Restless. Irreverent. Greylock GP. @GreylockVC
San Francisco, CA
Joined March 2008
- Tweets8K
- Following3.5K
- Followers10.7K
- Likes15.2K
Pinned Tweet
I don't blog often, but when I do, I buy the domain names. 😜 TheNewMoats.com UnitOfValue.com VCrisks.com TheOnDemandEnterprise.com EvolutionOfCloud.com CastlesInThe.Cloud OpenSourceVsCloudCastles.com TheNewNewMoats.com
No company journey is easy and problem solving requires trust, communication, and a strong working relationship with the board and team. @martin_c_mao successfully navigated @chronosphereio through the chaos of Covid and into the AI era.
A founder can be right about the market, right about the product, and still run out of money before the bet pays off.
When runway gets short, founders and investors may see the same risks differently.
A down round buys time, but at a cost. Waiting preserves the chance for a better outcome, but also risks leaving the company with fewer options.
The hardest part is knowing how much confidence to place in each other's judgment. The numbers can tell you what's at stake. They can't tell you whether someone is seeing the business clearly or simply believing what they want to believe.
That understanding comes from years of working together, especially seeing how someone handles setbacks and bad news.
When the numbers turn ugly, the founder-investor relationship gets its real test.
@chronosphereio co-founder and CEO @martin_c_mao and Greylock partner @jerrychen revisit how they navigated some of the company's most consequential decisions.
0:00 Intro
01:03 Chronosphere's Early Growth
02:22 Building the Next Product
03:26 A Changing Market
04:54 A Critical Quarter
06:28 Weighing the Options
09:06 A Conversation After the Board Meeting
09:34 The Turnaround
Jerry Chen retweeted
A founder can be right about the market, right about the product, and still run out of money before the bet pays off.
When runway gets short, founders and investors may see the same risks differently.
A down round buys time, but at a cost. Waiting preserves the chance for a better outcome, but also risks leaving the company with fewer options.
The hardest part is knowing how much confidence to place in each other's judgment. The numbers can tell you what's at stake. They can't tell you whether someone is seeing the business clearly or simply believing what they want to believe.
That understanding comes from years of working together, especially seeing how someone handles setbacks and bad news.
When the numbers turn ugly, the founder-investor relationship gets its real test.
@chronosphereio co-founder and CEO @martin_c_mao and Greylock partner @jerrychen revisit how they navigated some of the company's most consequential decisions.
0:00 Intro
01:03 Chronosphere's Early Growth
02:22 Building the Next Product
03:26 A Changing Market
04:54 A Critical Quarter
06:28 Weighing the Options
09:06 A Conversation After the Board Meeting
09:34 The Turnaround
Jerry Chen retweeted
Nobody decides to lock their data in. It happens by default.
Start with a cloud warehouse like Snowflake, and you inherit its defaults: a proprietary storage format and catalog. That's fine until you want to bring a data science engine like Apache Spark or a machine learning engine like Ray to that same data. The same default that looked harmless at first is exactly what makes that very difficult.
That's the real trap in engine-first architecture. You don't pick the wrong engine and regret it immediately. You pick an engine, and its defaults quietly become your data's defaults too, before you've ever brought in a second one.
Flip the order: put the data in open formats first, and Spark, Ray, or whatever comes next all get to be a choice instead of a fight.
#Lakehouse #ApacheSpark
Jerry Chen retweeted
Around half of customers coming to niteshift.dev are coming to us after their internally built cloud coding agent from earlier this year failed to get real traction internally.
Most teams should be targeting > 50% of PRs from cloud agents, if not 80%
To be direct about it: every company that considers themselves a “tech” company is currently building a much worse version of something they could buy off the shelf, and furthermore they often have several teams competing on this front
Then they launch it after 3-6 months of building in a silo and can’t get their colleagues to use it for entirely predictable reasons
Coding agents and their adjacent products are among the most competitive markets of all time. The benefits of a highly competitive market accrue to customers by virtue of this intense competition among suppliers. Econ 101
Is the internal team working on your agent platform really keeping pace with the market? Even if they put in the same hours (they don’t), they don’t have the breadth of customers and hard contact with the market to get the feedback necessary to make the product any good
Big corporates that call their devs “IT” don’t delude themselves in this way. They buy the best, recognize their relative advantage (vendors will build almost anything for you because the competitive pressures demand it), and get back to work
Jerry Chen retweeted
Announcing Dagger 1.0 beta.
Replace Jenkins and GitHub Actions with CI that matches the speed and scale of coding agents without wasting compute or tokens.
Featuring: test-native CI, dev shells, local and remote execution, official modules, and more.
dagger.io/blog/dagger-10-bet…
Jerry Chen retweeted
Grateful to @evansammccann for the months of thoughtful reporting that went into this story
Building @8vc has been my life’s work so far. What means the most is seeing so many people I’ve been lucky to work with take the time to share their experiences. Thank you all!
The Quiet Force. My profile of @AlexKolicich is out.
Raised by a single mom in Hamilton. Google engineer. Peter Thiel’s math guy. Founding partner of 8VC.
“8VC is 8VC because of Alex.” I was told.
I spent months reporting his story in SF, Palo Alto and Orange County, talking with founders and the people who knew him at Clarium, Mithril and 8VC.
Highlights
- Peter Thiel playing blindfolded chess
- Turning down Geoffrey Hinton’s lab twice
- Dreaming about oil
- Back story on 8VC's investments in Quince, Cognition and Anduril
Link to read below.
Jerry Chen retweeted
Today we’re releasing OCRia-1, a frontier VLM for document parsing that sets a new SOTA in accuracy, cost, and latency.
It not only sets a new high score on ParseBench (87.4), but also leads in three 3rd-party benchmarks: olmOCR-bench* (83.5 overall, 92.9 on tables!), PulseBench-Tab (84.1), and fr-bench-pdf2md (72.1).
OCRia-1 is really, really good at parsing complex tables: nested headers, merged cells, irregular grids, scanned tables, and more. It is also fantastic at parsing complex unlabeled charts (line, bar, pie, technical diagrams), and preserving complex formatting. Most real-world enterprise documents contain at least one of the these, whether it’s a big tax withholding table or a public economic report.
It is a 4B model and outperforms all other models within a similar parameter class: LightOnOCR-3, Chandra-OCR 2, Mistral OCR 4.1, and more. It is also extremely fast.
All credit goes to our research team, who spent countless hours on curating a diverse dataset, tuning our training pipeline, and squeezing every last drop out of model optimization so that it can process billions of pages at scale.
OCRia-1 as a base is powering all of our modes within LlamaParse (cost-effective, agentic, agentic plus), so you can access its capabilities today!
If you’re processing millions/billions of documents at scale, we’re offering enterprises that book a demo up to $3k in credits for a guided evaluation.**
*olmOCR-bench scores are adjusted based on unifying rendering at 1,288px for all models
**enterprise POC credits apply to eligible organizations that book a demo
Jerry Chen retweeted
Watch the full episode youtu.be/0VbD6YYN-vM?si=c3vp…
Jerry Chen retweeted
Founders learn to obsess over product-market fit. Greylock Partner @jerrychen says to look just as hard at who should write the first check. He calls it founder-VC fit.
In 2018, Jerry read a technical blog about M3, an open source project built at Uber, led by engineers Rob Skillington @roskilli and Martin Mao @martin_c_mao. At the time, they hadn’t even decided to start a company.
What followed was a five-month courtship.
Jerry was trying to convince Martin and Rob to take his investment. As first-time founders, they were diligencing him just as closely.
Martin sent him an 85-page business plan to see who would really do the work, understand the technology, and align with where they wanted to take it.
Rob and Martin eventually started @chronosphereio, with M3 as its technical foundation and Jerry wrote the company’s first check.
Seven years later, the company was acquired by @PaloAltoNtwks for $3.35B.
For founders, choosing the early investor can shape the company’s trajectory for years. The relationship deserves as much diligence as the financing.
Full story in the comments.
Working with @martin_c_mao and @roskilli and the @chronosphereio team was a wonderful journey. Excited to have Martin share the story in this podcast!
Founders learn to obsess over product-market fit. Greylock Partner @jerrychen says to look just as hard at who should write the first check. He calls it founder-VC fit.
In 2018, Jerry read a technical blog about M3, an open source project built at Uber, led by engineers Rob Skillington @roskilli and Martin Mao @martin_c_mao. At the time, they hadn’t even decided to start a company.
What followed was a five-month courtship.
Jerry was trying to convince Martin and Rob to take his investment. As first-time founders, they were diligencing him just as closely.
Martin sent him an 85-page business plan to see who would really do the work, understand the technology, and align with where they wanted to take it.
Rob and Martin eventually started @chronosphereio, with M3 as its technical foundation and Jerry wrote the company’s first check.
Seven years later, the company was acquired by @PaloAltoNtwks for $3.35B.
For founders, choosing the early investor can shape the company’s trajectory for years. The relationship deserves as much diligence as the financing.
Full story in the comments.
Jerry Chen retweeted
Thank you to @eglyman @karimatiyeh and the @tryramp team for constantly being positive sum
Releasing the secret sauce behind Ramp's velocity, so every company can move fast, and save money
Jerry Chen retweeted
When attackers use new sophisticated #AI, enterprise defenders cannot rely on legacy #cyber with AI only bolted on top.
I’m excited about what @CylakeAI is building from the ground up, with customers keeping full control of their data and infrastructure.
Thanks René Bonvanie for the insightful conversation.
Cybersecurity has always been asymmetric: attackers only need to find one opening, while defenders have to protect everything that matters. Now, AI is changing the scale and speed of that equation.
@chandna, Partner at @GreylockVC and Cylake board member, believes we're entering an era of machines fighting machines, where increasingly sophisticated attacks powered by “bad AI” will require equally capable “good AI” on defense.
In Episode 6 of Inside Cybersecurity, Asheem joins @renebonvanie to discuss what AI means for the cybersecurity stack, why truly AI-native architectures need to be built differently, and what three decades in cybersecurity have taught him about the people and companies that successfully navigate major technology shifts.
🎧 Watch Episode 6 of Inside Cybersecurity: cylake.substack.com/p/good-a…
Jerry Chen retweeted
Last week, Databricks CEO Ali Ghodsi and I debated Iceberg and open table formats on a post. Then I caught myself.
Why are we debating each other while Snowflake, BigQuery and Redshift still default customers to closed formats?
Two years of Iceberg announcements. Still no public retirement date for their closed formats.
Open-source projects get scrutinized for interoperability on every release. The largest warehouses should meet the same bar.
We’ve spent a decade building open formats and data lakes. I think it’s time we asked.
ALT Anakin and Padmé meme. Snowflake, BigQuery and Redshift in 2024: “The future is open. Use Iceberg.” Data community: “So you’ll retire your closed formats?” The same warehouses in 2026 say nothing, and the data community asks: “You’re still planning to retire them... right?
Jerry Chen retweeted
Today we are open-sourcing 🥕Karotte, our framework for building RL environments.
We've used it for the past year to build MLE RL environments for frontier labs, and it's been hardened through 1M+ evaluation runs and red-teaming.
Jerry Chen retweeted
This is why you should run your agents in Linux VMs in the cloud with a team (niteshift.dev) that has obsesses over benchmarking performance and pushing the major sandbox providers on faster and faster guest VM perf
Jerry Chen retweeted
To squeeze the performance out of our runtime components, we implemented a good amount of concepts from - SEDA: an architecture for well-conditioned, scalable internet services dl.acm.org/doi/10.1145/50205…
The seminal work at MIT around writing high-performance server software and the gift that keeps on giving.
Jerry Chen retweeted
Lots of really great talks at @modal Runtime. My favorite so far is the talk by @smehmood (cofounder of @niteshiftdev). He argues that you should treat agent sandboxes as a lossy cache, not the source of truth. Persist the essential state, let agents rebuild after failures, and keep the speed of running the harness inside the sandbox. This gives you reliability without sacrificing latency. Also really interesting hearing how @ListenLabs is building autoresearch agents and hillclimbing on their evals with Niteshift.
Jerry Chen retweeted
Today we announced “BYOC 2.0” at Modal Runtime for niteshift.dev!
BYOC 1.0 meant “runs in your AWS account” and so BYOC 2.0 means “runs in your Modal account”
Today, @modal is the only cloud provider with top-tier products for both sandboxing and inference — the two most important primitives for agents.
We think BYOC 2.0 unlocks fundamentally new use cases for cloud coding agents. Your agents can finally process sensitive customer data without it being stored outside of your account. Customers are using niteshift.dev today to resolve production issues and improve their agents against real production data.
If you want a cloud coding agent platform that keeps your most sensitive data inside your Modal account’s security perimeter, please DM me for access to our Private Beta!
Lots of really great talks at @modal Runtime. My favorite so far is the talk by @smehmood (cofounder of @niteshiftdev). He argues that you should treat agent sandboxes as a lossy cache, not the source of truth. Persist the essential state, let agents rebuild after failures, and keep the speed of running the harness inside the sandbox. This gives you reliability without sacrificing latency. Also really interesting hearing how @ListenLabs is building autoresearch agents and hillclimbing on their evals with Niteshift.