The worst part was not knowing what caused it. You upload something, it seems fine for a while, and then it disappears or gets buried. I spent days editing videos that ended up dead in the water because of things I couldn\u2019t see coming.
I started building a tool to help with this. Just something to check the content before uploading, to see if it might get flagged. That eventually became frametwo.
frametwo scans your video and flags the kinds of things that might trigger platform filters. It looks at visuals, language, and metadata. It doesn't rewrite your content or suggest edits. It just gives you a clear report, so you can decide what to do before your content gets punished.
I built the analysis system myself while balancing school and work. It\u2019s called Nextros, and it\u2019s focused specifically on moderation risks, not general-purpose AI. I didn\u2019t want a black box. I wanted something that helps creators understand what might go wrong, without guessing or overcorrecting.
The site is at https://frametwo.com
I\u2019d appreciate any feedback. Especially from people who have run into the same issues, or have thoughts on how this could be more useful to creators."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Frametwo \u2013 Tired of my videos getting flagged, so I built this"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://frametwo.com/"}},"_tags":["story","author_andrewjustus","story_44384584","show_hn"],"author":"andrewjustus","created_at":"2025-06-26T06:00:40Z","created_at_i":1750917640,"num_comments":0,"objectID":"44384584","points":5,"story_id":44384584,"story_text":"Over the past year I uploaded a bunch of videos to my personal channel. Most of them were short documentaries, news-style edits, or content covering current events and war. A lot of it ended up getting demonetized, age-restricted, or removed completely, usually without any clear reason or explanation.
The worst part was not knowing what caused it. You upload something, it seems fine for a while, and then it disappears or gets buried. I spent days editing videos that ended up dead in the water because of things I couldn\u2019t see coming.
I started building a tool to help with this. Just something to check the content before uploading, to see if it might get flagged. That eventually became frametwo.
frametwo scans your video and flags the kinds of things that might trigger platform filters. It looks at visuals, language, and metadata. It doesn't rewrite your content or suggest edits. It just gives you a clear report, so you can decide what to do before your content gets punished.
I built the analysis system myself while balancing school and work. It\u2019s called Nextros, and it\u2019s focused specifically on moderation risks, not general-purpose AI. I didn\u2019t want a black box. I wanted something that helps creators understand what might go wrong, without guessing or overcorrecting.
The site is at https://frametwo.com
I\u2019d appreciate any feedback. Especially from people who have run into the same issues, or have thoughts on how this could be more useful to creators.","title":"Show HN: Frametwo \u2013 Tired of my videos getting flagged, so I built this","updated_at":"2025-06-27T10:37:46Z","url":"https://frametwo.com/"},{"_highlightResult":{"author":{"fullyHighlighted":true,"matchLevel":"partial","matchedWords":["nexthop"],"value":"nextos"},"story_text":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"Like many people on HN, I practically live inside a text editor, a terminal, and a web browser (to read documentation). I handle these with a tiling window manager.
What are good sweet spot monitor aspect ratios and resolutions for plain text?
I used a 16:9 (1366x768) 11 inch screen for long time, but I realize it is limiting as it is too short and too narrow to split into several windows.
Some time ago, I switched to a 16:9 2k (2560x1440) 27 inch screen. A mainstream setup now. It has been a great increase in productivity as I can now comfortably split the screen with StumpWM or Emacs and have 2 ample windows side by side.
4k screens are typically 3840 x 2160, so still 16:9. With 1.5 scaling, that's the same as a 2k screen. Just more smooth text. If cost is not a concern, it is generally better. Unless scaling is still problematic in your platform.
What about other aspect ratio-resolution combinations?
I have heard great things about 21:9 3440x1400, where it is easy to fit 3 windows side by side.
Sadly I haven't seen any 2k 16:10 screens (2560x1600), or 4:3 (2560x1920) which would be great for having two taller windows side by side.
However, there are some intriguing 1:1 2048x2048 and 1920x1920 panels around mostly used for air traffic control. Eizo does sell one 1920x1920 for end users, and it is not too expensive."},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: What is the optimal monitor aspect ratio for programming?"}},"_tags":["story","author_nextos","story_20087403","ask_hn"],"author":"nextos","children":[20087576,20087847,20087961],"created_at":"2019-06-03T19:33:55Z","created_at_i":1559590435,"num_comments":4,"objectID":"20087403","points":3,"story_id":20087403,"story_text":"Like many people on HN, I practically live inside a text editor, a terminal, and a web browser (to read documentation). I handle these with a tiling window manager.
What are good sweet spot monitor aspect ratios and resolutions for plain text?
I used a 16:9 (1366x768) 11 inch screen for long time, but I realize it is limiting as it is too short and too narrow to split into several windows.
Some time ago, I switched to a 16:9 2k (2560x1440) 27 inch screen. A mainstream setup now. It has been a great increase in productivity as I can now comfortably split the screen with StumpWM or Emacs and have 2 ample windows side by side.
4k screens are typically 3840 x 2160, so still 16:9. With 1.5 scaling, that's the same as a 2k screen. Just more smooth text. If cost is not a concern, it is generally better. Unless scaling is still problematic in your platform.
What about other aspect ratio-resolution combinations?
I have heard great things about 21:9 3440x1400, where it is easy to fit 3 windows side by side.
Sadly I haven't seen any 2k 16:10 screens (2560x1600), or 4:3 (2560x1920) which would be great for having two taller windows side by side.
However, there are some intriguing 1:1 2048x2048 and 1920x1920 panels around mostly used for air traffic control. Eizo does sell one 1920x1920 for end users, and it is not too expensive.","title":"Ask HN: What is the optimal monitor aspect ratio for programming?","updated_at":"2024-09-20T04:20:51Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"andrewjneumann"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["nexthop","ai"],"value":"Couldn't find an iOS client that supported gpodder or Nexcloud (NextPod), so I built one. While it's flutter/dart built the focus of development is iOS and ecosystem features eg WatchOS support, CarPlay, AirPlay streaming, etc. You can download it from the Apple App store (iOS/iPadOS/macOS) for a few bucks (TestFlight is also available) or compile yourself from repo (tested on iOS & Linux, should work elsewhere too).
Adding podcast search next and thinking about adding standalone account support (keeping podcast data local and not requiring gpodder sync).
What features is it missing? What do you want to see from your perfect podcast app? What do you wish your current podcast app didn't do?
https://asecretcompany.com/yourpods/ has more market-y details."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Built YourPods an open-source iOS gPodder / Nextcloud podcast client"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/asecretcompany/yourpods-source"}},"_tags":["story","author_andrewjneumann","story_46799018","show_hn"],"author":"andrewjneumann","created_at":"2026-01-28T17:56:10Z","created_at_i":1769622970,"num_comments":0,"objectID":"46799018","points":1,"story_id":46799018,"story_text":"Couldn't find an iOS client that supported gpodder or Nexcloud (NextPod), so I built one. While it's flutter/dart built the focus of development is iOS and ecosystem features eg WatchOS support, CarPlay, AirPlay streaming, etc. You can download it from the Apple App store (iOS/iPadOS/macOS) for a few bucks (TestFlight is also available) or compile yourself from repo (tested on iOS & Linux, should work elsewhere too).
Adding podcast search next and thinking about adding standalone account support (keeping podcast data local and not requiring gpodder sync).
What features is it missing? What do you want to see from your perfect podcast app? What do you wish your current podcast app didn't do?
https://asecretcompany.com/yourpods/ has more market-y details.","title":"Show HN: Built YourPods an open-source iOS gPodder / Nextcloud podcast client","updated_at":"2026-03-05T23:27:56Z","url":"https://github.com/asecretcompany/yourpods-source"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"randomname4325"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["nexthop","ai"],"value":"Here\u2019s the clean \u201cshort thesis\u201d on Oracle (ORCL), with the biggest swing factors and near-term catalyst checks.
The bear case in 8 bullets
Valuation looks frothy vs peers. After the post-earnings melt-up, ORCL trades around 45x forward EPS, well above Amazon and Microsoft near 31x, which leaves little room for execution hiccups. \nReuters
Backlog quality and concentration risk. Oracle\u2019s remaining performance obligations (RPO) spiked 359% to $455B after four multi-billion contracts with three customers. That is enormous concentration. If any mega customer delays, resizes, or staggers deployments, recognized revenue could lag the narrative. \nOracle Investor Relations
Capex and free cash flow pressure. Management guided FY26 capex to about $35B, up sharply from FY25 levels, to build AI data centers. FY25 operating cash flow was $20.8B, so sustained heavy build could keep FCF tight or negative near term. High spend plus a big dividend tightens the cushion if growth moderates. \nYahoo Finance\n+1
Capacity, power, and execution constraints. Oracle has been \u201cwaving off customers\u201d while it races to add regions, which underscores how sensitive growth is to supply build-out. Large new campuses may need temporary gas generation and carry hefty opex until grid hookups land. Slips here hurt revenue timing and margins. \nCIO Dive\n+1
Customer concentration in AI labs carries headline risk. Reports highlight very large OpenAI commitments and a 4.5-gigawatt \u201cStargate\u201d build with OpenAI. If AI spending normalizes or financing tightens, these multi-year ramps could be resized. Note that the dollar figures are reported by press, not detailed by Oracle in filings. \nReuters\n+1
Oracle Health (Cerner) remains complex. The VA kept the Oracle Health EHR effort moving, but the program has been contentious and timelines have shifted. Any renewed operational setbacks can weigh on sentiment and mix. \nVA News\n+1
TikTok exposure is a policy wild card. Oracle\u2019s \u201cProject Texas\u201d role ties some revenue to an app under ongoing U.S. legal and political pressure. Divest-or-ban deadlines have been extended, but the policy path is not fully settled. Changes to hosting could impact related workloads. \nSupreme Court\n+2\nPolitico\n+2
Balance sheet leverage is not trivial. Total debt sits near the $100B mark by recent tallies. Higher rates plus larger capex means interest and refinancing risks matter more if growth expectations cool. \nCompanies Market Cap
What could crack the story near term
Capex intensity vs margins. Watch Q2 and Q3 FY26 gross margin and FCF. If capex ramps faster than revenue recognition, you get negative FCF and EPS drag. Guidance pins capex at ~$35B, so any lift above that or slips in activation would be bearish. \nYahoo Finance
Backlog conversion and logo disclosure. The market is paying for the $455B RPO. If conversion to revenue is slower than implied, or top deals prove smaller than press chatter, multiples compress. \nOracle Investor Relations
Build-out bottlenecks. Signs of power constraints, permitting delays, or forced use of expensive temporary generation at big new sites would pressure margins and push out capacity. \nTechRadar
Oracle Health headlines. Any fresh VA or DoD EHR stumbles. \nNextgov/FCW
Policy shock. A TikTok outcome that reduces Oracle\u2019s hosting role or data residency scope. \nSupreme Court
Common bull rebuttals you\u2019re betting against
\u201cBacklog is the tell.\u201d Yes, RPO is huge, but it is concentrated and subject to timing. The company itself only said four multi-billion contracts, not the detailed dollar amounts making the rounds. You are taking the other side of aggressive conversion and smooth execution.
\u201cValuation is justified by 77% OCI growth.\u201d True that OCI is the engine, but it is coming off a smaller base and needs flawless build to hit long-dated targets. Slower ramps or cost overruns bite faster when you are at 45x forward."},"title":{"matchLevel":"none","matchedWords":[],"value":"The short case for Oracle (via ChatGPT)"}},"_tags":["story","author_randomname4325","story_45211523","ask_hn"],"author":"randomname4325","created_at":"2025-09-11T13:39:36Z","created_at_i":1757597976,"num_comments":0,"objectID":"45211523","points":1,"story_id":45211523,"story_text":"Here\u2019s the clean \u201cshort thesis\u201d on Oracle (ORCL), with the biggest swing factors and near-term catalyst checks.
The bear case in 8 bullets
Valuation looks frothy vs peers. After the post-earnings melt-up, ORCL trades around 45x forward EPS, well above Amazon and Microsoft near 31x, which leaves little room for execution hiccups. \nReuters
Backlog quality and concentration risk. Oracle\u2019s remaining performance obligations (RPO) spiked 359% to $455B after four multi-billion contracts with three customers. That is enormous concentration. If any mega customer delays, resizes, or staggers deployments, recognized revenue could lag the narrative. \nOracle Investor Relations
Capex and free cash flow pressure. Management guided FY26 capex to about $35B, up sharply from FY25 levels, to build AI data centers. FY25 operating cash flow was $20.8B, so sustained heavy build could keep FCF tight or negative near term. High spend plus a big dividend tightens the cushion if growth moderates. \nYahoo Finance\n+1
Capacity, power, and execution constraints. Oracle has been \u201cwaving off customers\u201d while it races to add regions, which underscores how sensitive growth is to supply build-out. Large new campuses may need temporary gas generation and carry hefty opex until grid hookups land. Slips here hurt revenue timing and margins. \nCIO Dive\n+1
Customer concentration in AI labs carries headline risk. Reports highlight very large OpenAI commitments and a 4.5-gigawatt \u201cStargate\u201d build with OpenAI. If AI spending normalizes or financing tightens, these multi-year ramps could be resized. Note that the dollar figures are reported by press, not detailed by Oracle in filings. \nReuters\n+1
Oracle Health (Cerner) remains complex. The VA kept the Oracle Health EHR effort moving, but the program has been contentious and timelines have shifted. Any renewed operational setbacks can weigh on sentiment and mix. \nVA News\n+1
TikTok exposure is a policy wild card. Oracle\u2019s \u201cProject Texas\u201d role ties some revenue to an app under ongoing U.S. legal and political pressure. Divest-or-ban deadlines have been extended, but the policy path is not fully settled. Changes to hosting could impact related workloads. \nSupreme Court\n+2\nPolitico\n+2
Balance sheet leverage is not trivial. Total debt sits near the $100B mark by recent tallies. Higher rates plus larger capex means interest and refinancing risks matter more if growth expectations cool. \nCompanies Market Cap
What could crack the story near term
Capex intensity vs margins. Watch Q2 and Q3 FY26 gross margin and FCF. If capex ramps faster than revenue recognition, you get negative FCF and EPS drag. Guidance pins capex at ~$35B, so any lift above that or slips in activation would be bearish. \nYahoo Finance
Backlog conversion and logo disclosure. The market is paying for the $455B RPO. If conversion to revenue is slower than implied, or top deals prove smaller than press chatter, multiples compress. \nOracle Investor Relations
Build-out bottlenecks. Signs of power constraints, permitting delays, or forced use of expensive temporary generation at big new sites would pressure margins and push out capacity. \nTechRadar
Oracle Health headlines. Any fresh VA or DoD EHR stumbles. \nNextgov/FCW
Policy shock. A TikTok outcome that reduces Oracle\u2019s hosting role or data residency scope. \nSupreme Court
Common bull rebuttals you\u2019re betting against
\u201cBacklog is the tell.\u201d Yes, RPO is huge, but it is concentrated and subject to timing. The company itself only said four multi-billion contracts, not the detailed dollar amounts making the rounds. You are taking the other side of aggressive conversion and smooth execution.
\u201cValuation is justified by 77% OCI growth.\u201d True that OCI is the engine, but it is coming off a smaller base and needs flawless build to hit long-dated targets. Slower ramps or cost overruns bite faster when you are at 45x forward.","title":"The short case for Oracle (via ChatGPT)","updated_at":"2026-03-05T22:38:59Z"}],"hitsPerPage":50,"nbHits":51,"nbPages":2,"page":0,"params":"query=Nexthop+AI&tags=story&hitsPerPage=50&advancedSyntax=true&analyticsTags=backend","processingTimeMS":16,"processingTimingsMS":{"_request":{"queue":1,"roundTrip":15},"afterFetch":{"format":{"highlighting":1,"total":1},"merge":{"mergeLoop":{"prepareNextHit":5,"total":5},"total":5},"total":5},"fetch":{"query":7,"scanning":2,"total":10},"total":16},"query":"Nexthop AI","serverTimeMS":20}