Stick to boring tech when building a start-up

[pingu]

Reddit CTO: Stick to Boring Tech when Building Your Startup - The New Stack

 

[Daniel Duffy]

Internet has killed creativity.

 

[pingu]

Internet has killed creativity.

I'm not sure if that's what the article means.

 

[Daniel Duffy]

I'm not sure if that's what the article means.

It was a comment. What does the article mean?

I think they are talking about tools for business and not the business itself.

 

[yetanotherquant]

Reddit CTO: Stick to Boring Tech when Building Your Startup - The New Stack

Good article!
I can share some of my experience.

Prologue:
@Daniel Duffy : do you remember your opus about €200 - €500 hourly rate?
I did found a pretty talented developer in Ukraine, who works for €100/mo (+ of course an xx% share in project).
We decided to develop not a financial supermarket but rather a more focused project and we progress with a good tempo.

Story:
The first year of any new startup’s life will be “scalability hell,” he said. “This is the problem every tech startup faces.”
Yes! Still we do it in PHP, because a) my developer knows this language better than any other server-side language, b) I started my career as Perl/PHP programmer and c) using PHP is probably the fastest way to get a working prototype.
We'd better rewrite the project from scratch if we encounter scalability hell (and we do know how to do it) but so far we'd prefer to have something working as quickly as possible.

On the other hand my developer also wants to learn a new demanded technology (in this case even if our start-up fails, it will not be a waste of time for her). So I don't oppose when she uses this rich .js stuff on client-side. I would prefer to do it in the old way (as it was done 10 years ago) at least for the first prototype, because in this case we would pace faster and one can still create a usable site without these new tricks. But if I had insisted on the "obsolete" technology, her motivation would deteriorate.

 

[Daniel Duffy]

Sorry, yetAQ, the last time I did hourly consultancy business was 1992. I have no idea what consultants ask/get.

I did found a pretty talented developer in Ukraine,
Good for you.

We decided to develop not a financial supermarket but rather a more focused project and we progress with a good tempo.
What % of budget is Skype? Overhead? A potential problem is that IT/CS folks tend to have difficulty with the maths end of the project, thus slowing it up. Unless they are working on GUI/DB.

(+ of course an xx% share in project).
Sounds logical. Otherwise why work for the scrawny wages?? feed them peanuts and you'll get monkeys.

I doubt if there is a long-term business model for garage quant shops. Compare to CAD, the garages dropped from > 200 to just 5 in 5 years in Europe.

But if I had insisted on the "obsolete" technology, her motivation would deteriorate.
Let's say you get a mega project to update a VC++ 6.0 app to VS2015? What do you do? Is she indispensable? (80% of s/w in world _is_ legacy, haha).

BTW what is your hourly rate?

// You probably missed the internet boom end 90's those were the big days for developers and consultants.

 

[yetanotherquant]

What % of budget is Skype? Overhead?

Well, larger than I wished they to be but still manageable.
At the beginning it was really a bottleneck but we analyzed the way we communicate and radically improved its efficiency.

A potential problem is that IT/CS folks tend to have difficulty with the maths end of the project, thus slowing it up.

Yes!

Unless they are working on GUI/DB.

This is exactly what she does (well, I did DB myself because it was easier than to explain her about tickers, ISINs, OHLC prices and so on).

And as to math, I also found two probabilists (in Ukraine and Belarus).
(BTW, it is really hard to find an inexpensive programmer (or sailor) in Ukraine, but as to other jobs, €100 is a mean monthly salary).
Their ability to research the market on their own is (so far) limited, but if I state them a more or less pure mathematical problem they can solve it better than me, myself. In particular, one of them has found a very promising method to detect the market reversal points, we are currently backtesting it exhaustively.

Of course, such constellation requires a person like me, who knows both IT (in particular web technologies), math, financial markets and last but not least has a soft skills to manage the team. But once there is such person, the offshore development does work!

(+ of course an xx% share in project).
Sounds logical. Otherwise why work for the scrawny wages?? feed them peanuts and you'll get monkeys.

Or offer an interesting job and fair participation in profits and you get good people.
BTW, one of the mathematicians I work with worked for Alexander Sokol (I am sure you know who Sokol is).

But if I had insisted on the "obsolete" technology, her motivation would deteriorate.
Let's say you get a mega project to update a VC++ 6.0 app to VS2015? What do you do?

Sorry, cannot make out what do you mean with this question.

BTW what is your hourly rate?

"Ladies do not answer such questions because gentlemen do not ask them. "
(c) Meri Poppins
And my contract prohibits disclosing this info, so let's put it so: there is no such thing as MY hourly rate. There is a MARKET rate (which is a range, not a single number). With an exception of my first job after the graduation I was always above average but never on top because in Germany the have a socialism and are not used to reward top specialists with top salaries.That's why top guys and gals often move to Switzerland. I had no such opportunity before but now I do have it and I will make use of it.

// You probably missed the internet boom end 90's those were the big days for developers and consultants.

I remember it very well (I started earning with webprogramming in 1998)
However, those were big days for the developers in the 1st world countries only.
I was a good developer but first of all I was a Russian guy and it was hard to get contracts from abroad even for a ridiculous rate.

 

[Daniel Duffy]

Fair play, go for it! Best of luck.

The VC++/VS2015 remark is a bit similar to choosing work that is used by all companies compared to stuff that early adopters need. It's a mix of steady cash flow versus new ideas.

//
I know the name Sokol but am more familiar with the works of numerical analysts such as Marchuk (also head of Soviet Science), Smirnov, Il'in, Ladyzhenskaya, Janenko, D'jakonov et al as I did my PhD in Numerical Analysis based on their works. At the time I learned Russian in order to read their work. It was a golden age in Soviet Science. My supervisor was a friend of Marchuk.
What I created was exponential fitting for time-dependent PDE and I then applied it to Black Scholes.
http://www.datasimfinancial.com/UserFiles/articles/daniel3.pdf

Many results in the West from the 80s had already been discovered in CCCP 20 years earlier.

 

[sharan singh]

beginning is difficult for everyone ,later on it is easy to do.You know about your situations and enviroment of work.