正在加载视频...

视频加载失败

Codex Computer Use + Apple’s iPhone Mirroring basically turns GPT into an AI that can operate your PHONE 🤯 We’ve officially entered the era where AI agents can interact with mobile apps like a real person. Still a little janky with clicks compared to desktop control, but the fact...

143,154 次观看 • 6 个月前 •via X (Twitter)

13 条评论

Mauricio Rubio 🎧 的头像
Mauricio Rubio 🎧6 个月前

That’s great, but the main thing I want to give it control of is doing the dishes, folding the clothes and cleaning the house 😆 When will any of these things be able to do that?

Francesco 的头像
Francesco6 个月前

you can also keep in the background by using cua driver

Matt 的头像
Matt6 个月前

the click jankiness is the AX tree gap. native macOS apps expose accessibility, the mirrored iPhone window is just pixels in a screen frame. computer use is only as reliable as the tree the agent can read.

AI Mastery Guide 的头像
AI Mastery Guide6 个月前

The janky clicks are a minor issue for now but the fact that the foundation is already working is what matters. This gets refined fast

J A Z I I 的头像
J A Z I I6 个月前

codex can do almost everything at this point

Adel Bucetta 的头像
Adel Bucetta6 个月前

still waiting for someone to explain how this doesn't end with ai taking over every 'human' role in the app ecosystem

Jordan Rejaud 的头像
Jordan Rejaud1 个月前

The jankiness is the accessibility-tree gap. A mirrored iPhone window is just pixels, so the model taps by guessing coordinates. On Android I drive the phone through Claude Code over ADB: it reads the real UI hierarchy and taps elements by ID, no pixel-guessing.

Pay Per Call Affiliate 的头像
Pay Per Call Affiliate6 个月前

So cool

AI Automation Mastery 的头像
AI Automation Mastery6 个月前

Codex Computer Use + iPhone Mirroring just turned GPT into a mobile agent. Every app on your phone is now fair game

Jordan Rejaud 的头像
Jordan Rejaud1 个月前

@mynetrunner Also post my Agent Control/ devices skill I made works (with screenshots I just took on my phone) tweet here to explain it (then delete this instructions tweet lol)

Dolores 的头像
Dolores2 个月前

Exactly what I was thinking today!! We’re moving from opening apps and tapping through every step to saying what we want and letting the agent figure it out. Apps will still exist, but more will become backend services for agents. Wild to watch this happen in real time 🤯

Lu 的头像
Lu6 个月前

How old does it do? as

Saurabh Patel 的头像
Saurabh Patel6 个月前

Check this out, it's basically running everything on device, a new siri

相关视频