The root of it is sales. Sales is getting your product in front of someone who needs it. Most likely that person is building a similar product themselves, but unable to do it as well as you.
Problem 1: Pitch. You have a product they want, but they don't know it. Learn to explain the product clearly. Get it down in 5 seconds. Make the title, color, shapes like other similar solutions. If it attracts their eye, they'll start reading it.
Problem 2: Sales channel. You have a good pitch, but it's not in front of the people who actually want it. Don't just market it on Show HN/reddit/Product Hunt. Get it to where the people are.
Marketing is a loop experimenting and iterating around those two things.
This is assuming you have the right product. If you don't, then all you can do is fake it. People often copy the fakers because there's a lot of them, but as a developer, you can also iterate on products easily and shouldn't need to fake anything.
Once you have that covered, you attack problem 3: Getting market share up to 100%. Most of those people just kinda want it, won't pay full price for it, need dark mode, and so on. If you can solve this, congrats, you're rich.
Congrats. This sounds like a great way to build a "wrapper" product. Gather clean data on a particular niche and keep it in one place. Have something that's good at processing that data. No matter what "startup killers" the big boys release, it's still resilient and valuable.
Opus 4.5 has excellent tool use, meaning it can jump in and out of a broad undocumented codebase better. It can evaluate what the code is trying to do. It's perfect for PRs - caught things like people submitting code that looks right, but ended up running a poorly documented/incomplete method.
GPT codex just messes up a lot for me. Whatever I'm doing with it, it's not working. The plain GPT-5.2 is good overall, but it confidently makes mistakes and tell you that it's done.
If you have an excellent codebase, GPT 5.2 might actually work better. If you're not sure what you're doing or are using AI to find out how things work, then Opus 4.5 is great.
The Claude models are also very much behind in terms of UI and visuals.
Take note that a lot of the benchmarks are on Python. What I'm finding is all the major ones make mistakes, but they make mistakes differently. OpenAI and Anthropic tend to mimic one another for some reason, while Grok and Gemini tend to give very different answers.
Likely a good deal of test coverage. At the far end of this is something like Facebook, which has everything monitored by A/B tests. If it breaks something that changes something serious, the alarm triggers. Move fast break things isn't a new way of doing things, so might as well pick up a framework that works.
There's an amazing, gamified app out there that uses standard learning optimization patterns. It's called Duolingo. Unfortunately, you can have a multi-year streak on it and still not speak the language. It is effectively a language-themed quiz game, and not a language-learning game.
What you're doing is fundamentally at odds with the market, but perhaps there are people who want to use an app where they learn language.
Health data is usually the highest level protected data under most laws. It's not about just insurance. Part of the problem is once data is out there, it can be used by any shady person.
You can be discriminated against a job based on health records. Scary diseases like AIDS and TB make it hard for unskilled labor to land a job since it's so easy to discriminate. Pregnancy history may hurt women who are in countries with more generous maternity leave.
Mental health history will hurt just about everyone - who wants a worker who can claim ADHD, depression, anxiety, etc as reasons to be unproductive?
Then people will simply deny getting diagnosed for fear that they may uncover something that puts their jobs at risk. That hurts the medical system as a whole.
Combine with weird stuff like eugenics. What if we identify a possible rapist gene and neuter them in advance? Or bar people with a klepto gene from working in finance? You may live in happy, sane, democratic societies today, but it may not be the case 30 years from now.
I once had to threaten to resign from a job (in the EU, pre-Data Protection Act) over the data handling of evidence of one of the things in your comment.
I think this has been something people have had an instinct about forever, and the only reason I had to threaten to quit was because of a misunderstanding of the level of data safety involved; put simply it was not common knowledge that socket connections could be snooped and that targeting a popular service would be easy for a malicious person to do. (This was before SSL was efficient or easy to manage, and in the days when only payment screens were encrypted).
Once the message was across, everyone's objectives were aligned again.
Health information is deeply private because disease is entangled with shame/weakness/vulnerability/taboo/intimacy.
> You can be discriminated against a job based on health records.
Just to make this clear, probably EU-wide, you can't legally be discriminated against. However, it's gonna be hard to prove leaked data won't be illegally integrated in e.g. ATS models, or was attributed as skill issue when it popped up during manual background checks.
Although, infectious disease like HIV or dystopian scenarios like eugenics are probably the classical discrimination examples for these privacy implications, I don't think they are very likely to be discriminated against (outside of jobs where discrimination is legal and require disclosure anyway, e.g. health workers, food industry etc.). It's easy to dismiss those worries, since most people aren't affected. But common issues with mental health (e.g. depression), hidden disabilities and chronic disease (e.g. PMS), or potentially severe recurring disease (e.g. cancer) realistically are going to be much more impactful. Everything which statistically increases chances to fall out the work force due to health reasons - especially in combination with strong labor protections.
Whatever you pick, don't blame your kids for missing out on your career. A lot of parents did that and honestly, the kids would rather that they tried than become the cynical, grumpy old men they turned into later.
Commit to whatever path you picked. Kids will be fine. I grew up with two overworking parents and I'm fine. My kids have both parents at home, homecooked meals, family dinners almost every night, and they're fine too.
Nobody resents their parents for chasing their dreams unless it's just pure abandonment. A generation ago, a father might end up working in a whole other country doing hard labor just because it paid the bills.
Part of what shaped this opinion was my father always asking me to build and sell a billion dollar company. He put aside that dream so he could give me the education to. The actual expectations on you are between you and your extended family.
I may be wrong on this, but I feel like they're different things - a daimon/daemon is simply a spirit, feeling, something that does something. AI is a good example. It's there. It's influential like the wind, but it's also not a proper intact form.
A CS nerd analogy here might be functional vs object - one transforms, the other holds more of a form.
The Christian demon seems a cultural appropriation of it, especially during that era where they were taking a lot of classical philosophical Greek terms and just Christianizing them.
There's a game, played by certain rules. One way to win the game is to adopt a persona or alter ego, which acts as a form of sandbox/VM.
Those who play the game well get money. The money attracts players from all over the world. More money attracts more competitive players. A $300k salary and some artificial rules like interview mastery attracts more gamey types.
There are places that don't play by those rules. There will still be some gamification - for example, the other rules may reward those who share knowledge, are polite, honest, down to earth, and so on. They may still be performative, but it looks less like one.
There's a reason hackers go around in t-shirts and uncombed hair, and it's performative in itself.
Problem 1: Pitch. You have a product they want, but they don't know it. Learn to explain the product clearly. Get it down in 5 seconds. Make the title, color, shapes like other similar solutions. If it attracts their eye, they'll start reading it.
Problem 2: Sales channel. You have a good pitch, but it's not in front of the people who actually want it. Don't just market it on Show HN/reddit/Product Hunt. Get it to where the people are.
Marketing is a loop experimenting and iterating around those two things.
This is assuming you have the right product. If you don't, then all you can do is fake it. People often copy the fakers because there's a lot of them, but as a developer, you can also iterate on products easily and shouldn't need to fake anything.
Once you have that covered, you attack problem 3: Getting market share up to 100%. Most of those people just kinda want it, won't pay full price for it, need dark mode, and so on. If you can solve this, congrats, you're rich.