DeFi Monster(e/acc)

DeFi would eat the world .

[ranjan_raj @ranjan_vittal]

Today, I'm thrilled to announce Pramaana's $27M seed, led by @khoslaventures.
The foundational domains that hold the world together: tax, law, finance, healthcare; all run on certainty. Probabilistic AI can't give them that. We’ve been asked to accept wrong answers with AI as ‘hallucinations’, while in traditional software terms, it’s just a bug. And a wrong answer in such mission-critical domains is more than just a bug, it's a liability that could have catastrophic impact.

We built Pramaana to deliver a 100% trustable experience to the domains that run on certainty: AI that is provably correct, not probabilistically correct. We turn statute and regulation into machine-verifiable code, so every output ships with mathematical proof of correctness. Our mission is to make AI take ownership of it’s work.
Pramaana in Sanskrit stands for “means of valid knowledge”, and we’re going to achieve that by formalizing the world’s knowledge.

[Minn @minney_cat]

In the age of AI, human ingenuity becomes even more valuable. The likes of @elonmusk @spacex, @nvidia, @anthropic, @OpenAI were founded in the U.S. for a reason.

Over the years, I've met thousands of talented founders, engineers, and technologists, and many left everything familiar for the chance to build something new.

So we sat down to hear their stories.

First Landing 👇

[Sasha Malysheva @aimalysheva]

added another batch to the list of neolabs and people behind them I've been compiling

- @primeintellect (@vincentweisser, @willccbb)
- @merge (@SumnerLN)
- @kyutai_labs (@EXGRV, @ptrkprz)
- @poetiq_ai (@markiewagner)
- @GoodfireAI (@ericho_goodfire, @banburismus_)

https://t.co/mIJB2Ml8GO

bookmark the post, subscribe to the list, and let me know who else should be on it!

[Molly O’Shea @MollySOShea]

NEW: a16z bought Turpentine. OpenAI bought TBPN. Jack Altman joined Benchmark. Now Lightspeed makes its move..

The $40B AUM VC fund hired Claire Zau (aka 'Zauey Talks' on IG) to lead *new media* & co-host Lightwork, its weekly AI podcast

Zauey Talks Stats:
- 350K+ followers
- 10M+ monthly impressions
- Built in ~1 year

We cover:
- The tech media land grab
- Split investing + media role
- Tech X vs mainstream social media
- AI bubble?
- Why does GenZ hate AI?
- Who tells the AI story best?

Lightwork is co-hosted by @clairejyz & CMO @Machiz

Since 2000, @lightspeedvp has been an early backer of Abridge, Anthropic, Anduril, Castelion, Databricks, Glean, Mistral, Navan, Neko Health, Netskope, Thinking Machines, Reflection AI, Rubrik, Snap, Skild AI, Vinted, Wiz, & more.

𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒
(00:00) Claire Zau, Partner at Lightspeed Venture Partners
(00:54) Inside Lightspeed's new podcast studio
(02:37) How Josh Found Claire on TikTok
(04:34) Why venture expertise is missing from the mainstream
(08:23) The content that actually works
(10:19) What TikTok comments taught her about AI perception
(13:01) Keeping editorial independence at a mega fund
(14:42) The split role nobody thought was possible
(17:20) The show built to bring mainstream audiences into venture
(20:43) The AI bubble question everyone's asking
(22:00) What Claire is most excited about covering
(23:39) Why Gen Z hating AI makes no sense
(28:50) How big tech companies can fix their image problem
(32:16) Who tells the AI story best
(34:14) The Lightspeed portfolio companies to watch
(38:40) Two creators compare notes on building in media
(43:31) Getting recognized in the wild
(45:43) Expanding beyond ed tech into AI frontier

[Ellie in Space 🚀💫 @esherifftv]

Just thought I’d reshare my interview with the GOAT, 🐐 @lrocket

So happy for SpaceX!

[Jonathan Brebner @JPBrebner]

MINUS ONE TO SPACE GUN

with @natosaichek from @LongshotSpace @southpkcommons

(00:00) The 15-Kilometer Space Gun
(07:48) The History of Space Guns and Why the Idea Was Abandoned
(11:20) Why Rockets Aren’t the Best Solution (17:50) The Story of Longshot Space
(23:37) Nato On Leaving His Dream Job
(30:33) Fundraising for an Unconventional Idea
(42:22) The Perks of Having a Cofounder

[Andrew Côté @Andercot]

SF Deep Tech Week starts in less than two weeks.

There is 70 events on the calendar and the world's leading founders, everything from Fusion energy, Manufacturing, Defense, and the CEO's of publicly traded companies.

Folks like @aphysicist building the future of manufacturing
@bscholl creating next-generation supersonic jets
@TheaEnergy presenting Helios, their first-generation stellarator fusion plant
@EdwardMehr deploying high mix reconfigurable manufacturing
@jimbelosic gigascaling fast turnaround parts production
@jordannoone with AI accelerated CAD and CAM
@intel CEO Lip-Bu Tan, I mean, its the CEO of intel
@LongshotSpace building space cannons to LEO
@tybernstein deploying nuclear batteries for land, sea and space
and chief scientist of Pacific Fusion Will Regan, who's raised $950 million to commercialize magnetized target fusion

If you want to see the future being built today, come to SF for June 22nd-26th

Everything is free to attend.

[Shen Sean Chen @ShenSeanChen]

I spent an incredible afternoon in Shenzhen with @leon2mcp, ex-Product Lead at @Kimi_Moonshot and @capcutapp, diving deep into the future of AI-native collaboration.

Leon left ByteDance when Midjourney and ChatGPT blew up, joined Kimi during its darkest days, and witnessed its explosive growth. Now, he’s raised $40M+ to build @YouWareAI (a vibe coding tool predating Lovable) and @Bloome_im (an AI agent collaboration platform).

The energy in Shenzhen was electric. We had a raw, unfiltered conversation about the real differences between Silicon Valley and China, and how AI is fundamentally rewriting the rules of software and teamwork.

🎬 YouTube: https://t.co/5jasG7ImjW

In this deep-dive podcast, we discussed:

• The Next China Shock is here: While the US dominates Enterprise/B2B, China’s consumer AI execution speed is unmatched. Both regions are incredibly competitive, but in completely different ways.

• Coding is Everything: The release of Sonnet 3.5 was a turning point. It didn't just make developers faster; it empowered anyone to build. The coding agent war is the most intense battlefield right now.

• Stop @-ing humans, start @-ing agents: In Leon's team, the workflow has completely shifted. People don't interrupt each other anymore; they @ their agents. Asynchronous, flow-state collaboration is boosting dev efficiency by 5x.

• Agents need a social graph: "Man is the sum of social relations." If we treat agents as humans, not just tools, they need their own context, memory, and permission systems within an organization.

• Token-based pricing is dead: Charging by tokens is like paying dock workers by the sack. The future of AI monetization lies in outcome-based pricing and building ultimate "Trust" (just like Stripe did).

Huge thanks to Leon for hosting the entire venue and sharing such high-density, builder-level insights. The collision of minds in Shenzhen is truly something else.

You can build anything. You can learn anything. 💪

#AI #Startups #BuildInPublic #Shenzhen #AIAgents #FutureOfWork

[Westinghouse Nuclear @WECNuclear]

In the nuclear industry, 3D printing goes far beyond typical plastic filaments and resins. 3D printing, also known as additive manufacturing, uses a high‑powered laser or electron beam to fuse ultra‑fine metal powder into solid material, one thin layer at a time. Westinghouse is leveraging additive manufacturing to produce complex components that can withstand the extreme conditions inside a nuclear reactor... and that's just the beginning.

Discover how Westinghouse is advancing additive manufacturing innovation across the nuclear industry in our latest Insights article: https://t.co/xAH4GONvtR

[sourcery @sourceryy]

SpaceX's first employee Tom Mueller @lrocket says he first met @elonmusk by making amateur rockets in his garage:

" I've done a few things in my garage."

[Ilir Aliu @IlirAliu_]

MIT Press published a robotics textbook.
Then put it on GitHub for FREE. 📌
"Introduction to Autonomous Robots" covers everything:

kinematics, sensors, actuators, motion planning, localization, computer vision, and neural networks... from mechanisms all the way to algorithms.

It's written for undergraduates. Which means it's actually readable.

Most robotics textbooks assume you're already deep in the field. This one builds everything from the ground up, step by step, with real examples. Stanford's Mac Schwager called it "much-needed" (because it genuinely is).

Four professors at the University of Colorado Boulder spent years building it from lecture notes. MIT Press published it. Then they open-sourced the whole thing under Creative Commons.

PDF. Free. GitHub.

If you're trying to understand how autonomous robots actually work (not just the frontier research, but the foundations), this is where to start.

📌 [https://t.co/bw8zoK8MmB]

Share this with your fellow roboticist!

——

Weekly robotics and AI insights.
Subscribe free: https://t.co/9Nm01QUcw3

[Lena @dolylupec]

I’m heading to China this weekend for the next two weeks to explore the robotics frontier up close.

This is my first time visiting China, and I’m genuinely thrilled! Will be in Shenzhen, Hangzhou, Shanghai, and Beijing.

If you’re building in robotics or Physical AI — whether founder, researcher, engineer, or investor — in any of these cities and open to grabbing a coffee or chatting, I’d love to connect. DMs are open!

China’s robotics progress over the past year has been massive. After spending time following the robotics scene in the US and other parts of Asia, I’m really excited to experience this major ecosystem firsthand — and compare the differences and unique strengths to gain fresh insights.

Looking forward to meeting some of you and swapping insights 🔥

[Kyunghyun Cho @kchonyc]

sometimes , it really does take a decade for me to understand a paper and appreciate its insight and foresight. <Discovering Causal Signals in Images> is one such paper. wow ... david, @robertnishihara , @soumithchintala , @bschoelkopf & @LeonBottou really did see the future.

[Sonya Huang 🐥 @sonyatweetybird]

Did lemurs and dogs invade the set of Training Data, or did we have way too much fun with Google Omni?!?!?!

Super fun Training Data episode with @OfficialLoganK of @GoogleDeepMind on Omni, multi-modal models and generative media, vibe coding and the future of world models and video games, agent harnesses, the agentic Gemini era and more.

Enjoy!

00:00 Introduction
01:47 Agentic Gemini Era
03:05 Antigravity Agent Harness
05:07 Cannibalization and Outcomes
08:24 How Agentic Are We
14:22 Gemini vs Codex Claude
19:11 Vibe Coding Games
26:13 What People Build
27:07 Vibe Coding Games Soon
28:01 World Models vs Engines
29:29 Omni World Model Blur
31:10 Single Omni Model
33:50 Authentic Gen Media
35:19 Vibe Coding Android Apps
38:32 Scaffolding and Startup Edge
43:54 Inside DeepMind Culture

[Guilherme Penedo @gui_penedo]

Today we’re announcing Macrodata Labs.

Over the last few years, @HKydlicek and I have been turning a large part of the internet into some of the largest open LLM pre-training datasets. Through FineWeb, FineWeb2, FinePDFs, FineTranslations, and related work, we got a front-row seat to how scaling compute and data drove progress in LLMs.

We are starting to see a similar takeoff in robotics.

Building on advances in LLMs and VLMs, robotics is finally starting to scale. But physical data is messy in ways text isn’t: large video files, multi-rate sensors, many different formats, and open questions around what signals to record, which annotations matter, and how to turn all that context into better policies.

That makes data work in robotics especially important. Teams need to extract as much signal as possible from every demonstration, trajectory, video frame, and sensor stream, without rebuilding their whole data stack every time they change robot, sensors, format, or labeling method.

We think the right tooling for this is still missing. That is what we created Macrodata Labs to build. Our first step is Refiner, an open-source framework for processing robotics datasets.

We designed Refiner to handle a variety of robotics formats and help teams extract more signal from each demonstration. It is shipping today with support for hand-tracking, subtask annotation, and reward model scoring.

We are also launching a cloud version of Refiner, so teams can focus on their data instead of infrastructure. With a one-line code change, the same pipeline can scale on our platform, with sharding, checkpointing, model deployments, failure recovery, and detailed observability built in.

We’re fortunate to be backed by Air Street Capital, Drysdale Ventures, OPRTRS club, Kima Ventures, YG (Alex Yazdi), >commit, Thomas Wolf, and many incredible angels from top AI labs and technology companies.

I’m excited to keep exploring how better data work can push the frontier of AI, now in the physical world.

If @macrodata_labs sounds interesting to you, or if you are building in the space, I would love to hear from you.

[Azalia Mirhoseini @Azaliamirh]

At Ricursive, we are accelerating the inner loop of AI evolution through chip design. If you speed up training and test time by a multiplier, designing a new model can be sped up by a multiplier!

Thank you, @sallywf from EE Times for a thoughtful piece on @RicursiveAI!
https://t.co/Z2Yx4bx8pL

[Teddy Blank @teddypowday]

We now have VC-backed startups in geothermal power doing:
- Hydrothermal
- Super-deep hydrothermal
- Superhot/magma-adjacent
- EGS (fracking)
- EGS (electrostim)
- Closed-loop
- Geo-pressured
- CO2-plume
- Oilfield/retrofit
- Seafloor thermal vents

SEVERELY LACKING:
- Orbital

[thecollective @thecollectiv3e]

We’re super excited to announce our 2026 Fall Fellows!

Meet the fellows:

- @Sarah___Fan: USC, @uselemma_ai, @photon_hq
- @createdbymason: UC Berkeley, @byhackathonco
- @michelletliu: UCLA, @sundaysinla, @figma, @Apple
- @maya_l39: Western Univ, Vercel, 8VC, Notion
- @roxiegrunenwald: Yale, @hquotientxyz, @DormRoomFund
- @IlyssaYan: Cornell, @designmeetuphq, @MongoDB, @Google
- @imalexwang: Georgia Tech, Vert, @paraformtalent
- @Christine Wu: Boston College, content creator
- @bellaxandrade: Harvard, @massaicoalition
- @lindictive: NYU, founder at @ohtnyc, @tryscope_app
- @tinajyao: UVA, Girls Who Innovate, content creator
- @catarina_chia: Santa Clara, @ToborlifeAI
- Maria Eberhardt: Boston Univ, Resolution, @tiktok_us
- Olivia Liu: UPenn, Venture Lab, Moso Capital
- Siddharth Rout: Columbia, founder at Stealth
- Miraya Jain: UIUC, Product Space, @DormRoomFund

The Collective exists to create spaces for ambitious young people shaping the world. We hand selected fellows from over 750+ applications due to their understanding of culture, tech, and innovation.

This is one of the most talented groups we’ve ever seen and we’re excited to see what they accomplish.

Congrats to each one of them!!

[Sasha Malysheva @aimalysheva]

bio batch, as promised:

- @periodiclabs (@khoomeik, @agarwl_, @LiamFedus)
- @IsomorphicLabs (@MilesCcc, @m_schaarschmidt)
- @LilaSciences (@GVMaltzahn)
- @chaidiscovery (@jackdent, @joshim5)
- @EvolutionaryScale (@alexrives)
- @cusp_ai (@wellingmax)

https://t.co/TcE74ch0xF

[Harry Stebbings @HarryStebbings]

One of the best shows I have ever done.

1. The biggest problem today is power.

2. We will see large resistance to data centre buildout continue.

3. Micron will be worth more than Meta.

4. Export controls have meant China has developed their own architecture.

Alongside their ability to build data centres faster and cheaper, this makes them a real threat.

This and so much more on Monday but holy s*** this was like next level @AravSrinivas