I ask the question to agent from separate session without history to know more about my code

I use claude with remotion and ffmpeg to create final video.

Sounds cool! How easy to come out one with those three? And how much does it cost for the combo? BTW, is there a place to see the video? I'd like to have a look! Thank you!

Most of these apps work great in labs but fail in real life because everyone has their own different needs and the app should adapt to that. And rules, values and goals change day by day, so agents should adapt to those changes as well.

Agreed, but every lab has almost trained their own flavor of LLMs/agents being good and “generalized” at certain tasks rather than personalized for the user. (Like Anthropic models are great at design, GPT good at infra code, etc)

Also exactly “rules, values, and goals change day-by-day” is why learning needs to happen but across the apps/agents you’re using to personalize to you… not having to reconfigure every time for new agent or app that’s good at a certain task..

I think they must be using the Fable or maybe more capable model for internal use purposes.

You could use some smaller models, like parakeet, to figure out the users.

Yes it is and now it's a daily driver for me.

I liked the idea and I'm impressed, but here's what I'm thinking—what if the voice transcription gets it wrong? Like, voice can pick up similar-sounding words and transcribe them differently, you know?

This uses the new onboard TTS model from Apple, which is amazing, even with background noise.

Parakeet TDT 0.6B v3 (int8 CoreML, about 470 MB) runs on-device through FluidAudio's AsrManager, pinned to CPU plus the Neural Engine.

I think it works great on Mac but lags on Windows. The accuracy on Windows is lower than Mac.

How about digital twin of eon.io :)

We have one there! We actually started without it but found it useful for our own development and testing so we added it.