Cong Wang

Building physical AI for science 🤖🧠 · Robot Learning · Embodied AI · AI Agents · Researcher @ Harvard

[Cong Wang @wan73184085]

@sama anyone who can help with my account verification?

[Cong Wang @wan73184085]

@thsottiaux Can you help me with my account?

[Cong Wang @wan73184085]

@elonmusk Why not compare with GPT-live?

[Cong Wang @wan73184085]

Cool swimming robot

[Cong Wang @wan73184085]

@LuisSil24776614 @OpenAI @derrickcchoi Me too

[Cong Wang @wan73184085]

Claude Code for anything is possible

[Hanming Ye @DozenDucc]

Introducing Waddle Labs: Claude Code for robots.

Connect our API to your robot and enter a prompt, then our agents write code to achieve the task in 20 minutes.

@yiding_song @theWaddleLabs

[OrcaDub @OrcaDub]

Introducing OrcaDub 1.0🐳

A frontier speech-to-speech video dubbing model for natural multilingual localization.

Unlike traditional dubbing pipelines, OrcaDub preserves voice identity, emotion, prosody, timing, and lip sync—producing natural-sounding localized videos end to end.

Evaluation
🏆 4.83/5 MOS ↑
🎙️ 96.8% Speaker Similarity ↑
🌍 92.7 COMET ↑
📝 3.1% WER ↓
🎭 98.5% Speaker Attribution ↑

Built for production:
• End-to-end speech → speech
• Emotion & prosody preservation
• Idiom-aware localization
• Song translation
• Per-word forced alignment
• Background music preservation
• Free web Studio to post-edit
• OpenAI-compatible API

Available today.
PAYG at $0.60 per minute with 10 minutes free—no subscription required.

Try it: https://t.co/VRpwGtm7nZ
Model card: https://t.co/fHldhlUjPS
OrcaRouter: https://t.co/8mqH6skCXb

[Kimi.ai @Kimi_Moonshot]

Releasing the model weights and technical report of Kimi K3.

Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.

New model architecture: 2.5x the intelligence per unit of compute, not just more params.

Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.

Model weights: https://t.co/7m7eEg6Y0B
Tech report: https://t.co/yeu6cjpMCT
Tech blog: https://t.co/YTfiMSNM1f

[Y Combinator @ycombinator]

AI is moving into the physical world.

We're excited about a new wave of startups rebuilding the systems that power the real world, from education and healthcare to defense, finance, infrastructure, and work itself.

https://t.co/QCIz6DnQnN

[Black Forest Labs @bfl_ai]

Introducing FLUX 3.

One multi-modal model for Image, Video, Audio and Action-Prediction. Creations are truer to life in every kind of style.

FLUX 3 Video is now available in early access (link below).

Jointly trained in one unified architecture, our model can be extended to predict actions for robotics. See our work with mimic and Audi in the thread.

[Generalist @GeneralistAI]

Why create robot intelligence for just one hand, when we could have it learn from many?

GEN-1, our latest embodied foundation model, now supports a broad range of end effectors from 5-finger hands, to specialized tools, and everything in between.

[Cong Wang @wan73184085]

@TokenGremlin @thsottiaux can it remove the reset button

[Cong Wang @wan73184085]

One phone, two Codex accounts. Our cx and cx2 profiles share a host daemon, so the same mobile login can reach both while quotas stay separate. Claude Code Remote Control is stricter: the phone and local process must use the same Claude account and org.

[Cong Wang @wan73184085]

agent native software

[mewlist @mewlist]

https://t.co/Ygh0OhVeft

Pricing Update:
We want to remove barriers to innovation, so we are making these tools free! Both the CLI and MCP will be free, with no concurrency limits on the MCP server.

[Cong Wang @wan73184085]

the claude sub finished, and codex only looks good

[LiveKit @livekit]

Read the blog: https://t.co/28ALCXxdPA

[Cong Wang @wan73184085]

@Google where is the Gemini pro?

[Cursor @cursor_ai]

We've doubled usage limits for all individual and teams plans!

These limits apply to Grok, Composer, and any new Cursor models.

[Google @Google]

Today we’re expanding the Gemini family with three new models built to be faster, more token efficient, and reliable at scale.

Meet the new Gemini models ↓