I was expecting: PHP + MySQL + a bunch of shell scripts to install stuff on servers + reading logs via command line by ssh as root. Instead I found: Redis, Rabbitmq, DataDog, ES, React + Redux, Ansible, PagerDuty... I wouldn't call that "boring technology". Sure, you are not using Docker nor K8s... but still.
A sqlite db is just a file. So at that level it is exactly like a spreadsheet. It's just a much more powerful storage format. But for the app opening it it can be presented much like a spreadsheet.
I'm currently reading logs via command line by ssh as you describe for a simple service and not satisfied with this. What would be a step up without a lot of extra infrastructure?
I’ve personally used Graylog[1] with success in the past.
However, I’ve had an excellent time with Grafana and have been following their Loki[2] project.
My company uses other solutions, so I haven’t needed it, but the Grafana stack might suit your use case.
Run vector [1] on each host to collect logs and send them to New Relic's free tier.
No need to manage extra infrastructure for logging, New Relic works pretty well and also has alerts, and Vector makes it easy to move to a new provider or self-hosted solution without reinstalling agents: just add a new sink to your config file.
PHP has evolved, it less of an injection vulnerability factory now, and MariaDB is the new MySQL. As usual computers became cheaper and more powerful, maybe not as much as we could have hoped in 2010 but SSDs made up for the slow rate of progress in CPUs.
All that to say that PHP + MySQL/MariaDB has evolved will get you much further than it did in 2010 without losing its simplicity. Current trends are backed by tech giants working with millions of servers and billions of users, and they push for technologies that fit their use cases. They goal is not to make a one man shop.
> Most of my time is spent on talking to other human beings, replying emails (30%~40% of my time), and thinking (!!!), which is not considered as “real work” by engineers :)
listennotes has become my go to site when I want to find podcast for my daily commute (about 30min). Its got good UI/UX. Pretty well executed for a one man company.
I wonder though if the niche search engine market can fetch them enough money to be profitable.
Regarding the profitability, the founder discusses in other posts, that they get the bulk of their revenue from their API (they also have some cash flow coming from ads). I think he was for a while bootstrapping, and then got some VC funding. Still he pays for a not that cheap office at a WeWork location in SF and from the API service numbers I saw, if they are legit, he should have more than enough ARR to live well from his site (and that's exactly the impression one gets from reading his blog posts). Nonetheless, he is totally against posting openly about the financials of his business, and showing actual data.
Great article, but I think he's far too hard on himself and reading too deep into his rejection (plenty of immigrants, non-native English speakers and people in their 30's from non-Ivy League schools succeed at building businesses). He's built a bootstrapped site that is probably going to be able to sustain himself for years.
But small businesses and big businesses are fundamentally different. I come from the restaurant space, and all my favourite restaurants are 'small' and likely can't exist in any place other than where they are. They can't scale. But they're much better qualitatively than restaurants that can scale. Some of the other products and things in my life that I enjoy also come from small artisans and could never scale. And that's OK.
He didn't build a massively scalable business but he did build a (seemingly) good one.
I think his earlier expectation that without a greencard he was summarily rejected might be true. It's a lot of risk to assume of "the founder/entire team might be deported".
Other than that, I think you are correct he was reading too much into his demographics/education.
Idk if he is even hard on himself. He actually sounds like someone happy with how the site ended up evolving. And maybe even looking back in hindsight to the YC experience, he realized that they might have not been a good fit anyways.
I mean, he says it quite bluntly, VCs are in the business to do money, they don't finance you as a charitable exercise or because they like tech for the sake of tech. He found a nice ARR through the API service, and if that sustains him well, and as he says, does not cause him much stress nor makes him work long hours, it's probably an even better outcome that being pressured by the VC to fulfill some exorbitant growth.
To me it sounds really cool to be a one-man band, and his posts describing the tech stack are really delightful to read (he's definitely someone who knows what he is doing, I guess that comes with being 'old' and having previous 'not-serial-entrepeneur' tech experience as a developer.
Obviously, things have changed quite a bit in the three years since this was written. I'm curious what we now consider boring that was "exciting" then. First thing that comes to mind are the managed kubernetes services that weren't really "mature" back then. Are those boring? Some would say yes, I personally would say "not yet".
I wouldn't consider managed kubernetes services as boring. But I think nowadays more people accept that Kubernetes is an overpowered solution for most projects, especially side projects.
Interesting post to see how this solution is accomplished, but how is this a one-person company, if there are contractors? How are the contractors utilized and how much do they cost on average? Are the contractors needed or could this be done by a single engineer?
Would you consider someone delegating tasks to a personal assistant on a billable hours basis a multi-person company? How about paying a freelancer once for the initial design of the website? Both these tactics are discussed in "Start Small, Stay Small" a book aimed at single developer software product "startups."
It's pretty clear that if you want to shorten time to market, you leverage existing assets - when it's people using React or FLOSS databases, no one bats an eye, but farming out tasks here and there that aren't worth your burn rate (pay someone cheaper who can do it faster), how is that different?
When the government hires contractors they're not considered part of the government workforce (ie, civil servants).
A single engineer owner with multiple contractors is different from multiple employees under a single engineer owner in what way? It is dishonest to claim that one is “one-person” while another is not when they are organizationally the same.
Reminds me of people that call themselves “retired early” when they are effectively entrepreneurs with their own businesses and side hustles.
Kudos to the original poster though for making a business that they own by themselves I assume. However, taking all the credit for themselves despite using contractors to build their company is disingenuous.
"Hey, look at this guy over here, claiming to be a 'one-person company', when he pays for a VPS instead of constructing a server from hardware, then self-hosting it himself. Pfft, I'll bet he doesn't even personally service the AC that is used to cool the building the server is in either! Probably pays an ISP too, instead of creating his own Internet from scratch."
I think comparing the use of contractors to the use of cloud services is pretty disingenuous in this context.
The reason that "one-person company" is emphasized in the title is because people will expect to find a list of technologies that make developing and maintaining all of the tech required to run a company do-able for a single person.
Advice about which cloud services to use is exactly what I would expect to find on such a list. Hiring external developers is exactly the opposite of what I would expect to see, and kind of undermines the premise.
UI design is for an artist, not a developer. I can build whatever someone wants with the same stack that the OP is using (Django, Postgres, Lucene-ish search db, etc), but I will freely admit to people that I suck at website design in the way that they think of website design (what it looks like). Bootstrap was made for dolts like me who don't know what buttons and forms are supposed to look like.
Is it cheating to hire an artist to design the front facing site components?
It depends. If the contractors are working ad-hoc or part-time, IMO it's reasonable to say there is one person working full-time on this company, therefore it's a one-person company.
I think of it this way: if I'm solo and I hire a CPA to do my taxes or a bookkeeper to balance the books, does that make me a two-person company? That doesn't feel like an accurate description from my perspective.
I ran my company as a single founder with nine contractors working from a few hours a month to maybe 15 hours a week. All little task-based projects like loading up new content from a repo or answer support emails. Technically, letter of the law, we were 10 people. But from a cost/profit perspective (which I would posit is the most important aspect) it was very much a one-person company.
It sounds like he's using contractors for "surge" capacity. Normally, he's the only employee. But, if he has a major set of features to build/release, he'll contract out some of the work. Assuming they're all short-term jobs, I think it's fair to continue claiming it's a one-man company. But, if there's a steady flow of contractors that roughly adds up to a single FTE, the he's stretching the truth.
