Problems I Live

i wanna make something thats personalized per used and when shot a vid of showing hard proof, people should be saying wow, building an agent / api can be smart too, i just want to make it a bit more data centric if selling to others, tho im not selling to indie hackers nor schoolers cuz theyre broke, i dont wanna build a dashboard cuz those will die, ai is becoming the core layer to most people and we need to make something ais can run only if we’re not going for pure b2c which i would actually love btw, i love working out and i know a fuckload about it and people’s problems, and i know some friends etc who would genuinly use my app to do stuff if i gave them access, so that can be a way too, also goes viral easily, or an agent layer, not sure

• ai context problem, what ryva tried to surface as well, but in a different way
• making agents more friendly for non technical people, the setup of openclaw etc are complex and most people dont know about it, can be marketed in social media platforms to gain traction
• tracking calories is still not viable, cal ai and myfitnesspal exists but its hard to track sauces etc
• homework, school work etc too long something that automates these - tho most students are broke so they wouldn’t pay for it
• marketing sucks, something that can actually make people gain 10x more tration with proof, i just need it to be automated and codex automations etc suck, sales suck as well, a cmo agent could be smart
• tracking workouts is too manuel, would be better if more automated
• scheduling stuff is still boring and actually sticking to those routines is hard
• something thats truly agentic like a operating system but something agents can read entirely like ibx but in terminal etc as well would be good
• ai limits hit too fast, would be great if we can somehow reduce the costs / reduce context usage beyond basic skill files
• making 3d websites is painful, most ai website generators cant make them, something that can generate 3d models / websites like igloo.inc would be cool
• there is website and app makers but high quality game maker ai space is untouched, could blow
• ai cant write like me and forgets everything all the time
• learning to code is still very hard and courses not enough, also most courses still just teach normal coding in the agentic era this is a bit dumb
• something just knows and remembers everything thats also easy to setup would be amazing, similar to egeuysal.com/agents.json but not a web app, maybe a cli + chat sdk integration so you can chat and ask questions in discord slack etc, as well as others. just like an agent as a service
• threads open, no closure no “who do I follow up with today” losing warm leads
• publishing blogs, uploading images, generating og images etc take time, most solutions are incomplete despite p CLI
• School + startup context switching kills deep work
• “Motion vs. progress” is a recurring mental bug, Almost every diary entry has some variant of: “did a lot, but did it actually move the needle?”
• editing videos that look good take too much time
• Outreach context switching is brutal
• You write a diary entry, then manually adapt it to X, LinkedIn, Reddit, and your blog. You reduced image friction with p but the prose distribution is still manual. Your blog.json and diary.json feeds are machine-readable but nothing reads them automatically to draft channel-specific repurposed content.
• What it is: A multiplayer IBX. Teammates dump thoughts/blockers/todos in natural language. The AI subagent expands, prioritizes, and assigns them. No Jira config, no project hierarchies — just a shared execution queue. Deep Slack integration so messages can become tasks without leaving the thread.
• • You re-explain yourself to every agent you run. Each session starts from zero.
• Skill files have no home. Finding good ones, sharing them, knowing what exists — there’s no place for any of that.
• You built Brain but it only works for you. Nobody else has it. They’re still doing ICP research manually.
• The white-glove run doesn’t scale. It’s your best move and your biggest bottleneck at the same time.
• Warm threads die and you don’t know which ones until they’re already cold.
• Your strategy drifts and you don’t catch it. Your actions stop matching your stated direction and nobody flags it.
• Publishing is still split across too many steps. The content exists once. The work happens four times.
• The standup problem has no good solution. Tools that exist just move the same bad question to a different channel.
• You still don’t have a fast way to know if a problem is daily pain or monthly annoyance before you build.
• What moves the needle today? You answer this every night and start from scratch every morning.
• Your open source repos are a complete system that requires significant setup. The gap between the code and a running system is the entire product.
• ICP buying signals disappear into feed noise within hours and are structured nowhere.
• You spent 2 months building 70k lines of Go infrastructure before talking to a single user.
• You built Huesly for a problem that happened once a month. You had no fast way to test frequency before building.
• You chased 100k Reddit impressions and got 4 signups. Reach and conversion are completely different signals and you had no tool that distinguished them.
• You spread across 5 channels simultaneously and owned none of them. Dilution looked like leverage until it didn’t.
• You were pitching before listening. Every first-wave outreach was a mini landing page nobody asked for.
• You built your own AI API because existing ones had rate limits and costs you couldn’t predict. Inference pricing is still a surprise every month.
• Your first hackathon project was better than 90% of submissions and you still missed the deadline because of deployment chaos.
• You learned on an iPad with no computer, SSHing into a VPS because there was no other way. Barrier to entry for real development is still broken for people without the right hardware.
• You over-engineered your own productivity system so badly it became the thing you were procrastinating with. The system became the work.
• You were using AI to generate entire projects early on and learning nothing. There’s no signal between “AI wrote it” and “you understand it.”
• Slack decisions disappear. PR context expires. Standup reconstructs state that should already be visible. You’ve written about this from 6 different angles and it’s still not solved at the tool level.
• You have 25+ deployed apps and no single place that shows what they are, their status, whether they’re still running, or when they were last touched.
• Agent context burns on long sessions. You built the caveman skill to compress it. Everyone else is hitting the same wall and solving it from scratch.
• Your school internet is unreliable and most tools assume connectivity. Offline-first is still an afterthought for developer tooling.
• You say no to family and friends to protect execution time but have no systematic way to know whether the trade was worth it on any given day.