[brief update] we have a new set of funds for @haystackvc announcing haystack 8 plus more, but it's really about my teammates @aashaysanghvi_ & @DivyaDhulipala thank you to all of our frequent ecosystem co-conspirators, our LPs, our co-investors, our founders. i'm very grateful 🙏 and this is the most fun role in the world! semilshah.com/2026/04/26/new…

Buying and deploying proprietary data is critical for the next frontier. But when it comes to buying real data from real people, trust is the first step in making sure the largest supply of real-world data can move into training as efficiently as possible. @integralprivacy ensures that your data will be monetized the right way. useintegral.com/get-paid

We built a time machine for the web. Introducing Exa Snapshot: an index of 400 billion historical snapshots of webpages that lets you search as if it's the past. Snapshot is already being used for backtesting prediction models, RL at labs, exploring the pre-AI web, and more.

Today we’re announcing we’ve raised $20M in seed financing led by @QuietCapital , with participation from @GradientVC , @haystackvc and @CompoundVC and other amazing folks to build the verification layer for AI-driven biology. We are also delighted to be welcoming @vqctran who is joining us from @GoogleDeepMind as a co-founder. AI is not going to cure all diseases without a verification substrate - a scalable environment where models can test predictions and learn against living human biology. AI is becoming increasingly capable of generating biological hypotheses, therapeutic candidates and experimental designs. But generating ideas is advancing much faster than our ability to test what those ideas will actually do in human biology. Existing systems force a tradeoff: the most biologically relevant approaches are slow and expensive, while faster in vitro and computational systems often lack the fidelity needed to capture human response. Polyphron’s bet is that tissue is the fastest, cheapest, most parallelizable verification substrate that still contains the biology that matters and we believe the path to biological superintelligence runs through closed-loop interaction with these living and simulated human systems. Thank you to @axios for covering the announcement. Link to our new website in the thread.

Excited to share the @Nominal_io x @aerovironment partnership - one of our biggest yet. Driving insights across the entire hardware development lifecycle with an intelligent platform. All systems Nominal! 🚀

exciting!

Beautiful, precise engineering.

Private school education if you’re having kids soon will cost roughly $1.5-2M all in in the Bay Area, $2-2.5M if you include college.

Think big. That’s always been a driving motivator for us, and now we have a partner to match our ambition. Today, we’re proud to announce Applied Intuition’s strategic collaboration with HUMAIN to deploy physical AI across Saudi Arabia, starting with autonomous trucking. This is a major step toward our goal of making a billion machines intelligent. Together, we will deploy thousands of autonomous trucks throughout key Saudi logistics corridors by 2030. This will make it the largest autonomous trucking network in the world, and our Self-Driving System and Vehicle OS technology will be the centerpiece of this collaboration with HUMAIN. This is physical AI at national scale. appliedintuition.com/blog/ap…

Introducing BioSecBench-Function, a verifiable benchmark for testing whether AI agents can infer the functional properties of viruses, bacteria, and toxins from data. Opus 5 with Claude Code leads in endpoint pass rate (50.4%), while Grok 4.6 with Grok Build leads in overall pass rate (44%) when refusals are counted as failures. The benchmark contains 111 deterministic evaluations spanning viral, bacterial, and toxin systems. It covers five primary threat axes: transmissibility, immune escape, virulence or toxicity, drug resistance, and persistence or fitness. Each evaluation is grounded in a published study or dataset and reviewed by domain experts. Tasks draw on deep mutational scanning, Tite-Seq, surface plasmon resonance, X-ray crystallography, free-energy calculations, and related assays. Agents receive the relevant data and produce structured answers graded against study-derived ground truth.

In 2013 I could take an Uber from SF to Emeryville everyday for $6. Using AI personal agent apps right now has that same feeling. This cannot be the real price.

Legend of @breeves08 just beginning!

Both valued at $2.5 Billion today: Linear, having crossed $100 million ARR with 177% net revenue retention. Instinct, the viral AI assistant that launched last week.

Some of the earliest money in Cursor (the largest vc-backed acquisition ever) turned a $750,000 investment into a ~$1 billion return. @davidtisch and his midas touch at Box Group strike again! Great profile here of one of the best early stage investors who wins while going against the grain @agarfinks fortune.com/2026/08/26/david…

This is blowing my mind: BART spends $1.10 in electricity per passenger ride, average ride is 15 miles. My Tesla Model Y is far more efficient: it costs $0.76 to go 15 miles. And it can carry 5 people, so could actually be far more efficient if not traveling solo.

I'll be going on TBPN today. I'm at the MTS Live studio. I just announced a new $2.42B round on Sourcery. I'm on a redeye to London for 20 Minute VC. Just did the MOTS Pod pushup challenge. I'm whiteboarding with Dwarkesh. Just did Senra. I'm on Invest Like The Best.

I’m super excited to announce Synthefy’s $6.5M seed round, led by Wing Venture Capital and joined by Haystack, Samsung Next, Canonical, Lightscape, and angels and operators from OpenAI, Microsoft, and Meta. Setting the price of every Uber ride in real time. Spotting the early signs of disease from thousands of patient measurements. Blocking a fraudulent payment before money moves. Keeping the power grid stable as supply and demand change every second. These decisions are not waiting for a better chatbot. They require Machine Intelligence, systems that can learn from massive volumes of structured data and make predictions beyond human scale. At Synthefy, our vision is to advance Machine Intelligence, enabling enterprises and users to make predictions on their troves of structured data that enable them to make critical business decisions. Nori is our first foundation model in a class of AI that will transform prediction. It already unlocks brand new ways of looking at prediction. But this is only the beginning of our journey towards zero-shot prediction that outperforms any bespoke ML. Our mission is to give customers predictive answers from their structured data that they cannot get from today’s tooling, 100× faster and at one-tenth the cost of bespoke ML. We aim to make Nori the universal prediction engine that enterprise software and agents call whenever they need to forecast, score risk, or make a decision from structured data. I’m deeply grateful to my co-founders, the Synthefy team, our early users, and investors. Special thanks to @ai for backing us from day -10, when I was still learning what PMF meant, and to Gaurav (Wing) for seeing the ambition and helping us take it further.

We are announcing our $6.5M seed round led by @Wing_VC, with @haystackvc, @samsungnext, Canonical @ai, @LightscapeVC, and angels from OpenAI, Microsoft, and Meta. We started Synthefy around a simple observation: most data that runs the economy is not text. It is transactions, sensor readings, trades, inventory, customer records, and time series. Yet every prediction problem built on that data, from forecasting and fraud detection to pricing and predictive maintenance, still requires a bespoke pipeline and months of data and ML work. We are building foundation models for structured data to change that. Our first model, Nori, puts predictive capability into its pretrained weights, so customers do not need to train and maintain a separate model for every dataset and use case. Our mission is to give customers predictive answers from their structured data that they cannot get from today’s tooling, 100× faster and at one-tenth the cost of bespoke ML. We are grateful to our team, early users, and investors for believing in our mission. Try Nori at synthefy.com/product/tabular. Full announcement in the comments.

I now ignore all inbound decks & pitches that are 90+% AI. Initially I was open-minded, but as the aphorism goes: how you do anything is how you do everything. If someone is sloppy or lazy with a high leverage fundraise, it's hard to build conviction on execution in other areas.

Legendary investment for @deedydas Maybe 1st of venture career? Huge 🎉 sir!! 14x 1700% IRR In 1.5 years

At @Nominal_io, we are committed to delivering the best-in-class experience for every part of a hardware engineer’s day. Turning human expertise into efficient, repeatable, scalable, and machine-actionable operations. Data-driven procedures (and yes, the agents that interact with them too—more on that soon) are one important piece of the puzzle. They’re powerful and flexible, and they respect the complexity of our users’ work. And they live in the same place as your critical manufacturing, test, and operations data. Build procedures on Nominal! ⚙️