Silicon Valley hiring is not a meritocracy (2013)

An Internet search engine for the safe for work and legal Internet content, searches people want to do, and results they want to find where keywords/phrases work from poorly down to not at all.

So my work is for a part of search served at best poorly now. Thus, my work is not really in competition with anything else out there now.

Keywords/phrases often work very well, and there my work is rarely as good.

While the broad idea for the users and the UI are simple and intuitive, the business model, UI/UX, data from the user, database in the server farm, and the manipulations of all that data are wildly different from anything else in search.

Can read all you want on the Internet about various ways to build a search engine and still won't get even a hint about how my search engine works, not the UI and not the core server side software.

Broadly part of the idea is to give the user search results, that is, Internet content, that has a useful approximation to the meaning the user wants. So, the search engine is trying to get at or approximate meaning.

For this meaning, there's some math involved -- original with me, right, with some theorems and proofs, typed into Knuth's TeX, and with some advanced prerequisites. I especially thank A. Kolmogorov.

From my Ph.D. and published papers, I'm supposed to be able to do such research in math, and I have.

Of course the users will not be aware of anything mathematical -- the math is all buried in the deepest part of the server side software.

The key part of the whole project is just the math: If stare at the theorems and proofs of the math, then can get convinced that it should work quite well for its intended part of search.

Given the math, the software is routine and even next to trivial.

Maybe it is appropriate to call the project a case of information technology, but the software is so simple that it's not fair to say that it is a software project. Instead, plainly, it's a math project, with some original math, but might be called an applied math project.

The work is intended to do very well honoring user privacy.

The revenue is from just the most obvious source -- ads from ad networks. And I have some ideas for especially good ad targeting, still doing well on user privacy, I intend to implement later.

The core software has looked solid for a long time; it runs as just some Windows console applications that communicate via TCP/IP sockets and de/serialization of instances of a few simple Visual Basic classes.

The timings of the core software look quite good and much better with some selected use of solid state disks -- write rarely, read frequently.

There are good log files for the Web site and also the core software. Fairly simple processing of the log files should give a good view of what the users are doing and the performance of the server farm.

What's left to do is putting a Web site in front of that core software. The software for the Web pages is also nearly done: I should be able to get the last bugs out of the Web pages this weekend. The Web pages are so simple that they should present no performance challenges. E.g., the most complicated Web page sends for only about 400,000 bits.

Fixing the last bugs in the Web pages, then the alpha version of all the software should be done. Then I will give a critical review and do the beta version.

I wrote no prototype code; when the beta version looks good and is nicely free of bugs, that will be the production version 1.0.

The code I wrote is as solid as what a team I was on at IBM's Watson lab wrote and shipped as IBM Program Products, IBM's highest quality software product category.

My Web pages are so simple and small (exactly 800 pixels wide but a user could get by seeing only the left 300 pixels of width) that they should look good on any screen from a work station down to a smart phone.

As we know very well by now, Internet search is just super important, in the US and around the world. Well, a significant fraction of search is served at best poorly, and my guess is that nearly all Internet users will want a good search engine for that fraction and will be able to see that my work is by far the best for that fraction. So, that's the source of my confidence that my project might be successful.

If the project works at all well, then I shouldn't have any trouble, say, buying a new car!

While I'm a solo founder and did all the work, on this project I'm an entrepreneur, not a programmer. If this project is successful, then I will want to hire some people for some routine programming, and I will train them and, then, let the first such batch set up a training program, e.g., with video lectures, for later programmers. But looking at the job ads, it's clear I'd never get hired as a programmer.

Moreover, the venture capital people and YC want nothing to do with me and won't until I have traction. But for the amount of traction they want, I would have revenue enough not to need or be willing to accept equity funding. No, I don't have four kids and a pregnant wife.

Also I'm terrified of the idea of reporting to a BoD. I just can't think of a realistic BoD that could understand the core work. When at an eary Board meeting I gave the math for the ad targeting to be done, most of the Board members would get so upset they'd leave a smelly trail all the way to the little boy's or little girl's room. I'd have hose off and totally redecorate the Board room and everything on the path to the little rooms. They'd try to execute their fiduciary responsibilities and, too soon, totally blow their fuses.

One of the more unpleasant experiences in life is listening to a math lecture with theorems and proofs where really do want and need to understand but don't have even 1% of the essential prerequisites.

My guess is that any usual startup BoD would have my company totally dead in less than 18 months. They might try to save the company by, say, selling cupcakes on-line!

So, I am 100% owner and should stay that way.

It's been noted more than once in recent months that it's now possible for a solo entrepreneur to bring up a successful Internet based business. Actually, so far I've seen no reason to hire people. Instead, it's my business and up to me to make it go.

Nah, I meant the second person plural.

We should have stayed with thou/ye instead of you/you. It was more clear.

I would be absolutely fascinated to see whether you can build a search engine in Visual Basic!

It seems you have a very pragmatic and humanistic mindset, but at the same time all the conventional wisdom of software development tells me you can't run a successful startup on VB. Good luck and I look forward to learning how it turns out! Hopefully you'll smash my preconceptions :)

I love the term "HR Pros". Because everyone else is a "resource" but they don't like applying the term on themselves.

I agree with you it doesn't work this way, and my wife would say "that's why you don't get good employees." And she'd be right. It doesn't take half a day to give a short 20-30 minute skill assessment. And if you're going to pay this person $50k and up a year, both of you should find the time to do it. The cost of onboarding a new employee can be tens of thousands of dollars in salary and training costs before they're an effective employee. If you have 5 people on your short list, you can spend an entire day adequately testing them and still come out ahead.