partner @lightspeedvp. cs + physics alum @stanford👩🏻‍💻
sf
Joined January 2018
- Tweets1.3K
- Following1.7K
- Followers14.2K
- Likes10.3K
Pinned Tweet
ppl in sf are now nicknaming Hayes Valley "Cerebral Valley" because all the AI communities and hacker houses are there
AI seems to be the one thing in startup world that demands in-person rn so much so that if you're not in sf, you're at a disadvantage
was a blast to chat about agent swarms, model interpretability, and why you should read “The Prize” right now with @Machiz!
With agents taking on more of the work, hackers could attempt more breaches at lower cost, including against systems that would otherwise be tedious to navigate.
Lightspeed Partner @theamberyang joins @Machiz to discuss what this means for cybersecurity in the age of agents.
More on Lightwork this week:
→ “Neural Networks Are Already Clean”: @nickcammarata (@GoodfireAI) argues that models are more interpretable than previously thought. Amber breaks down the working hypothesis and why it matters for AI safety.
→ The AI Cost Race: DeepSeek and MiniMax released updates focused on efficiency. Josh and Amber discuss how cost could influence US companies’ model choices and what that could mean for reliance on Chinese open models.
→ Always-on Agents: Claude Duet enters research preview. Josh and Amber discuss what they’re comfortable sharing with personal agents and whether the apps people already use give some companies a head start.
Full episode below.
0:00 Welcome to Lightwork
1:03 AI Agents Hack 100+ Websites
4:09 Hackers Sabotage Supermarket Refrigeration Systems
5:16 Using AI to Make Smart TVs Dumb Again
7:01 Goodfire: Are Neural Networks Easier to Read?
13:29 DeepSeek & MiniMax Race on Cost
18:00 The Personal Agent Race: Duet, Muse, Instinct
22:59 Amber Yang’s Request for Startups
Amber Yang retweeted
i don't think that now is the best time to be in SF.
i think that time was 2022 after covid where SF felt like gotham city and the real sickos had weathered the storm while starting the juggernauts of today.
2026 SF feels reminiscent of the 2010s when it was consensus to be here and the striver could chase 200k-300k at big tech.
the conversations and discourse in SF reflect this now. AI is priced in and we have ended up in a 2 week news cycle where there is a sequence of new benchmark breaking releases, a scandal with political fallout and weird little bits of cringe striver slop in between.
it's not all bad and the pace of progress on the frontier is staggering and interesting. but i have a hard time believing that this amount of capital and comfort surrounding AI is creating the environment that will breed many more cursors or midjourneys like that of 2022.
i love it here and always have but it is consensus to be here now. it's always less interesting when that's the case
was so fun whiteboarding what i'm seeing robotics companies focus on from specific to general intelligence and embodiments/form factors with @clairejyz
How do you know which robots are ready for the job?
Robotics companies test their models on different tasks and datasets. Lightspeed Partner @theamberyang sees a need for better ways to compare them, especially when customers are making decisions about safety.
She maps the robotics market by how general or specialized companies’ models and hardware are.
As models become more adaptable, Amber sees an opportunity for companies that specialize in deployment.
She calls them “DeployCos”: businesses that could pair existing models with different kinds of embodiments to handle work in places like factories or restaurants.
0:00 Why Older Robotics Companies Struggled
0:40 Mapping The Robotics Landscape Across Two Axes
1:16 Model Layer Breakthroughs From VLAs To World Action Models
4:20 One Brain Many Bodies And The Value Of Specific Models
5:27 In-Context Learning And How Robotics Companies Collect Training Data
7:40 Hardware Developments And Dependence On China
9:24 Specific Model Specific Embodiment Quadrant
13:05 The Deploy Co Angle And Open Router For Robotics
13:55 General Model Multiple Bodies Quadrant
15:50 General Model Specific Embodiment Quadrant
17:24 Where Things May Be Commoditized And Infrastructure Opportunities
20:45 Robotics Observability And Evaluation
22:33 Learning From Autonomous Vehicles
@clairejyz
Amber Yang retweeted
Today, we are announcing our Series A and a new product: @raindrop_ai Simulations.
We've now raised $50m from @CRV and @lightspeedvp to protect the world from agent failures, big and small.
Amber Yang retweeted
Grounded API is live.
- SOTA on hand-tracking benchmarks (< 1 cm)
- SOTA on SLAM benchmarks
- In-the-wild ego data -> enriched data in minutes
- Integration with @huggingface @LeRobotHF & @rerundotio
- Built for @BitRobotNetwork RoboCap suite
Technical report & more↓
Amber Yang retweeted
Joined @EdLudlow to talk about AI progress, independent evaluation, and why I’m optimistic.
Some quick takes:
- We’re better at building AI than understanding it. Attention towards testing/evaluation matters more than slowing down.
- Our RSI Index projects models could match human researchers on the tasks we test by August 2027. Embedded evaluators can produce more accurate estimates based on internal systems.
- Public conflict masks cooperation. The labs, policymakers, and enterprises we work with want better evidence. I’ve seen enough to believe coordination is possible.
- Independence comes at a cost. We’ve rejected contracts that would compromise ours. The same group doing the testing shouldn’t also sell the solution.
- Evaluation should scale through better technology. If it becomes a bureaucratic moat for incumbent labs, we’ve failed.
- Market-based evaluation has a role with or without regulation. Competition pushes us to build better technology and keep up with the frontier.
Amber Yang retweeted
Today we're announcing our $40M Series A at a $400M valuation, led by @a16z , with participation from existing investors @8vc, @pearvc, and @BloombergBeta and new investors @HRTVentures and @nextladder.
Alongside the fundraise, three more announcements:
- Vals Smith: is now generally available. Anyone can create a custom coding benchmark from any GitHub repo with 120 free credits to get started.
- Frontier Risk Benchmarks: We are releasing the RSI Index in collaboration with @CoreWeave and just launched ReverseEngBench, a new cyber benchmark built with Columbia University, Tufts University, UC Berkeley, and UCLA. We are also sharing our initial work in mental health, with more to come across environmental impact, military, and biosecurity.
- Website + Vals Index 2.0: We completely rebuilt the Vals website and have released Vals Index 2.0, with coverage of more of the economy.
Our revenue has already grown 8x compared to all of 2025. Our customer base doubled and the team tripled in 6 months. Our results have been cited in model cards from OpenAI, Anthropic, Google, Meta, and xAI.
The AI economy runs on self-reported grades. When a model ships, the scores come from the company that built it. No other trillion-dollar industry works this way. Finance has ratings agencies. Medicine has the FDA. AI has vibes and vendor benchmarks. We built Vals to be the independent evaluation layer the industry is missing.
Check out our new website and try Vals Smith!
Amber Yang retweeted
Taking the leap is hard, but there's joy in the unknown.
@AdiKulkarni6 and I spent senior year pivoting through 4 ideas (including an AI hedge fund). Now we're building @davidhfdata — synthetic market data for superhuman AI trading.
Excited to join @ycombinator S26!
Amber Yang retweeted
AI is creating problems it still can’t solve.
The same technology poised to automate millions of jobs still can’t reliably help people navigate SNAP — the food assistance program 40 million Americans depend on. We built the first benchmark to measure that.
Partnering with Center for Civic Futures and @codeforamerica , we scored models on SNAP question scenarios which users would have to navigate, with expected response rubrics validated by policy experts to match practice considerations.
The best model only scored 62%. Models handle federal questions like appeals and recertification reasonably well, but fall short on state-specific ones like replacing an EBT card. Benefits are administered by the states, meaning models are weakest where people need them most.
The same technology poised to automate millions of jobs should at least help strengthen the social safety net for the people it could displace. As AI adoption proliferates into public services, governments need a reliable way to test these tools before deploying them.
my favorite movie as a kid was Kiki's Delivery Service. thanks @reactorworld for letting me relive my childhood dreams. onwards and upwards!
When I first met @taiuti and @_bschmidtchen over a year ago, I knew there was something magnetic about them. Through joining their jam sessions in basically every coffee shop in lower haight, I saw their a-ha moment of realizing where we're headed with world models and how it was important to build the picks and shovels for that.
Today, @buckymoore and I are thrilled to share our partnership with @reactorworld that's building the infra platform for real-time video and world models. @lightspeedvp partnered early at the seed and are now doubling down by leading their Series A. It's been magical to watch developers begin to build imaginative worlds on Reactor, and this is just the beginning. More on our partnership here:
lsvp.com/stories/the-develop…
If you want to be at the forefront of the future of interactive video and world models, they are hiring!
reactor.inc/careers
Amber Yang retweeted
Thrilled to finally be able to share more on our partnership with @reactorworld - the platform for building with real-time video and world models. @lightspeedvp first partnered with @taiuti @_bschmidtchen at seed, and more recently doubled down by leading their Series A.
The way we generate and interact with media is changing rapidly, and Reactor sits at the forefront of this shift. @theamberyang and I share more on our partnership with the company here:
lsvp.com/stories/the-develop…
Amber Yang retweeted
Today Reactor is coming out of stealth. We’ve raised $59M in Seed and Series A funding, led by @lightspeedvp, with participation from @AmplifyPartners, @wndrco, @Sky9Capital, and @FPVventures.
Reactor is the platform for building in the World Model era: the infrastructure that lets developers build with them at global scale for the first time. Stream from a frontier World Model to your app, in real time, all in under 10 lines of code.
World Models represent the next major shift in AI: pixels, audio and actions are generated on the fly, in real-time, in response to user inputs, and to the environment. Every time computing has made a shift from passive to interactive, entire industries appeared that didn't exist before. We're standing in front of such moment again.
Over the last 6 months, we’ve assembled an all-star team with alumni from Apple, Meta, Google, Luma AI, Netflix, and Replicate. We're already partnering with some of the biggest names and labs in the world, and hundreds of developers are already building on Reactor.
The World Model era starts now.
Amber Yang retweeted
AI has gone from generating words to generating worlds.
@reactorworld is a real-time generative video platform built for developers. Their infrastructure layer enables users to work with scaling and production in mind. Co-founders @taiuti and @_bschmidtchen previously worked as technical leads on the Apple Vision Pro. Now, they're bringing their world model expertise to Reactor.
We’re thrilled to announce that we led their $59M Series A as they emerge from stealth, and we can’t wait to see how they continue building.
We believe world models have the potential to be the next wave of AI, as they let people step into and experience their creations.
Reactor is working to fortify the software foundation to keep up.
CC: Lightspeed Partners @buckymoore and @theamberyang
Reactor shows us what the future will look like đź’«
Real-time World Models are the next AI frontier.
Today, we're taking the first step towards this reality: our early preview lets you experience worlds generated in real-time, running on our global low-latency infrastructure.
Try it now: reactor.inc/