Crane retweeted
Come to the next ML Breakthroughs Breakfast on Tuesday in London hosted by @RMBattleday and @jcrwhittington luma.com/ml-breakthroughs-br…
Crane retweeted
"If MongoDB and ClickHouse had a baby, it would look a lot like RawTree.”
@bnevilleoneill and Enzo judged the Long Horizon Agents hackathon and saw how builders dumped raw events, JSON, and agent telemetry into RawTree with no schema, then queried it with SQL.
Crane retweeted
Great conversation with
Anandamoy Roychowdhary (MD, Crane Venture Partners) and Joanna Ossinger (Managing Editor, APAC Digital, CNBC) at SuperReturn Asia in Singapore on The New India Story: Investing in AI and Deep Tech.
We spoke about how deeptech founders in India are already innovating fast, capital is flowing in, and winners are emerging from India. Getting there will take patient capital and long-term thinking.
We mapped the three technology waves India has gone through: IT services and BPO, which made India the world's technology partner, then the second with digital startups building products at scale for India.
The third is beginning now, as India builds the underlying technology itself in semiconductors, space, AI, defence, biotech and more. It builds on decades of software capability, and increasingly hardware.
The point I made is that this third wave differs in one crucial way.
Policy sat on the sidelines in the first two waves, but in the third it is existential, because frontier tech needs an enabling ecosystem.
That's why bringing founders, VCs, policymakers, academia and ecosystem enablers to the same table is at the heart of what we do at Startup Policy Forum.
The key takeaway from the panel: the India story is real. With the right enabling environment, it will take off.
@smdcmc @crane_vc @CNBCi @CNBC @CNBCtech @ossingerj @scott_sage @krishnav @Sid_1_0 @SPF_India
Crane retweeted
Today, we are launching the world's first decentralized security and authorization layer designed specifically for securing agentic AI. The agentic era requires a new paradigm to secure agentic AI - and we’re building it here at Outerlimit. outerlimit.com/company-news/…
Crane retweeted
Thrilled to finally be able to talk about @OuterlimitAI
They emerge today out of stealth with $16m in pre seed.
The first time I met Tony, we talked about almost everything but security - the human side of building a business, the scar tissue and lessons learned building Egress, and so much more.
Tony and his cofounders Neil and Peter are truth seekers. It has been incredible watching them operate and building the team to date.
The future needs a completely new security layer - and thats exactly what they are building.
Today, we are launching the world's first decentralized security and authorization layer designed specifically for securing agentic AI. The agentic era requires a new paradigm to secure agentic AI - and we’re building it here at Outerlimit. outerlimit.com/company-news/…
Crane retweeted
Forward Deployed Engineer a pie de calle =>
javisantana.com/fde/
He intentado resumir 8 años de aprendizajes intentando no ser demasiado sofisticado ni añadir demasiado fuego artifial.
Está muy orientada a empresas de producto B2B, aunque creo que hay algunas lecciones que sirven en otros contextos.
Gracias a todo el mundo que lo ha revisado
Crane retweeted
The work is just getting started.
We upgraded the models powering PR scans and Whitebox Pentests on Hacktron last night. Early results indicate far fewer false positives and more true positive findings flagged that were previously missed.
On Monday we're also dropping a new UI for Whitebox Pentests that allow you to see what actions our dynamic / runtime validation agents are performing and verify it's browser use behaviour.
DMs open for whoever may be affected by HEIF Heist or are thinking about better security for your software.
Crane retweeted
Yesterday, our Founder, Alain Richardt, sat down with @GarthBray from the @nzherald at the GD1 (Global From Day One) "Believe Bigger" Conference.
They discussed how we're discovering synthetic alternatives to reduce reliance on geopolitically constrained supply chains in months, not decades.
"3 random dudes”
1. hacked apple, again and again.
httpvoid.com/Apple-RCE.md
httpvoid.com/Hello-Lucee!-Le…
httpvoid.com/Hacking-Apple-w…
2. your github enterprise is our github enterprise.
httpvoid.com/GitHub-Enterpri…
3. get a discord message from me, get pwned.
hacktron.ai/blog/discord-rce
youtube.com/watch?v=R3SE4VKj…
4. oh yeah, at one point we basically had shells across the electron ecosystem. check the DEF CON research.
media.defcon.org/DEF%20CON%2…
5. your supabase database is my database.
hacktron.ai/blog/supapwn
6. we got the posthog prod database.
hacktron.ai/blog/posthog-rce
7. react2shell? vercel paid us $170k for helping secure their waf.
hacktron.ai/blog/react2shell…
8. your palo alto vpn is my vpn.
hacktron.ai/blog/cve-2026-02…
9. ai ides? we got shells for you, antigravity
hacktron.ai/blog/hacking-goo…
10. windsurf rce.
youtube.com/watch?v=23Mz7qcR…
11. turning cluely into malware.
hacktron.ai/blog/hacking-clu…
12. ai browsers? sure, uxss: your perplexity browser is my browser.
hacktron.ai/blog/perplexity-…
13. openai atlas too. kinda uxss
hacktron.ai/blog/hacking-ope…
14. hey, it’s not even our first time hacking discourse.
projectdiscovery.io/blog/dis…
15. adobe coldfusion: pre-auth rce. because apparently we needed another one.
projectdiscovery.io/blog/ado…
there’s a lot more. go dig.
anyway, yes: “3 random dudes.”
and @HacktronAI is full of more random dudes like these.
Crane retweeted
A bug-hunting independent security research team used Anthropic’s Claude software to gain access to an OpenAI employee’s ChatGPT account, giving them a way to read and suggest changes to the company’s private cache of software. on.wsj.com/4h8vaBP
Crane retweeted
Precision-3 is live. Diarization error down 10.4% vs Precision-2 (14.35 vs 16.02 DER across 15 benchmark datasets), with most of that gain coming from better speaker attribution: the model gets the wrong speaker 18.7% less often.
Crane retweeted
How many companies can you hack through an image parser?
We spent the last few months finding out.
OpenAI was one.
So were Slack, Meta, GitHub Enterprise, Rails, Next.js, ImageMagick, and many more.
We called it HEIF Heist. 📷
Crane retweeted
We found some stuff. Actually, quite a lot of stuff.
Back in July, our team @rootxharsh, @S1r1u5_, and @iamnoooob managed to compromise multiple OpenAI employees’ ChatGPT accounts.
In theory, any ChatGPT connector can be affected. This includes Slack, github, email, and other services. 🧵
Crane retweeted
$5.3M seed for @ExpanseCompute , led by @crane_vc with PXN Ventures. Expanse predicts exactly what an AI workload needs before it runs, so teams get as much as twice the work from the same hardware. Full story: expanse.sh/blog/expanse-rais…
Congrats!!
We’ve raised a $5.3M seed for @ExpanseCompute, led by @crane_vc with PXN Ventures.
Expanse predicts exactly what an AI workload needs before it runs, so teams get as much as twice the work from the same hardware.
Why we started it: expanse.sh/blog/expanse-rais…
Crane retweeted
~500M of fresh capital raised to back the greatest founders in all the different parts of the globe we can find them . as we like to say at @crane_vc " first to believe and the last to leave" @scott_sage @krishnav @Sid_1_0 @anshulbhide
We Believe the Future Belongs to Synthesists.
Today we announced $484 million across multiple investment vehicles - nearly half a billion dollars to invest from inception through seed across Europe, the US, India, and Asia-Pacific.
We are excited about the capital. We’re even more excited about the much bigger search space it creates.
Specifically, the ability to find more synthesists.
A synthesist isn’t a generalist. They go deep in one discipline but refuse to stay inside its boundaries. The physicist who ships software. The biologist who thinks like a machine learning researcher. The materials scientist who understands computation.
We think more of the world’s important companies will be built by people like this.
For most of history, becoming fluent in a second discipline was expensive - another degree, another career, or an unusual set of life’s stepping stones (h/t @kenneth0stanley). AI is collapsing that cost, at the exact moment many of the hardest problems in science are becoming computational problems too.
The interesting opportunity isn’t on either side of those combinations. It’s in the space between them.
The next great category may not exist yet because the disciplines required to create it have only just learned how to speak to each other.
This is the idea behind our Meridian Principle: human progress accelerates when extraordinary people connect ideas across disciplines, cultures, and generations.
@BoydVarty writes in The Lion Tracker’s Guide about following the track rather than fixating on the destination. Venture at its best has that quality. You follow exceptional people. You notice unusual intersections. You stay curious about things that don’t yet fit a category. Occasionally those tracks converge into something much bigger than anyone planned.
Consensus is good at pricing categories that already exist.
Venture returns come from recognising the ones that don’t.
The search and adventure space just got a lot bigger.
We intend to spend the next chapter following the most interesting tracks and stepping stones we can find. Thank you to all of the Founders and LPs that have entrusted us to be their partners.
Crane retweeted
We Believe the Future Belongs to Synthesists.
Today we announced $484 million across multiple investment vehicles - nearly half a billion dollars to invest from inception through seed across Europe, the US, India, and Asia-Pacific.
We are excited about the capital. We’re even more excited about the much bigger search space it creates.
Specifically, the ability to find more synthesists.
A synthesist isn’t a generalist. They go deep in one discipline but refuse to stay inside its boundaries. The physicist who ships software. The biologist who thinks like a machine learning researcher. The materials scientist who understands computation.
We think more of the world’s important companies will be built by people like this.
For most of history, becoming fluent in a second discipline was expensive - another degree, another career, or an unusual set of life’s stepping stones (h/t @kenneth0stanley). AI is collapsing that cost, at the exact moment many of the hardest problems in science are becoming computational problems too.
The interesting opportunity isn’t on either side of those combinations. It’s in the space between them.
The next great category may not exist yet because the disciplines required to create it have only just learned how to speak to each other.
This is the idea behind our Meridian Principle: human progress accelerates when extraordinary people connect ideas across disciplines, cultures, and generations.
@BoydVarty writes in The Lion Tracker’s Guide about following the track rather than fixating on the destination. Venture at its best has that quality. You follow exceptional people. You notice unusual intersections. You stay curious about things that don’t yet fit a category. Occasionally those tracks converge into something much bigger than anyone planned.
Consensus is good at pricing categories that already exist.
Venture returns come from recognising the ones that don’t.
The search and adventure space just got a lot bigger.
We intend to spend the next chapter following the most interesting tracks and stepping stones we can find. Thank you to all of the Founders and LPs that have entrusted us to be their partners.
Crane retweeted
I'm proud of the speaker line-up our team has put together for the inaugural Long Horizon physical AI summit in SF Oct 20-21.
Join us to see talks from @carolina_parada @deepakpathak @alexgkendall @Ben_Burchfiel @ipoupyrev @QuanVng @PimDeWitte @SebastianThrun @qasar @AnguelovDrago, and many others.




























