Offical twitter account of Omri Casspi, Swish Ventures and a former NBA player.
🇮🇱🇮🇱🇮🇱
Joined July 2010
- Tweets4.4K
- Following672
- Followers79.6K
- Likes2.6K
Pinned Tweet
We’re proud to announce Swish III, a $250M early fund, bringing Swish Ventures to $800M in total AUM.
I started this firm to spend my career supporting the artists who build and shape the future.
The technology shift underway is unlike anything we’ve seen. We plan to stay true to our colors: few investments a year, real concentration and partnership.
Grateful to the founders we get to work with, who inspire so much of how we work, and to our LPs for walking this path with us.
We are more locked in than ever.
Omri
Omri Casspi עומרי משה כספי retweeted
Stable php2shell aslr bypass exploit.
@ace__pace @Tenzai_Labs is cooking.
Sign up at tenzai.com to test your infra.
Boom! 🎆
Runs stable in our eval env!
@Tenzai_Labs 🤖
ALT Tenzai screenshot showing the agent successfully exploiting PHP to shell
ALT Log from the Teznai agent showing it using the attack, reproducing it, and checking health
Shift-1 is unique because it routes requests, not tasks. this allows it to adapt much more effectively to long running agentic workloads.
Introducing Shift-1, a frontier coding router trained on AC2 that reduced coding agent cost by up to 55% with no noticeable drop in quality. Try it at shift.ac.
Omri Casspi עומרי משה כספי retweeted
A few months ago, I shared that I was stepping into my next chapter.
Today, I’m excited to share that we’ve launched @trace_cap. 🚀
Trace is a private investment firm. We make concentrated investments in a small number of exceptional businesses, partner closely with them as they scale, and own them for the long term.
Trace draws on more than fifteen years of working alongside founders and over a decade of investing. Now we get to build an entire firm around that work, with an incredible team.
I’m grateful to the founders, partners, investors, colleagues, and friends who have trusted us and helped us get here.
We’ve been quietly building and investing over the past several months. Look forward to sharing more about Trace and the companies we’re backing.
Omri Casspi עומרי משה כספי retweeted
Nemotron 3.5 Super VL by @NVIDIAAI is now supported for both training and inference on AC2.
In early post-training runs for our customer, we've been impressed with the quality of the base model for its inference efficiency.
Omri Casspi עומרי משה כספי retweeted
We received early access to the latest @NVIDIAAI Nemotron 3.5 Super VL and we are really impressed with the post-training results using Era
era.eon.io
Nemotron model matched frontier models at 88.9% accuracy, with 27× lower cost per correct answer using Spot GPUs.
We then used Era-generated data to post-train Nemotron for just $609, improving its ability to recognize when information is missing while further reducing inference costs.
The most exciting part to me is the potential to close the loop between benchmarking and post-training. Find where a model fails, generate the data to improve it, and test the new version against an entirely different simulated company.
This could fundamentally change how enterprises customize open models for their specific needs.
With Era, customers can apply this workflow to evaluate, improve, and deploy models tailored to their own enterprise environments.
Great collaboration with the NVIDIA team, and much more to come!
Full research:
eon.io/blog/improving-nemotr…
Omri Casspi עומרי משה כספי retweeted
Introducing Spork, an efficient post-training recipe which increases our solve-rate on chess puzzles by 100% when compared to our best GRPO setup.
It uses a GRPO warmup, splits task-specific datasets further into partitions, trains specialists, and then uses multi-teacher on-policy distillation. Like a Spork, it leverages gains by simply consolidating specialists.
Omri Casspi עומרי משה כספי retweeted
AC2 now supports decision model training. Teams can build custom Jev-like models for use cases like request routing, content classification, and anomaly detection.
Reach out to get started with your own classification task!
You can now train your own decision model in our platform AC2! Jev-like models, custom fit for your use case :)
We implement decision models through a language-model backbone with a small learned output layer called a decision head. The backbone processes the input, question, and answer choices, and the head produces scores which are softmaxed to produce a probability distribution over answer choices.
Our software interface makes it easy to add a variable-sized decision head to frontier open-weight models. Choose any open source model, and train it as a decision model using Brier loss or cross-entropy loss. Training updates both the decision head and the language backbone.
We tried it on 50k sampled training datapoints from the Civil Comments dataset, starting from Perplexity’s pplx-decider-v1-27b, and F1 for detecting toxic comments increased from 0.482 to 0.677 on 20k sampled eval datapoints. The training job took less than an hour on four B300 GPUs.
Reach out to get started with your own classification task: docs.appliedcompute.com/plat…
Omri Casspi עומרי משה כספי retweeted
Thank you to Shaun and Ron, and to everyone who came on Monday.
To the 600+ who asked for a seat we didn't have, we owe you a bigger room, and until then Era.eon.io has space for all of you
More than 700 people asked for a seat at the fireside on Monday night, in a room that holds 80.
The ones who got in heard @shaunmmaguire say that before he'd trust an agent with sensitive data, the first thing he'd want is a backup.
(We promise we did not write his answers.)
Ron Kimchi, our Co-Founder and CTO, was asked where engineering effort gets wasted in AI, and his answer was "talk less, build more."
Era went live that same morning and is free at era.eon.io.
Thank you to everyone who came and shared this amazing night with us
Omri Casspi עומרי משה כספי retweeted
@NoetiveAI research blogpost ... building Intelligence for the physical economy with recursive self-improvement (RSI)
ofir is one of the most important voices in the Israeli tech ecosystem, so good to have him here on X
You can’t safely test an enterprise agent on a real company’s data.
So we built a company for it to work in.
Era is live today, and it’s free.
It generates a complete simulated enterprise that behaves like a real one across Salesforce, Slack, Jira, Zendesk, Gong, Deel and more, along with cloud databases and storage. Agents interact with it through live MCP and API interfaces.
People leave. Deals change. Records get duplicated. Permissions differ across systems. And because Era generated the company, it knows the exact ground truth.
Test, benchmark and improve agents against realistic enterprise workloads, use the failures for targeted post-training, then rerun the same environment to measure the impact.
Huge thanks to our research partners @NVIDIA, @Decart, @Composio, @openlayerco, @Deel, @Eragon and @Plurai, with more coming soon.
enterprise agents need something better than toy datasets and empty sandboxes to test against.
You can’t safely test an enterprise agent on a real company’s data.
So we built a company for it to work in.
Era is live today, and it’s free.
It generates a complete simulated enterprise that behaves like a real one across Salesforce, Slack, Jira, Zendesk, Gong, Deel and more, along with cloud databases and storage. Agents interact with it through live MCP and API interfaces.
People leave. Deals change. Records get duplicated. Permissions differ across systems. And because Era generated the company, it knows the exact ground truth.
Test, benchmark and improve agents against realistic enterprise workloads, use the failures for targeted post-training, then rerun the same environment to measure the impact.
Huge thanks to our research partners @NVIDIA, @Decart, @Composio, @openlayerco, @Deel, @Eragon and @Plurai, with more coming soon.
this is a must watch
My conversation with @AlexTaubman, co-founder and CEO of Long Lake.
Long Lake is 3 years old, has 30,000 employees, $4 billion in revenue, and just took Amex Travel private.
Alex and his partners have spent the last three years building Long Lake with almost no public profile. They started in HOA management and expanded into five service lines. Their latest deal: $6.3 billion for American Express Global Business Travel.
Long Lake buys great businesses and makes them better with AI. Their engineers sit alongside the people doing the work, understand what slows them down, and build tools that give them more time to serve customers.
We talk about what he learned growing up in a family of entrepreneurs, why Long Lake built its own AI platform, what Henry Singleton and the Rales brothers taught him about compounding, and why the best partners make you more ambitious.
Alex wants to spend the next 50 years on this. We talk about how far it could go.
0:00 The $6.3 billion Amex GBT deal
1:21 Bringing AI to the real economy
6:17 Why greater productivity can mean more jobs
11:29 A prepared mind and a tight acquisition filter
16:01 Nexus and the AI platform behind Long Lake
17:32 Put the engineers where the work happens
20:36 Why own the businesses instead of selling software?
22:28 Build tools people want to use
25:24 A high bar for talent, acquisitions and technology
28:03 What his grandfather taught him about removing friction
36:40 Learning capital allocation at Goldman Sachs and Oaktree
41:01 Buying founder-owned businesses with permanent capital
44:35 How the Long Lake team came together
47:26 The best partners make you more ambitious
49:28 More than $4 billion in projected revenue
53:32 Why going public could expand Long Lake’s ambition
54:23 Henry Singleton and staying flexible
55:57 John Malone and the businesses that will endure
57:57 Why Long Lake doesn’t plan to sell
1:00:13 Could Long Lake become a trillion-dollar company?
Includes paid partnerships.
Omri Casspi עומרי משה כספי retweeted
@Tenzai_Labs finished #1 on @Hacker0x01's VDP business leaderboard again this quarter.
Last Q we ran on US frontier models; this quarter we used GLM 5.2 & 5.3, open-weight models from z.ai.
We still came out on top, with just 3.5% false positives, and roughly 30% lower cost per report.
For us, it shows the harness matters more than the model underneath it - we built Tenzai to work with a choice of models.
The bigger point is the warning. Open weight models are freely available and highly capable to power offensive AI operations at scale. Attackers can and will build malicious agents with these models.
@nytimes covered it today: nytimes.com/2026/10/01/scien…
If you're defending applications, networks and infrastructure the window to test them continuously, before someone else does, is closing.
Full write-up on how we did it: tenzai.com/blog/an-open-weig…
#1 on @Hacker0x01's business VDP leaderboard - for the second quarter in a row, and this time using open weight model that anyone can download.
So Tenzai is #1 on @Hacker0x01's business VDP leaderboard - for the second quarter in a row.
Here's the news: this time we did it using an open-weight model anyone can download. Same result as the frontier model, ~30% cheaper per report, and only 3.5% false positives.
And here's the real news: These models are highly capable to power offensive AI operations at scale. Attackers can and will build malicious agents with these models.
The write-up: tenzai.com/blog/an-open-weig…
Omri Casspi עומרי משה כספי retweeted
13,000+ internal images. 300+ organizations. 900+ repositories. Publicly exposed on GitHub.
New research from Glow Labs found AI coding agents exposing customer records, internal financial systems, unreleased product features, and other sensitive development work.
No attacker involved. No credentials stolen. Just an agent trying to finish the job it was given.
That’s what makes PixelLeak important. AI agents can find their own ways around technical limitations, creating security risks that existing controls weren't designed to catch.
Our latest Glow Labs research breaks down what we found, how PixelLeak happens, and what security teams can do to find and prevent this exposure.
Read the full research: glow.io/blogs/how-ai-agents-…
#InfoSec #AISecurity #AppSec #EndpointSecurity #GlowLabs
Omri Casspi עומרי משה כספי retweeted
Today in Bloomberg Deals: Long Lake CEO outlines his AI strategy, and Dangote targets 10 million investors with IPO marketing blitz. bloomberg.com/news/newslette…
