much, but I still fell for it.

So, now I had a paid version of Sparrow, and it handled my tags properly, and I was once again completely sold. I just pitched to every person I met, and did manage to sell it to a lot of them as well (including a lot of people at my company). Things were once again good, I guess you know the drill by now.

The sad finish

About a month later, our internet was running slow, and we moved on to a new leased line. However, things didn’t really improve when it happens, and I cussed at Tata Indicom on every chance that I got. It again took us a few days to realize what the problem was. The new leased line had committed hara-kiri due to Sparrow’s magic. The new brand new issues on the Sparrow app were that it took us an indefinite amount of bandwidth and CPU.

Here is the CPU usage, 1 hour after start on my Macbook Pro

This is what it was doing 8 hours after reboot

Sparrow had taken up 180% more than Google Chrome, and 84% more CPU that the second highest most CPU intensive application running on my mac.

Sadly we had to send out a company wide email to stop using Sparrow with immediate effect. I really like this software and I hope soon they fix this issue. Or at least I hope that their iPhone app works better.

Till then, it just me and Chrome.

Our recruitment process is a simple five step procedure

1. Find resumes you like
2. Ask them to fill up this questionnaire
3. Have a telephonic interview with people who send in nice answers
4. Call them for a personal interview
5. Ask them to write a piece of code in 2-3 hours

It seems to work well for us. Since our questionnaire is quite generic, we have been able to use them with designers and testers alike. The last step gets left out with those two, though I’m sure we will build up a way to improve that with time too.

Hiring is a big problem here in India and rere are some things that we consider before we recruit anybody.

1. Don’t hire people who switch companies a lot

A lot of younger developers in India just pick up whatever job gets offered first and then start looking around. The problem is that most people don’t even consider this to be a bad thing. The origin of this problem starts from our schools and colleges. People get pushed to take up science or commerce based on what somebody else thinks is right for them. The same thing happens when they join a college. Eventually, the first time that they have to take a decision is when they have to join a company, and they are just not experienced with it. So they take up the first thing that they come across, and soon find that it doesn’t work for them. Once they start switching, they realize that you get a salary jump in moving from company to company without really doing anything.

So, If a junior developer has spent less than 6 months at their last company, we just discard their resume, no matter how good it looks.

2. Check if they know why do they want to work in your company

Another very clear way to finding out if the person is the right fit for your company, is to ask them ‘Why would they like to work for your company’.

If they are just randomly looking around, they will give you some crap like, ‘we want to work with your esteemed organization’ or ‘I love the work that you do’ or ‘I want to work on this technology’. Whenever you get any of these or similar answers, you need to dig deeper with this like ‘Which product did you like best’ or ‘What do you think we can improve on the one that you like’.

The smart ones always look up who you are, and will know little details about your company. Many times they would have talked to somebody at your company as well.

The best ones will also be able to tell you WHY you are a better choice that your competitors. They will also ask you some really insightful questions

3. Write your job descriptions well so that the right people find you.

People have researched that if you use words like ‘ninja’, most developers don’t like it. ‘Programmers’ is also another bad word. You can find more details about what to use and what not to use here

4. Experience with huge services companies makes (most) people useless for startups

Some people don’t agree with me on this but I just don’t hire people who have worked at the really huge services firms (I’m talking about you TCS, Infosys, Wipro). I’ve put this lower in the order because this is not the most important deciding factor, but it still is very important. If you’ve ever worked at any of these companies, you would know that working culture if not really good. Most developers even with decent work experience, can’t sit in a chair and code for 3 hours straight.

Also, when their friends start going onsite, they will just want to go too and in smaller companies it is not always possible to keep sending them.

The only person who I have ever recruited with such a background was one who wanted to move into designing because he just wasn’t interested in working as a developer anymore.

5. If they can’t type fast, don’t take them

I’m not sure where I read this (found it), but I’ve found it to be really true. All great developers type really fast. It doesn’t mean that if you type fast, you would be a great developer, but I have never come across any great developer who typed slow. Ever.

Also, I have noticed that the people who type slow, never put in comments. I guess it just takes them really long to do it if they keep typing in long comments too.

6. Ask them to write code

This comes in really low not because it’s not important, but because everybody knows this. Still, most people don’t follow it. I myself used to do this a lot earlier. But now, we just don’t do it anymore. It takes slightly more time, but you need to get them to write a component is 2-3 hours before you can understand how they will deal with bigger problems when they actually join your company.

7. Check if they have a github profile

You can use rapportive to find out if they’ve used Github or not. If they have, give them a lot of extra points because you can go and check out that code even without talking to them. Even if they have just forked another repository and never added anything to it, give them a few marks for that because they would know how to use git. It shows that they are trying to be better developers and have read in places that you need to work with repositories.

All in all, no matter how bad-developer-proof you try to be, some bad-developers will still get in. When this happens, I would like to point you to something really important that I heard from Anand’s (from Gluster) speech at Unpluggd

If you put a lot of emphasis on hiring, you also have to put a lot of emphasis on firing. And you have to ensure that you fire the right people frequently, soon they would be 20% of your company and then you can’t do anything about it

How do you handle your recruitment?

If you need people, you have to get the smart ones, and help them get better. The colleges won’t do it for you. (And there’s certainly no dearth of smart people)

The next came soon after when he started making socks with flowers imprinted on them with the help of the new, Rs9 lakh machine in 1991.

He’s been doing it for 20 years!

There are no shortcuts.

There has been a lot of buzz lately in the tech space about the 4G mobile internet technology and a lot of people are starting to speculate that it is the next best thing. The 4G mobile internet technology no doubt promises a lot of great benefits such as its remarkable speed, the ability to work in harsh weather conditions and its reliability. The problem however is that like with any new technology most internet users are confused as to whether they should start using the 4G mobile internet technology or not. I’ll be giving you a few reasons why you should start using the 4G internet as soon as it is generally available below.

It Works for Computers and Mobile Phones Alike

The number one reason why you should be a 4G mobile broadband internet user is that it makes it completely easy to use the internet on both your mobile phone and your computer. We are already familiar with the 3G technology and the fact that it makes connectivity to the internet using your computer possible. The great thing about 4G internet also is that it works for both computers and mobile phones alike – you can easily connect to the internet using your mobile phone while at the same time tethering it with your computer.

I’m also suspecting that companies will start creating 4G modems soon that will make it extremely easy to connect to the internet using a 4G enabled sim card in a 4G area. The good thing about this is that you don’t have to be paying two separate ISPs to use the internet on your mobile phone and your computer – all you need to do is stick with a 4G network and tether it with your computer.

It Provides Advanced Security Compared to Other Forms of Broadband Internet

Another major reason why you should start using a 4G mobile broadband internet connection is because it offers exceptional security when compared to other forms of broadband internet services. Unlike the wireless broadband internet connection that is easily exposed to any hacker or the cable internet connection that is always on the 4G mobile internet technology is highly robust and protected against most form of attacks. When it comes to security the 4G mobile internet technology is like an improved version of 3G – and what this means is that you no longer have to worry about your security being compromised.

It Provides Great Speeds that Makes it Easy to get Things Done

There is no point in going after a particular internet connection in the world we live today if it isn’t extremely fast. The 4G internet connection is super fast with download speeds as high as 100mbps. It is also very great not only because it is super fast, but because it is consistent with its speed. There is no point in using an internet connection that is fast today and slow tomorrow and you can be rest assured that you won’t experience any such problems with a 4G internet connection.

The blog was probably started on June 11, 2005 and has had a lot of successful and unsuccessful moments over the last six years. Now, it seems that Mike Arrignton is being unceremoniously shown the door (Though it’s not half as bad as what happened to Carol Bartz). The game is still on.

DHH is out to bash TC in a nicely written article and I really hope I could join him this time, but since I run a services business 6000 kilometers away, I couldn’t care less.

Anyways, here is what TC looked like at the end of 2005. How many of you have seen this version of the site?

Ever since Apple’s AppStore caught everybody’s attention with their 15 billion app downloads, we have been caught in an exploding frenzy of third party application stores. The concept is really catchy, because it allows third party developers to build applications and sell them without spending too much time on marketing the apps.

However, things are not hunky-dory with most app stores. Other than the iOS appstore, which really helps developers make serious money (at least some of them), the others mostly suck. The Android app store, which is probably the second biggest store after Apple’s appstore, is quite pathetic when it comes to the amount of revenues it can generate (there are a few people who made money there too). Blackberry is almost insignificant, and I wouldn’t even want to talk about the ones from the telecom operators. I have never met a company or a developer who has made money off the other AppStores, and I doubt if many will.

The thing that brought this post on was this post on MediaNama. Now, I usually ignore news which talks about another AppStore, but this one caught my eye. The article said that this was started by Upasana Taku and Bipin Preet Singh, these guys are also behind ZaakPay, which is undoubtedly one of the hottest startups around.

So, I headed over to MobiKwik via the link on HackerStreet and I think I’m a little disappointed. It’s a single page before you sign up, and even though there are 11 points about why one should sign up, there are only a few things which you can’t see in most places. These are some of them – You can withdraw your payments anytime, a security mechanism to prevent piracy and that the developer has full control over the look and feel of the app pages.

I’m sure there are a ton of things I still don’t know about the MobiKwik AppStore and given their history, I’m sure Bipin and Upasana will over time turn it into something worthwhile. I wish them the best of luck.

But at the same time, as a developer, it’s time to define a set of rules based on which we can define if we should be submitting apps to that store or not. So here we go

1. SALES

It is the job of the AppStore to mention HOW and how MANY sales are they making. The HOW defines how the company is marketing their AppStore and what are they doing to promote it. How MANY will define what kind of sales are they already making. If you can’t share exact numbers, you need to tell us how is the top developer doing on your network. I would like to see links and testimonials to people who are using it and are making money there, so that the developers and the bloggers can go and talk to them to see if they are really making money.

2. PAYMENTS

The MobiKwik AppStore makes it quite clear that they allow devs to withdraw payments to bank accounts as soon as possible, however as we all know things are not as simple with the Indian Government. If I’m going to make serious money on an AppStore, I need to know how is Service tax going to handled, and what about TDS. What happens when somebody in the US purchases via the AppStore. Also, if the way we withdraw money will change, how much time notice will we get.

3. PRIVACY

I need to know what exact steps are being taken to prevent piracy because I do not want to spend 3 months building an application which is going to be cracked in 3 days. If the stores targets multiple operating systems, can you ensure that the app won’t be cracked. If you say that it won’t and then it does, can I sue you?

4. LITIGATION

As we have seen with Lodsys, I want to know that if somebody sues me for using a service that the AppStore provides, how exactly will you help us. Many of the patents these days are generic enough to be targeted at any Store. If you’re not going to do anything, tell me before I start selling.

5. POLICY CHANGES

If your policy is going to change, how much notice will I be given so that I can adapt to it. For example, the Amazon AppStore for Android started a Free App A Day promotion, in which there were going to give the app out for free, and pay the developer 20% of the price of the app for the same duration. However, they just changed the policy about paying developers and there were developers who sold 1,00,000 copies of their app, without making a single rupee.

I guess these are good enough to start out with. If you’re selling apps on any AppStore other than iOS and Android, I would love to hear about the kind of issues you face.

Image via Cristiano Betta

UPDATE There is a great discussion on at HackerStreet between Bipin, Arjun (Who runs TaazzaGO, based on his website, I think I like his taste!), Upasana, Vijay and lots of others.

Usually when a project starts, everybody is on a high and it’s difficult to say if the project will go out on time or if will rot with delays from both sides. However, by the time that you reach the middle of the project, things start getting much clearer. A couple of times I have tried to write down the warning signs, which might tell me if the project is going out on time or not, but the number of cases are so many that there are still cases when things go wild.

So, this post is basically a list of all warning signs which tell you that the project is going down.

1. Non-specific specification

Specifications are supposed to be specific. If they are not, it is usually a strong indicator that things might go wrong. If you think that this is obvious, you have never really worked in a services company before. The problem is not having a spec, there always is a spec. You can ensure that before you start the specifications are complete before you start. The problem is that the client has the authority to change the spec as time progresses. The reasons for it are many. There are always features which weren’t thought out before, some other website does the same thing in a better way and you want to do it.

2. The wrong people on the project

Another important reason for delays is putting the wrong person on the project. Each developer and manager have their own style of doing things and you need to assign them appropriately. Some people require specs set it stone to work really quickly, and some are really fast at improvising whatever was initially discussed. Both of sets of people are really helpful, and putting the wrong person on the wrong project gets everybody in a mess.

3. Getting most of the payments at the end of the project

This one is slightly tricky. Most clients believe that if say it’s a 3 month project and we charge 50% upfront and 50% at the end, we will have a lot of incentive to push towards closing it. This sadly, isn’t always the case. Most of the time, if a 3 month project, pushed to a 4 or 5 month project, even if you don’t want, you need to move developers to another project after the first few months because somebody has to pay that month. So, the incentive to work on the project is always higher, when the payment date is closer.

4. Late Engagement of the Client’s Testing Team

The client usually doesn’t put in the the complete testing team on the project until the very end. This means that many of the things which we were sure about suddenly are counted as not done when their testing team comes into play. It not only makes it impossible to define timelines, it plays havoc with the team’s morale. You need to push them to start testing as early as possible.

5. It does take a few months to create the spec

Most of the clients assume that they know exactly what they need and that specifications and mockups will only take a couple of weeks at most. Sadly, this never happens. For a project which will need three months to get developed, it doesn’t take less than 6 weeks to prepare and finalize the specifications and mockups. Even then, there are always modifications to it as soon as the project starts, midway during the project, and at the end. So the total time required for mockups and specifications is approximately equal to 2/3 of the time required to develop the project (for new products).

6. Unstable External APIs

Most mobile products require to connect to an external API through which it interacts with the world. If the API that your product needs to work with is not stable, life’s going to suck for the developers, managers and the clients. This is because everything that starts working once, will fail when the API is changed even slightly. There is nothing which kills the morale of the dev team faster than writing against a changing API.

These are some of the reasons which really derail the speed at which the project runs. If you seem to run into some more, do let me know.

Image by queenvanna creations