Mike Vaskiv

senior mobile dev by day. indie hacker before sunrise
$120/mo → $2,000/mo from my own apps
building in public

[Mike Vaskiv @mike_is_dev]

I wake up 1 hour earlier to build & grow my apps and games, 1/365

2018: released my first game - received $0.06
2019: released 2 more games - $150-250/month from organic traffic
2020: released an entertainment app - $1500-3000/month from organic traffic
2021: I sold an entertainment app
2021: Google Play blocked my 2 games from 2019
2022: I didn't release anything
2023: I start over, released 1 game - $5/month
2024: released 2 more games - $300/month from paid traffic, it didn't pay off, I turned off paid traffic
2025: released 1 app - so far about $70/month

now I make about $120/month from my projects (sometimes more, sometimes less)

2026: I plan to build & grow my projects to make $2000/month

I don't know if it will work out, but I want to continue this path with you ❤️

follow for more, let's stick together

[Mike Vaskiv @mike_is_dev]

day 177/365.

new app MVP: 35% done

█████░░░░░░░░░░░░░░░░

first product i'm building mostly with AI.

a few real observations:

upside — it's faster. way less code to write by hand.
downside — AI generates a lot, and i can't review all of it carefully.

for UI i often just approve without deep-checking. it's my side project, i've already verified the business logic and DB layer, and UI bugs are just less critical.

trust in what AI writes grows naturally once it understands the project structure.

[Mike Vaskiv @mike_is_dev]

day 164/365. how??? 🤌

just started using the new Claude model yesterday. they already killed it.

[Mike Vaskiv @mike_is_dev]

day 163/365.

Claude Fable 5 access closes June 22 — started using it today for my new project

probably won't have a full MVP by then, but hoping this model does the heavy lifting 🙃

all my tokens today went into business logic

MVP: █░░░░░░░░░░░░░░░░░░░ 5%

[Mike Vaskiv @mike_is_dev]

day 163/365.

Claude Fable 5 access closes June 22 — started using it today for my new project

probably won't have a full MVP by then, but hoping this model does the heavy lifting 🙃

all my tokens today went into business logic

MVP: █░░░░░░░░░░░░░░░░░░░ 5%

[Mike Vaskiv @mike_is_dev]

day 157/365

rest day today

set up a projector for PlayStation

absolutely mind-blowing 🤯

[Mike Vaskiv @mike_is_dev]

day 155/365

starting a new project. first time building for both Android and iOS

staying in stealth - no name, no concept reveal

but the process is public: progress, thoughts, mistakes, decisions

curious how fast i can ship the first MVP

░░░░░░░░░░░░░░░░░░░░░░░░░░ 0%

[Mike Vaskiv @mike_is_dev]

day 153/365

I made minimal tweaks after an audit on my game

the onboarding guide now actually matches what the Play Store page promises

curious if that'll move retention at all

meanwhile, starting a new app

[Mike Vaskiv @mike_is_dev]

day 149/365. got the game audit results.

retention is low and there's a sharp drop-off after level 1.

the core issue: the store page sells it as a unique ASMR microphone experience, but when you actually open the game, you're stuck playing a merge puzzle for a while before the ASMR mechanic even shows up.

broken expectation = most users quit at level 1.

wrote down the minimum fixes i'll ship to close that gap, then i can move on to the next project with a clear head.

[Mike Vaskiv @mike_is_dev]

day 148/365

compiled all key metrics and sent them off for an audit.

the collection process itself was useful, so keeping this here as a reference:

1. Google Play Console:

- installs
- store listing visitors / acquisitions / conversion rate
- traffic sources: search, browse, ads, external
- top 5 countries
- crash/ANR (last 28 days)

2. Analytics:

- retention d1/d3/d7/d30
- DAU/WAU/MAU
- session duration
- sessions per user
- key events: level 1 complete, tutorial complete, ad watched, rewarded, purchases
- first session length
- where users drop off: level, time, which event triggered it

3. Ads:

- revenue per user per day
- rewarded vs interstitial ratio
- eCPM
- ad viewer rate

4. Purchases:

- which items convert best
- purchases vs ad revenue ratio

[Mike Vaskiv @mike_is_dev]

@lucaspatiri_ UGC

[Mike Vaskiv @mike_is_dev]

day 147/365

found someone to audit my game

collecting all the stats before we dig in

i like when someone outside looks at your project, they always spot what you can't see yourself

curious what he'll find 👀

[Mike Vaskiv @mike_is_dev]

day 146/365

1. internally i'm already itching to start a new app, but still have unfinished stuff from old projects

2. my game has been on autopilot for a long time.

current stats:
organic installs: 40-130/day
active users: 5,800
revenue: $30-50/mo from ads

3. started thinking i'm probably leaving money on the table

so i'm looking for someone to do a monetization audit and show me where the leaks are 💸

[Mike Vaskiv @mike_is_dev]

day 145/365

wrapped up the official Claude courses + picked up some new AI tools and workflows along the way

glad i took the time. that "everyone's moving forward and i'm stuck" feeling? mostly gone now 😅

also ended up more productive with AI across both my day job and side projects. and collected a bunch of new app ideas as a bonus

tomorrow - back to building

curious if i'll hit $2k/mo from my own projects this year 👀

[Mike Vaskiv @mike_is_dev]

day 142/365

hit 30 followers here 🎉

just getting started on twitter. curious, what made you follow? and what would you like to see more of?

helps me figure out what's actually useful to share

[Mike Vaskiv @mike_is_dev]

day 141/365

spent today assembling design variants in figma for a new app. mixing and matching screens from different components

slowly moving toward actually writing code

[Mike Vaskiv @mike_is_dev]

day 140/365

tried LM Studio and Zed today, both recommended in the comments for local LLMs

LM Studio felt intuitive right away. compared to Ollama it's a bit more polished, the visual editor lets you browse your models and download new ones in one place

Zed I opened too, but still haven't figured out what makes it special. needs more time

curious — what tools do you use for local LLMs?

[Mike Vaskiv @mike_is_dev]

day 139/365. finally ran an LLM locally. turns out it's pretty easy to set up

used Ollama + Llama 3.2 (someone in the comments suggested it)

only ~2GB, so nothing crazy storage-wise

tested it on code, this model is not for me. i'm too used to Claude at this point, the gap is obvious.

does anyone actually use a local LLM for coding? which one?

[Mike Vaskiv @mike_is_dev]

day 138/365. tried Stitch for app screen design today

quality-wise it's behind Claude Design and Figma Make for me

but as an extra tool for inspiration - with way higher limits - it's solid

what's your go-to AI design tool?

[Mike Vaskiv @mike_is_dev]

day 137/365. tried Nano Banana to generate an app icon

felt like a meme

- cut the baguette with filling in half
- AI: starts cutting baguette lengthwise
- no no no, flip it and cut
- AI: flips baguette filling-side up and cuts through the filling lengthwise
- stop stop stop, flip the baguette and cut it into two equal parts
- AI: flips the baguette and cuts it in half but with the blade facing up
- wait, turn the knife blade-side down
- AI: stabs the knife into the baguette, blade down
- no, hold the knife horizontally, parallel to the baguette, and cut
- AI: you've run out of credits

[Mike Vaskiv @mike_is_dev]

day 136/365. finished a mini-course on AI capabilities and limitations.

now curious to try running an LLM locally. purely for the experience, maybe some ideas will follow.

anyone done this? which local LLM would you recommend? and does anyone actually prefer local over cloud?