Verified humans now have a new base. $KGEN is live on @base.

Our elder @manishdiesel is heading back to Seoul for KBW 🇰🇷 A panel discussion, plenty of conversations, and a whole lot more. See you all around.

🤖KGeN : 이번 추석에 로봇이 우리 집에 온다면?🤖 ( @KGeN_IO #KGeN #PhysicalAI #한가위 ) 우리집에 로봇이 온다면 송편을 좀 이쁘게 빚어줄래??? 저희집은 추석때가 되면 송편을 빚어요. KGeN은 아시다시피 사람을 검증하고 그 검증된 사람들의 데이터로 AI를 가르치고 데이터를 제공한 사람들에게 보상을 제공하는 프로젝트죠. 그래서 단서가 붙습니다. 떡은 무조건 이쁘게 만드는 것이 중요해요. 그렇기 때문에 이쁘게 로봇이 많이 만들어줬으면 좋겠어요. 송편을 빚으면서 어른들과도 대화를 유도해서 잔소리를 좀 적게 듣게 해주길 원합니다. 이것은 일거양득~~!!!! 🤖우리집 로봇의 추석 업무 1. 송편 재료 준비하기 2. 송편 반죽하기 3. 송편 예쁘게 빚기 4. 송편 빚으면서 어른들과 대화하기 -> 화재 전환으로 내가 들을 잔소리 감소, 송편 다량으로 만들기 4. 송편 찜통에 찌기 저와 가족들이 송편 빚는 일에 시간을 덜 보낼 수 있고 어른들은 간단한 응대를 해주면 적적하지 않고 잔소리를 들었던 아이들은 잔소리에서 벗어날 수 있어요~~

16,000 hours recorded in just 24 hours. That’s what a verified human network can do.

Context Matters. Your kitchen might teach a robot one way to perform a task but a restaurant can teach it something completely different. Same task. Different layout, tools, and conditions. For Physical AI to work reliably in the real world, it needs to learn across different environments, not just one controlled setting.

What is BRIDGE ASR 2.0? My intern asked me to explain it like they were 10. Here's my honest attempt. Say you've got a robot at home or on a factory floor. The real question was never "does it work in the demo." It's this: if you tell it to turn off the stove, does it actually hear "turn off," or something close enough that it does nothing? In the physical world, "pretty good" isn't a real number. You need something close to 99.999999% accuracy. One task out of a hundred going wrong isn't a rounding error. It's a stove left on. This isn't me saying BRIDGE ASR 2.0 is beyond argument or question. It's an honest invitation. If you work in voice AI, come challenge it. If we got it wrong, point it out. My DMs are always open. Check out the full report, the methodology, and the datasets here: humynlabs.ai/bridge/ASR/2.0

Physical AI sees the world through you.

Find $KGEN here:

A sneak peek of what our community in LATAM has been upto 👀

[Speaker Announcement] We’re pleased to welcome Manish Agarwal, Co-Founder and CEO of KGeN, as a speaker at The Bridge Summit 2026. Manish brings deep experience across gaming, consumer technology, and large-scale user networks. Before KGeN, he served as CEO of Nazara Technologies, helping lead the company to India’s first gaming IPO and a $1B+ peak valuation. At the summit, he will share how verified human networks can turn global participation into trusted data infrastructure for AI, gaming, and the next generation of digital economies.

Physical AI can see. Hearing the real world is next. Introducing BRIDGE ASR 2.0. The only global independent ASR benchmark evaluating 23 models for real-world deployment across a 6-metric stack. Tested across 23 languages on real dual-speaker overlap, interruptions, code-switching, and conversations. See where your model actually stands. humynlabs.ai/bridge/ASR/2.0

KGEN'S (@KGEN_IO) VERIFIED HUMAN DATA NETWORK FOR PHYSICAL AND MULTI-MODAL AI LAUNCHES ON BASE

Over $2M was paid out across 15 LATAM countries, just for living their lives.

“What do you like about KGeN?” Glad you asked ▶︎ •၊၊||၊|။||||။၊|။|||။|||။၊| 4:54:11

[Platinum Partner Announcement] The Bridge Summit 2026 is proud to announce @KGeN_IO as our official partner. KGeN is the world's largest verified human network — 80M+ users across 60+ countries. The network sources diverse, multimodal data to train Physical AI and LLMs. $KGEN ties that growing revenue directly to on-chain supply reduction — automatic and verifiable by anyone. KGeN is backed by Accel, Prosus Ventures, Jump Capital, Aptos Foundation, and Polygon.

the fly can even record egocentric data

The little things matter. Reaching for a product, picking it up. Looking at the label. They may seem like ordinary moments, but these everyday actions help AI understand how humans actually move through the real world. The future of Physical AI is built on moments like these.

If a robot watched your daily routine, what would it learn from you?

You see the robot perform. We see the 99% training behind it.

Teach AI how humans actually see and interact with the world, around the world. Every perspective you capture helps build better physical AI models. With KGeN, you don’t just contribute, you get rewarded for it.