Загрузка видео...

Не удалось загрузить видео

На главную

I gave Hermes OS one voice command and watched it build a website “Create a website for a hair salon.” My agent started writing code, moving through windows, and building the site right in front of me. I could follow the whole process in real time I just added...

16,194 просмотров • 4 дней назад •via X (Twitter)

Комментарии: 12

Фото профиля Ziwen
Ziwen3 дней назад

give my man a huge up

Фото профиля Maurice | KI-Vater
Maurice | KI-Vater3 дней назад

voice is exactly the part i'm still chasing on the watch side. install works and my chats show up on a real apple watch, but dictation to reply is still open. what broke first when you added native voice, or was it mostly latency tuning?

Фото профиля Gregor
Gregor3 дней назад

watching it move through windows in real time is the wild part

Фото профиля witcheer
witcheer4 дней назад

Tony Stark

Фото профиля isc0o0
isc0o03 дней назад

Is this a live chat! Iand how about your voice data privacy in case not fully offline

Фото профиля Luke The Dev
Luke The Dev3 дней назад

It’s your own computer, your own iOS. All data stays in your computer

Фото профиля Abearica
Abearica3 дней назад

LOL nice man 😂 @thevoiceofcash this dude built a website!

Фото профиля M3Labs
M3Labs3 дней назад

Looks like absolute slop

Фото профиля 肖勤立
肖勤立3 дней назад

Hermes只有python对接google agycli最省钱。

Фото профиля Mohit Nagaraj ☄️
Mohit Nagaraj ☄️3 дней назад

watching it move through windows instead of just editing files directly is the part that would freak me out, feels like handing someone your keyboard

Фото профиля Richard J. H. Ditbenner
Richard J. H. Ditbenner3 дней назад

O projeto tá muito legal, Luke! Parabéns! 👏🏼 Vi que você estava rodando ele no Linux. Nesse último vídeo está executando no macOS? Qual stack você está usando nele?

Фото профиля Siftloom
Siftloom3 дней назад

Watching it move through windows is what makes the demo land - headless agents ask you to trust a transcript, this lets you audit the run. Can you interrupt by voice mid-task or does it queue until the step finishes?

Похожие видео