The journey can be really tough. But engaging with your community and understanding what's ahead keeps you grounded and reduces anxiety caused by unnecessary FOMO. Keep your head down building and solving hard problems. We can all win.

Find an important unsolved problem you care about.

Then use AI to solve it. Go deep! Talk to people. Build a community.

It might take you months or years, but always know that AI capabilities will only keep improving.

Build for now and for the future.

@omarsar0 Everyone’s saying this, but nobody’s actually shipping anything that lasts.

@omarsar0 100% I've find custom benchmarks and sle fimproving agents as a topic worth investing years into because it will pay off at the end

@ChainZenit @omarsar0 We’re moving past the flashy prototype phase into the real workflows. Exciting times.

@omarsar0 ask yourself, would you still build it if AI stopped improving tomorrow? if yes, you've probably found the right one.

@omarsar0

@omarsar0 this is basically the only strategy that makes sense now

adapt or get left behind basically

@omarsar0 Excellent advice 🎯

@omarsar0 good advice but "go deep, talk to people, build community" is 90% of the work idk why people skip it

@omarsar0 building unified infrastructure for AI work history and looking to bring together serious Claude Code and Codex power users.

@tonyspiro @ChainZenit Exciting times indeed. There are so many unsolved problems to tackle.

@OskarKohler great way to put it too. but i won't bet on ai not improving. i have sort of seen a good percentage of what's ahead and it's going to get pretty incredible

@omarsar0 problem is most people dont know what they care about yet

ai wont fix that part for u either

@TomasMann1878 i am all in on custom evals. we should all be building these for our domains

@omarsar0 the best time to start was 2 years ago. the second best time is now. but the worst time is "when AI gets good enough," because by then everyone else is building it too

@omarsar0 Wise words said, fully agreed.

@omarsar0 the deep part isn't the ai anymore. the ai is just the routing layer. the deep part is finding the proprietary data that actually makes the routing layer worth using

@omarsar0 This is the right frame. The problem you care about matters more than the tool, because caring is what gets you through the months when it's hard. AI keeps improving, so the gap you can't close today might close next year. Pick the problem first. Let the capabilities catch up.

@omarsar0 cool, but don’t ignore the current users