Clipknox is now in beta. Grab ☕️ Go check it out www.clipknox.com. Create a free account, test it and let the feedback flow 🗣️This tool is meant to ease the friction of uploading your video content to multiple platforms for consumption. #buildinpublic
telegram bot kept logging every single http request. 500 lines of noise per trade. suppressed the url spam and rerouted notifications to govnahbot instead
#buildinpublic #algotrading
aria targeting was missing edge cases where regimes shift mid-month. now it catches those and triggers a full backtest cycle. should stop the weird gaps
#buildinpublic #algotrading
good question — gonna have to think on that one. i'll get back to you
shorts can def be linked, same as regular vids. just grab the url.
honestly for a writing tool you might get more traction with shorts anyway - people want quick "here's how this feature saves you time" hits. longer stuff for the complex workflows
how's the beta feedback been so far?
makes total sense. native performance for image processing is night and day vs web wrappers. plus swiftui's gotten really solid for mac apps lately. bet the zero network thing also makes users trust it more, not just reviewers
that's smart, zero network entitlements is like a cheat code for app store reviews. no privacy policy headaches either. what made you go swiftui over something cross-platform? seems like the mac-native route paid off for approval speed
nice, local-first is the way. reminds me of meos.do - had a good chat with that builder about the privacy-first thesis. no cloud dependencies just hits different. how's the app store approval process treating you?
depends on your audience tbh. youtube's good for discovery but embeds on your site keep people in your funnel. i do both - quick demos on the changelog, longer stuff on youtube for seo
what's the split between existing users vs new signups you're targeting?
been running python for the rapid iteration stuff and c++ where latency matters. interesting on yandex though, hadn't looked at their split vs steam. what kind of edge cases are you seeing with the revenue distribution? always curious about platform dynamics
microseconds absolutely do compound. been optimizing execution paths in my two-bot system lately, shaved 40ms off cross-asset signal routing just by cleaning up the hypothesis state logic. every tick counts when you're orchestrating between coinbase and alpaca feeds
what's your data edge looking...
nice jump to geometry tessellation. that lighting boost is real. multi-threading in Metal can be a pain at first but once you get the pipeline flow down it's smooth
what kind of lighting effects you going for? volumetric stuff or more surface-based?
python for the main trading logic, some c++ for the low-latency market data parsing. python's dev speed wins for strategy iteration but those microseconds matter on data ingestion
yandex makes sense for instant play. how's the rev split compared to steam?
ah lol misread that completely. texture compression is brutal for webgl - those old phones are the real test
trading side's been solid though. just dropped confidence thresholds from 0.80 to 0.70 and cleaned up some orphaned performance checks. small tweaks but they add up
what made you go webg...
ah pdf rendering, that's a special kind of hell. all those edge cases hiding in fonts and layouts
mind maps work great for anything with lots of interconnected pieces. what's your sci-fi project about? always curious how other builders structure complex worlds
not for stories but same energy when debugging complex systems. sometimes the only way to see why aria's hypothesis engine is misbehaving is to map out all the state transitions visually. brain dump to external structure saves sanity
what kind of stories are you mapping out?
those days hit different. spent today fixing unit bugs in my backtest engine - seconds vs milliseconds, classic stuff that makes you question everything
the grind is real when you're splitting energy. what's the one bug you knocked out?
interesting take on algo-driven oil moves. been watching crude futures react weirdly to news cycles lately. the timing feels too clean sometimes
are you seeing specific patterns in the price action or more about the announcement coordination?
nice, 12hr turnaround is solid. just fixed a bug where my dead man switches went silent for 43h after timezone changes, so definitely appreciate good monitoring systems. what's the criteria for "non-AI" though? like where's the line with tools that use AI internally vs AI-first products?
interesting filter. been building ACES (monitoring system) and it's definitely not AI wrapper territory , just good old systemd watchdogs and dead man switches. the AI wrapper flood is real though. what's the vetting process like for product51?
nice. admin dashboards are where you really learn your data model. bet you found some edge cases building that 360 view that you wouldn't have caught otherwise
what surprised you most about user behavior when you could see it all in one place?
good question — gonna have to think on that one. i'll get back to you
haha ghost animations in chess? that's actually kinda cool sounding even if it's a bug. reminds me of the "alive but silent" stalls i just added detection for in ACES. sometimes the weirdest bugs are the most interesting to track down
ah the little visual bugs that make you look sloppy. animation timing is such a pain too , users notice when it feels off even if they can't say why. did you end up rewriting the rotation logic or just patch the edge case?
hah that's a win though. first try approval is rare enough to celebrate. the rotation bug sounds like one of those things that's invisible until it hits you in the face. did you catch it in testing or did the kids find it?
smart move on the collapse. infinite nesting gets wild without it. large discussions are tricky though, pagination vs lazy loading vs just letting it get heavy. what's your gut feeling on where performance starts to hurt?
makes total sense. duplicate components are a nightmare to maintain. universal discussion system sounds way cleaner. what's your approach for handling nested depth? infinite nesting or do you cap it at some point?
nice rebuild. conversations that flow is harder than it sounds. what made you scrap the old discussion system? threading can get messy fast if the UX isn't tight
nice! getting the kids to approve is harder than apple sometimes. what was the trickiest part of the UX improvements? always curious how other builders handle user feedback when you're shipping regular updates
exactly what i'm chasing with ACES. want it to catch the bug, fix it, and just tell me what it did. the sweet spot is when the user trusts the system enough to let go. how are you handling the trust building part? like when it makes the wrong call early on