Joel

@SentientAGI Mod

[Joel @Joelc_eth]

AI is no longer just chat models,the next leap is happening in physical intelligence and automated discovery.

From humanoid robots (PGI) and chip design agents to living neuron video models and autonomous theorem proving, 10 labs and their recent ships: 🧵👇

[Joel @Joelc_eth]

@SentientAGI @MIT @CarnegieMellon @Microsoft @Google @AlibabaGroup @amazon Seeing different teams build on EvoSkill shows how open research can extend an idea’s impact.

[Sentient @SentientAGI]

This year, EvoSkill has been cited by 60+ papers, including work from @MIT, @CarnegieMellon, @Microsoft, @Google, @AlibabaGroup, @amazon, and more.

Here are 14 of the most impactful papers building on EvoSkill ↓

[Joel @Joelc_eth]

10/10 Archgen Labs (YC F26) (@archgen_)

👀 Tracked By: shyamal (@shyamalanadkat - Product & Strategy @ OpenAI), Ravi Vijayakumar (@thehiphopswami - AI Researcher (ex-Google, OpenAI)).

• What it does: Semiconductor AI lab developing self-learning agentic EDA systems to automate physical chip design and accelerate silicon tape-outs (YC F26).

🚀 Recent Ships:
• Ranked #1 globally in the Hudson River Trading (HRT) & Partcl Macro Placement Challenge 2026 across 120+ international competing teams.
• Built self-learning physical design agents for semiconductors that capture engineering heuristics to autonomously optimize chip floorplans and PPA.
• Integrated with OpenROAD and open-source EDA ecosystems to compress custom ASIC and FPGA chip design timelines from months to weeks.

[Joel @Joelc_eth]

AI is no longer just chat models,the next leap is happening in physical intelligence and automated discovery.

From humanoid robots (PGI) and chip design agents to living neuron video models and autonomous theorem proving, 10 labs and their recent ships: 🧵👇

[Joel @Joelc_eth]

9/10 The Biological Computing Co. (TBC) (@biocomputingco)

👀 Tracked By: Kylie Robison (@kyliebytes - Senior AI Reporter @ The Verge), am.will (@llmjunky - Lynk, Starswap), Kaito (@kaycidot - Founder @Utopialabs ), Brian Schechter (@bschech - Partner @ Primary Venture Partners).

• What it does: Hybrid biocomputing lab harnessing living biological neurons to engineer neural-derived AI software that is 5x faster and significantly more energy-efficient.

🚀 Recent Ships:
• Unveiled the world's first neuron-derived video generation model, translating living biological neural firing patterns into 5x faster AI inference on AWS.
• Formed a strategic infrastructure partnership with Amazon Web Services (AWS) to bring biological wetware compute into cloud environments (featured in WIRED).
• Demonstrated live control of the Pollen Robotics Reachy humanoid robot powered directly by networks of living biological neurons.

[Joel @Joelc_eth]

7/10 Grounded Superintelligence (@groundedsi)

👀 Tracked By: S◎L Big Brain (@solbigbrain - Big Brain VC & Solana Investor), Amber Yang (@theamberyang - Partner @ Lightspeed), Jessy Lin (@realjessylin - Co-Founder @ Engram), Hersh Desai (@hersh_desai - AI Grant Director & Startup Investor).

• What it does: Robotics platform developing egocentric wearable hardware and 3D motion-tracking data infrastructure for generalist robot foundation models.

🚀 Recent Ships:
• Released 'Grounded API', achieving SOTA benchmarks on hand-tracking (<1 cm precision) and egocentric SLAM.
• Introduced 'RoboCap', a wearable sensor suite converting raw in-the-wild human manipulation into metric 3D labels ready for robot training.
• Integrated with Hugging Face LeRobot and Rerun to build a unified data foundation layer for generalist robotic foundation models.

[Joel @Joelc_eth]

The future won't belong to a handful of monopoly corporations; it will be built on independent, open, and innovative infrastructure labs.

Backing small labs early on is why this is critical.

If you found this breakdown useful, feel free to RT the first post and follow along so you don't miss early builder signals. 🫡

[Joel @Joelc_eth]

Top 10 new AI infrastructure &amp; Physical AI robotics labs to watch in 2026:

Multiplication-free additive LLMs, TPU vLLM acceleration, autonomous bio wet-labs, and self-accelerating AGI.

Tracked by OpenAI, Anthropic, and YC leaders , here are their latest ships: 🧵👇

[Joel @Joelc_eth]

10/10 Flourish (@flourishailabs)

👀 Tracked By: Soumith Chintala (@soumithchintala - Creator of PyTorch & Thinking Machines), elie (@eliebakouch - Co-Founder @ Prime Intellect), Christian Catalini (@ccatalini - Founder @ Cryptocat VC & MIT Researcher), Scott Gray (@scottgray76 - OpenAI Core Systems & Triton).

• What it does: Research lab engineering foundation intelligence architectures operating at biological efficiency to transcend GPU compute scaling bottlenecks.

🚀 Recent Ships:
• Announced a $500M funding round at a $2.5B valuation featured in WIRED to pioneer human-level intelligence with human-level efficiency.
• Initiated laboratory testing of cognitive model architectures designed to bypass brute-force parameter scaling via biological efficiency.
• Established an independent frontier research facility backed by PyTorch creator Soumith Chintala and prominent OpenAI systems researchers.