Naman

beautiful text animations are now built-in to the editor harness. no need to spend hours perfecting your motion text. you'll notice the difference immediately.

why is it costing you $6 per minute? mine does it in cents. extract frames every 100ms and then turn them into contact sheets of 5x6 grid. then make the agent read those sheets together instead of reading every frame at full resolution. i've also solved 4 by having the llm breakdown the footage across 8-10 axis of (what makes a video good) and write it out in pure text. then query over those for context. link in bio if you'd like to take a look at my ai video editor. does most of what you're saying + built in timeline editor for manual polish. happy to give a demo too.

the ideal version of Ai video editors is you give it a file full of footage and in one click you have the optimized version of this concept To do this you need 1. the Ai to be able to consume all video files accurately (This is possible with a very well tuned harness as my company already does it but it’s slow and costs about $6 per minute of footage (feasible for larger productions)) 2. Extract data that matters (Ais don’t actually know what is good content vs bad without custom datasets. Through our testing latest models can identify retention mistakes with 35~ accuracy. The best data get set up to about 70% currently) 3. Understand what the story is and what it optimally could be (This requires the ability to think beyond context it’s given like the best editors do - this is possible with good prompting) 4. Rapidly query the entire dataset. (Hours of footage is easily 10M tokens plus so the ai can’t look at it all at once. You need to query across the dataset but this only works if you have number 2 on lock and know what you’re looking for) 5. Creatively diverge and then judge quality accurately (diverging and judging only work insofar that you have the data that allows you to do it accurately) 6. Harness (the agent harness will be extremely complex with all the functions required. Again this is doable it’s just very far from solved currently)

There’s like 6 fundamental breakthroughs needed before ‘Ai video editors’ can work like you’d want them to

Imagine being so valuable that people turn to god to have the privilege of paying money to see you from 100ft away

@natepreneur @businessbarista I have something for you Nate

@businessbarista Have searched high and low for something like this and there doesn't appear to be one solution... But the components seem to be there, including solutions like Gemini or Twelve Labs' video vision AI.

I want to start editing my short-form & long-form videos programmatically with AI. A few months ago, AI video editing wasn't there yet. Remotion was the only thing remotely good, but it was focused on motion graphics. Is it still not there or are there any really good tools or GitHub repos that are 90% of the way there with video editing?

@businessbarista I think you might want to try @wovenvideo. I’ve spent a long time perfecting this. Native desktop. LLM powered. Any type of video 10x faster.

@Danhightower Banger 😂

How to ragebait a VC - Which podcast gave you that opinion? - How would you add value if you found a way to be valuable? - You look like you make all of the 2 but none of the 20. - Sorry, we aren't raising from you right now. - Which AI model did you use to vibe code your firm's website? - What startup role would you take if they'd have you?

@cramforce @grok explain this to me like I’m retarded

You need to go one floor down. The options are: taking the stairs or the elevator. The stairs are a bit away and so taking them is slower at p50. However, the elevator is sometimes busy and hence has bad p95. What do you do?

Me using Claude 4.8 ultra code to center a div

@pmarca Insecurity trying to optimize for “being disagreeable” than actually disagreeing.

Many of my favorite people are 100% for the psychological trait of Disagreeableness. But the thing is, they are Disagreeable even when consensus is correct. A difficult puzzle.

@rauchg woven.video. 5.5 + 4.7.

Show me the thing you’ve built with AI you’re most proud of. Reply with a working product URL and what model / agent you primarily used.

don't do this if you're serious about creating good videos. create skills for every step: script-plan > shot-plan > imagine > animate > voiceover > edit-plan > assemble-video iteratively update them as you create different types of videos for your brand

@AndrewPrifer Neat!

Dear frontend devs and UI designers. I bring you Liquid DOM, a complete and faithful implementation of Liquid Glass on the Web. - Shape morphing - All properties animatable - Dynamic refraction and reflection - Adaptive tint - Adaptive specular highlight - Dispersion - Full html integration - Super fast layout engine that works across Canvas and html - Pointer event handling - Framework and renderer-agnostic low level API - High level React API - Ootb @threejs and r3f integration And lots more. Read on for implementation details and demos.

does anyone know where I can find a curated collection of high quality examples of web shaders?

@huozhi beautiful

🧪 Experiment in building a minimal programming language. Focus on parsing and executing directives, less about language features. Inspired by the legendary c4 project. View everything on langsagne.vercel.app

@shuding Neat!

A deployed demo of the <Water> component: glsx.dev/example?shader…

@rauchg AI Gateway is great

Vercel's AI Gateway gives us a glimpse into real-world production AI and Agents usage. Google is king of production scale, Anthropic dominates in coding & spend, OpenAI is growing fast since 5.4, and OSS continues to gain ground. The AI race is a lot more fluid than it looks :)

@theo @ClaudeDevs What will be worse?

@ClaudeDevs Framing this as a free credit instead of a regression for users is wild. Now I have to make the Claude Code experience on T3 Code significantly worse in order to keep using the 25x subsidization provided on a Claude Code subscription. x.com/theo/status/20…

Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK

@jessepollak would you like to try Woven? I've spent months perfecting skills & tools to analyze, create, and edit videos with agents and put them all in this harness. link in bio.

who has good claude/codex skills for editing video into punchy vertical video clips? want to be able to go from raw footage, to refining content included, to adding captions, to adding hook, all in a chat convo with claude and/or codex

@robj3d3 you might also like Woven. link in bio.

I found a solution. We officially live in the future.

@rileybrown It’s already done

I’m telling you right now someone could build a $1B company that’s a full video editor that connects to codex, Claude code and cursor etc. A super app native video editor. Don’t even build the standalone app with an ai agent side panel. Waste of time. Just make it a plugin that works inside codex.