After having owned an AI chatbot company for more than a year, I finally did what I was supposed to do before I even started it; I did a Competitor analysis - And o'boy should I have done it earlier.
Having been the sole software developer in what's arguably the most hyped industry since the dot com bubble, I naturally thought I had to hurry to create great products if I was ever to be able to compete with NASDAQ top 10 companies.
Holy macaroni was I wrong ð
I took some of our largest competitors, and I compared these side by side with our stuff, and in case you're in the TL;TR mindset at this point, they're all sjait!
The comparison
I basically went to the website of some of the self-proclaimed largest AI chatbot companies on earth, and I compared their products with my product. An objective and neutral conclusion, based upon facts, numbers, and technology, concludes with that we're a bajillion times better than everybody on everything that's possible to neutrally measure ...
- Intercom AI chatbot versus AINIRO
- HubSpot's AI chatbot versus AINIRO
- Zendesk AI chatbot verus AINIRO
- Five9 AI chatbot versus AINIRO
The above 4 companies are employing a total of something like 5,000 software developers, and I run around around them in circles, alone!
Of the 4 above "competitors" the only one that provides any value what so ever is HubSpot. But even HubSpot is 5 orders of magnitudes worse than ours. The rest of the above "alleged AI chatbots" doesn't even have any AI, and are similar to the stuff I did on my Oric 1 in 1982 using BASIC - I was 8 years old at the time ... ðĪŠ
All of the above articles have videos with side by side comparisons in case you don't trust my judgment ...
Why NASDAQ fails
It almost feels as if it's literally impossible for large companies to create great AI chatbots. HubSpot's chatbot, the only one providing any value at all in the above list, was in fact created by a single software developer in 3 months, arguably proving my point. HubSpot's CTO took 3 months off from his daily duties to create their chatbot, and he was the only one who at least to some extent was able to assemble something resembling a legitimate AI chatbot ...
It's almost as if creating an amazing AI chatbot seems to somehow be impossible for project teams with more than 1 software developer ...
Quarter of a pizza team comes to mind ... ð
Mythical Man Month
Let's play with an idea a little bit. Leonardo DaVinci was brilliant, and he painted Mona Lisa. If today's great painters are about 10% of his brilliance, basic math dictates that if you fill a room with 10 painters, they'll create another Mona Lisa for you, right?
The above argument is (ofc) rubbish, still 99.99999999999999999999% of every single CEO in the IT world will somehow believe it's the truth when it comes to software development. As in, throw 100+ brilliant software developers into a room, and the end result will be the sum of their individual brilliance.
The problem with that math, is that it's not a summation thing, it's a multiplication thing. Implying, if you've got one painter that's got 10% of DaVinci's brilliance, this becomes a multiplication factor of 0.1. Adding another who's got the same brilliance, results in 0.1 multiplied by 0.1, giving an end result of 0.01.
Congratulations, by adding another dev to your project, you've now reduced its combined brilliance by 90%. Instead of now having 20% of DaVinci's brilliance, you've got 1%
Adding 10 brilliant developers therefore results in 0.0000000001% of DaVinci's brilliance!
How I run my company
I do everything. Sales, marketing, software development, ranting on DEV, I'm AINIRO's social media manager, and I attend sales meetings. Yet still, I'm somehow able to find the time to outperform the entire top 100 NASDAQ index on software quality, in the thing they're throwing hundreds of billions of dollars at trying to solve. In case you don't understand why, read this entire paragraph once more, since it explains why ...
When I talk to customers and partners, they provide me with feature suggestions. We discuss these features back and forth, I ask them why they need the feature, at which point I'm often capable of completely eliminating the entire feature, by either showing them how they can already achieve the same with an existing feature they didn't know about - Or, by implementing another feature than the one they requested, that solves their problem, in addition to 1,000+ additional problems.
If I was to run my company "by committee", 99.9999999% of all valuable feedback would be lost on its way from the client, through my product manager, into my tech lead, to the end software developer actually responsible for implementing it. In case you don't understand the problem here ...
The client asks for a green button, they get flying dragons, with break lights and candy crush games, popping out of their ears, while singing happy birthday to you in Spanish ...
When I talk to the customer, I don't need to waste 2 hours explaining my dev what he or she needs to do. I know and understand the business reasons for the request, and the required time frame of 1 week to implement, stabilise and deploy, has now been reduced to 11 minutes of coding, executing 1,000+ unit tests, for then to trigger my CD pipeline ...
A company with more than one single employee simple cannot compete, not even in theory ...
For the record, I'm not proclaiming I'm Leonardo DaVinci, but neither are the guys at Zendesk or Intercom ð
And they all suffer from the "let's hire 100+ more devs to increase velocity syndrome" ...
The one that got away
There's one competitor I take seriously. These guys actually scares me a little bit. They're obviously brilliant, and the dev who created this thing, obviously has some skills.
However, paradoxically, I am 100% certain of that this company probably has a single developer like me, sitting in his room, doing whatever he wants to do, without having to be bothered with committees or "software development by democracy".
Now if I could only convince these guys to hire 10 more devs ... ð
Conclusion
Winston Churchill once said; "Democracy is a terrible form of governance, but it's less terrible than all other forms we've tried, so we're kind of stuck with it". My anarchist heart of course resonates with his observations, though with a different conclusion.
When it comes to software development, democracy-based software development, is not only terrible, but it's a "junkware guarantee". However, please don't tell Five9, Zendesk, and Intercom - Because as long as they believe they can throw another 100+ developers into their AI chatbot project to increase its velocity, I get to keep my little playground alone, and can take over the world and smoke these guys in the long run ... ð
Because, in the long run, there is one thing that matters, and one thing only - And that one thing is quality!!
Top comments (4)
The VC system builds in this need for excess people. People represent a proxy for "power" in their minds. Hire a few hundred and "look" big. Overhead too high and you haven't been able to flip the company quickly yet to some new sucker? You can always fire them if you have to. It's all about illusion. And, why, yes, when you chase the greater fool principle into IPO magic, the more people you have, the more valuable you "look" ;).
I remember a time when big blue bought a competitor and gave them 500 developers for 2 years - oddly enough, not much happened to the product... Bet they had a lot of meetings though.
Hehehehe :D
It doesn't matter how much science we get about this, they'll still mess it up over and over again ...
Just adding 6 to my team, that'll make us 20. Much more and we'll be running at 1/2 the speed of when we were 6 :) Now to convince everyone I've "right sized" and I don't need to double the team next budget round.