Dylan Wang

Co-founder and chief architect at Pine AI
I built Pine Computer @pine_computer
DMs open for builders.

[Dylan Wang @HyuWang]

Today we're launching Pine Computer.
AI got smarter. Its computer didn't.
Pine computer is a computer built for AI. Your product hands it a job, the work crosses websites, files and business software, and the finished work comes back. Now in private beta for developers.

[Stanley Wei @StanleyWei4748]

AI is already smart enough.

Real-world tasks are still slow, expensive and unreliable, because we hand AI a computer built for humans, then wrap it in a heavy harness.

AI doesn���t need to get smarter. It needs a computer built for it.

Today we’re releasing Pine Computer

[Dylan Wang @HyuWang]

@zhilinjerrywag We're hosting hackathon next week, welcome if you're interested https://t.co/03Q62qRhDl

[Pine Computer @pine_computer]

Come build with Pine Computer for a day.

Saturday, Oct 17, 10am-8pm in SF. Credits for everyone, food all day, and our engineers in the room. Solo or teams up to four;

Demos at 6. Top three win $5000, $2000 and $500.
https://t.co/oAvo0YRfqy

[Dylan Wang @HyuWang]

@HaseebMir91 @StanleyWei4748 We'll share more details on self-hosting and model support as the on-premise version takes shape. For now, we focus on the hosted service side.
Stay tuned for updates from @pine_computer

[Dylan Wang @HyuWang]

@ashwaanthh @StanleyWei4748 We still have a vision-based path combined with application state. we're also continuously expanding support for more apps. There is feature request form inside console for additional needs

[Dylan Wang @HyuWang]

@sorayukinoyume 是的呢 现在在HF0孵化中

[Dylan Wang @HyuWang]

We're overwhelmed by the waitlist now. If you want a fast pass, come and join the hackathon.

[Pine Computer @pine_computer]

Come build with Pine Computer for a day.

Saturday, Oct 17, 10am-8pm in SF. Credits for everyone, food all day, and our engineers in the room. Solo or teams up to four;

Demos at 6. Top three win $5000, $2000 and $500.
https://t.co/oAvo0YRfqy

[Dylan Wang @HyuWang]

@BillyEnriz55084 @pine_computer Thanks for applying! We're getting a lot more interest than we expected, we'll do our best to get dev like you onboarded ASAP.

Would love to see what you build with Pine Computer!

[Dylan Wang @HyuWang]

Just run it yourself. Apply the waitlist at https://t.co/pkocMdp5wP
Ask me and I'll give free credits for app builders. @pine_computer

[Stanley Wei @StanleyWei4748]

AI is already smart enough.

Real-world tasks are still slow, expensive and unreliable, because we hand AI a computer built for humans, then wrap it in a heavy harness.

AI doesn���t need to get smarter. It needs a computer built for it.

Today we’re releasing Pine Computer

[Dylan Wang @HyuWang]

@a7vinx @StanleyWei4748 Thanks ! More technical deep dives coming soon ! stay tuned with @pine_computer

[Dylan Wang @HyuWang]

@EvanJahngm2e @StanleyWei4748 Auditability is critical for workflows like this.
We don't have access to your data inside Pine Computer, but we provide the diagnostic API, so you can use it for audit and troubleshooting yourself.

[Dylan Wang @HyuWang]

@tiachent @StanleyWei4748 Absolutely! we're planning to roll out more spec for heavier workloads.

For now our focus is still at Computer interface and OS layer to make it work better with model.

[Dylan Wang @HyuWang]

@jnru191610 @StanleyWei4748 We going to polish paper around it, stay tuned.

[Dylan Wang @HyuWang]

@DaSonovaBeach @StanleyWei4748 Yep, it's built on Linux, we're also planning to open-source more OS components and work the community to build an ecosystem around it.

[Dylan Wang @HyuWang]

@qwfasdzx @StanleyWei4748 We actually outperform them in many real-world tasks with smaller models.
We used to train the model but with Pine Computer, the whole runtime and OS is unblock model to do more, the the key is treating those as one system. Rather than work on model or tools along.

[Not Elon Musk @ElonMuskAOC]

The model got smarter.

The computer didn’t.

Pine Computer reads the page, not a picture of it.

[Dylan Wang @HyuWang]

@JinRixky @StanleyWei4748 We're already seeing teams using Pine Computer for marketing workflows!

This is exactly the kind of use case we're building for. Spin up computer sessions in parallel, then bring everything together into a deck.

[Dylan Wang @HyuWang]

@migcoliveira @StanleyWei4748 That's exactly one of the hardest parts.
We focus on verifying the actual end state against the original goal, using structured state and the resulting artifacts, rather than just trusting model's assessment.

[Dylan Wang @HyuWang]

@Fartn_Bozo @StanleyWei4748 🤣🤣🤣

[Dylan Wang @HyuWang]

@kimindiehacker @StanleyWei4748 It's true.
We're already seeing this with early enterprise adopters, which solve their internal platform pain so well, just plug it into existing workflow without having to build everything from scratch.

[Dylan Wang @HyuWang]

@OutcodedHuman @StanleyWei4748 Haha, VMs are just the infra underneath.
What we've built is an AI-native computing environment, with structured state, event-driven perception, and native actions across real apps.

The difference isn't the VM. It's what AI can actually get done with it, and at what cost.

[Dylan Wang @HyuWang]

@zhilinjerrywag Ya, let's keep in touch, we going to host Hackathon at SF soon.