Discover more from Liberty’s Highlights
211: Crowdstrike Q3, AWS re:Invent Keynote, Microsoft's Zoomier Teams, Facebook + AWS, Nvidia + ARM FTC Trouble, The Oil Drum (RIP), NFTs & Gaming, and Starlink
"I just don’t see how that works without ruining games"
If you put a spoonful of wine in a barrel of sewage, you get sewage. If you put a spoonful of sewage in a barrel of wine, you get sewage.
(I wonder if Munger had read that one and it inspired his famous “raisins” quip..)
🥇🥇vs 🥉🥉 Explained to my 7yo the concepts of win-win, win-lose, and lose-lose.
I don’t think it’s very hard for a kid to understand, yet I don’t think it’s taught at a young age nearly widely enough. He should be coming home from school with storybooks and homework about this stuff. It’s as important as anything else!
⬜️ ⛓ 🎮 I was thinking about the use case of using NFTs on blockchain to have portability for digital goods between games (I won’t say the m-word)...
Knowing a bit about game design and how hard it is to balance gameplay, tune things so they’re fun, create a very specific aesthetic and ambiance, etc. I think it'll be very hard to have portable digital objects between games (weapons, armor, gadgets, etc).
I just don’t see how that works without ruining games (or at least making them worse)…
Who wants to play on a Call of Duty server where some guy brings in his BFG9000 from some Doom game? ☠️
So it'll probably be cosmetic stuff, which makes it a lot more limited... It’s not nothing, plenty of character skins and clothes and such are sold every day, but far from the dream of having that special unique assault rifle that you bring from game to game.
And even if you could, each game-maker would have to integrate that weapon into their game pretty carefully for it to work and be fun and look good, and what are the incentives for them to do so — to allow you to give your money to someone else rather than them for items in your game?
💡 So many of my ideas for things to write about pop up when I’m not trying to write.
I think it’s because most of the time, I’m having input coming in — reading stuff, listening to podcast — and it’s when I can’t do that, like in the shower or mowing the lawn or unloading the dishwasher, that the flow reverses and the ideas start coming out.
I’ve written before about the importance of having some time without input so you can think your own thoughts, process things, and figure out what you actually think.
I should probably try to create more of that space for myself…
It’s easy to forget when there are so many browser tabs open, so many books on the Kindle, so many podcasts in the queue..
💚 🥃 If you like the food, don’t forget to let the chef know 👨🍳
Liberty’s Highlights is reader-supported. To support my work, consider becoming a paid subscriber. 💙
Investing & Business
Crowdstrike Q3 Highlights
Just some stuff that stood out to me. CRWD 0.00%↑
ARR growing 67% at $1.51bn run-rate (!!!)
FCF margin 32% (getting harder to re-invest all those gross profit dollars at 79% gross margin… the absolute numbers added every few months are getting pretty bonkers).
Non-GAAP operating income grew 168% (not that anyone is looking at that)
Added 1,607 new subscription customers, up 75%, Total of 14,687 subs (which may seem high, but the biggest security companies are in the 100,000s range)
CrowdStrike’s subscription customers that have adopted 4 or more modules, 5 or more modules and 6 or more modules increased to 68%, 55%, and 32% (extra modules are almost pure gross margin, once their agent is installed and the CAC has been spent on getting the relationship…)
From the transcript of the call (not afraid to name names of competitors):
This quarter, our win rates increased across the board, and we saw a record number of wins against both legacy and next-gen vendors with SMB, mid-market and large enterprise customers. [...]
We also landed a record number of wins and displacements over a recently public next-gen vendor, SentinelOne. To be clear, we define a displacement as removing an incumbent's product and replacing it with Falcon.
Let me share a recent example with one of the largest nonprofit hospital systems in the U.S. which had initially chosen the recently public next-gen vendor based on price and promised features that were never delivered. Just a few months into their multiyear contract with the other vendor, this organization realized the product fail to scale, cause major performance issues, prohibited critical processes from functioning properly and drove significant friction within the organization and its subsidiaries.
That is when they turned to CrowdStrike, purchased multiple modules in a multimillion dollar ARR deal and realized immediate improvement, gaining up to 30% performance increase on their servers alone and greater efficacy without intrusive false positive.
Here’s the arrows for Microsoft and Symantec:
We continue to see success winning new customers among our larger competitors as well. Take for example a leading health care system that was using a combination of 2 legacy vendors, Microsoft and Symantec when it was hit with a massive ransomware attack that disrupted their business. This organization turned to CrowdStrike's renowned incident response services team to remediate the breach. They also came to realize that while Microsoft attempts to check most of the boxes on paper and appeal to the CFO office, in reality, when every second counts, they needed CrowdStrike and standardized on Falcon Complete.
Don’t focus on just malware:
Cyber adversaries are increasingly attempting to accomplish their objectives without using malware. As we cited in our 2021 Threat Hunting report, based on recent customer data indexed by Threat Graph, 68% of detections analyzed were not malware-based. This is why companies need to employ a holistic breach provision strategy rather than overly relying on malware prevention regardless if it's legacy or next gen.
Public sector business should be a nice tailwind thanks to a renewed focus on cybersecurity, and a big reference account win at CISA:
record quarter in the public sector, including wins with educational institutions, states and local governments and the U.S. federal government. Our performance on this front is highlighted by a large win with CISA, the Cybersecurity and Infrastructure Security Agency, to secure a significant portion of endpoints and workloads for multiple federal agencies as they operationalize the White House executive order to improve the nation's cyber resiliency. [...]
We believe this significant win will create new opportunities and unlock additional business within the massive U.S. federal government.
They were previously very very excited about protecting cloud workloads, a segment that seems to be growing very fast for them:
On the cloud front, our footprint continues to grow even faster than our overall server endpoint growth with over 25% of the servers we protect now in the public cloud.
On their forays deeper into Zero Trust, in large part thanks to the Preempt acquisition:
Unlike any other competitor in the market today, our identity protection modules give customers the ability to prevent the spread of ransomware and stop lateral movement when credentials are stolen. This is a major advantage to winning deals in the field and increasing deal size. The third quarter marks the 1-year anniversary of our acquisition of Preempt. And in this 1 quarter alone, we generated more net new ARR from our Zero Trust modules than in the history of Preempt before the acquisition.
Zero Trust on the data layer, via the SecureCircle acquisition:
CrowdStrike already natively enforces Zero Trust protection at the device layer and the identity layer. By harnessing technology from our recent acquisition of SecureCircle, we plan to take data protection to a new level by enforcing Zero Trust at the data layer. [...]
after we combine SecureCircle's technology with CrowdStrike Zero Trust modules, customers will gain even more fine grain visibility and control as well as continuous risk monitoring to detect and respond to threats, whether they manifest as a device, identity or data layer.
The CFO had some fun, citing the Rule of 40 heuristic popular with SAAS companies:
We also continued to perform at a high level well in excess of the SaaS industry's Rule of 40 benchmark, achieving a Rule of 77 and when calculated on a free cash flow basis, a Rule of 96 at scale with over $1.5 billion in ARR.
“The Salesforce of security”:
if folks are looking at us as just an endpoint company, pigeonholed endpoint company, they're really missing the big picture.
We've proven that we are a platform company, as I said before, the Salesforce of Security, with 21 modules. You can look at the attach rates.
The Oil Drum (2005-2013)
I was recently reminded of this website, which I used to read between about 2006 and 2011.
It was about energy, mostly through the lens of peak oil, coming scarcity, and its impacts on our civilization. This was during a huge commodity supercycle (largely driven by China, and with some supply shocks like hurricane Katrina).
At the time, if you looked back a couple decades at the price of oil, it looked like this:
This seemed to confirm “peak oil” theory — the number of large oil fields discovered was down since the 1970s, new ones were increasingly expensive (deep water offshore, etc)…
The mind could easily extrapolate what this trend meant going forward.
But then, this happened:
(yes, including that brief dip in negative territory at the height of the pandemic)
The Oil Drum website stopped publishing in 2013, but I’m glad to see that they didn’t just wipe the server and pretend it never happened — all of the archives are still there, and you can go do some time-traveling to see what the headlines and arguments looked like at the time (f.ex. ‘Peak Oil, Carrying Capacity and Overshoot: Population, the Elephant in the Room’ from 2007).
I think it’s a great resource, and should be learned from, like old threads on Twitter/forums about convincing investment theses that didn’t work out the way they were predicted to...
Drowning in a river that is 4-foot-deep (on average), Redux
Microsoft is creating a version of Teams that isn't tied to Office (more directly competing with Zoom)
Microsoft is creating its first standalone version of Microsoft Teams for small businesses. Microsoft Teams Essentials will be priced at $4 per user per month and provide access to the core meeting features of Teams without the typical Office app bundling that requires a more expensive Microsoft 365 plan. [...]
The differences between Microsoft Teams Essentials and Microsoft 365 Business Basic plans are primarily around Teams functionality and cloud storage. Essentials only offers 10GB of OneDrive storage, compared to the 1TB available on Business Basic. Essentials also lacks meeting recording and transcripts functionality, real-time translation, breakout rooms, and whiteboard integration. (Source)
This is kind of a big deal in the space. Microsoft MSFT 0.00%↑ has such great distribution that I wouldn’t be surprised to see this low-priced product be quite widely adopted…
Facebook Picks AWS as “Key, Long-Term Strategic Cloud Provider”
Yeah, no, I won’t call them Meta (yet?). META 0.00%↑
Meta uses AWS’s proven infrastructure and comprehensive capabilities to complement its existing on-premises infrastructure, and will broaden its use of AWS compute, storage, databases, and security services to provide privacy, reliability, and scale in the cloud.
Meta will run third-party collaborations in AWS and use the cloud to support acquisitions of companies that are already powered by AWS. It will also use AWS’s compute services to accelerate artificial intelligence (AI) research and development for its Meta AI group.
In addition, Meta and AWS will work together to improve the performance for customers running PyTorch on AWS and accelerate how developers build, train, deploy, and operate artificial intelligence/machine learning models. (Source)
Very interesting, since Facebook is basically an hyperscaler cloud player itself, just not one that sells to third party customers like the other big ones do.
They’ll still use their own infrastructure for most of what they do, but this shows just how hard it is to compete with AWS, even if you have huge scale yourself. And how much they want PyTorch wider adoption to fight off Tensorflow…
To me, this is certainly positive for Amazon, not necessarily because of the dollar-amount involved here (unknown), but because of what it says about their competitive position.
AWS re:Invent 2021 Keynote
Speaking of AWS AMZN 0.00%↑, I watched the keynote of re:Invent, given for the first time by Adam Selipsky, now that Jassy is upstairs:
(I still think Selipsky should change his legal name to add “William” as a middle name… what great initials that would be)
They always announce a million things, so I won’t cover it all, but these are some of my highlights:
AWS Trainium, is the second machine learning chip built by AWS that is optimized for high-performance deep learning training.
Trn1 instances will deliver the best price performance for training deep learning models in the cloud for use cases such as natural language processing, object detection, image recognition, recommendation engines, intelligent search, and more. They support up to 16 Trainium accelerators, up to 800 Gbps of EFA networking throughput (double the networking bandwidth available in GPU-based instances), and ultra high speed intra-instance connectivity for the fastest ML training in Amazon EC2.
They are deployed in EC2 UltraClusters, which can be scaled to tens of thousands of Trainium accelerators with petabit scale, non-blocking networking. These Trn1 UltraClusters are 2.5x larger than previous generation EC2 UltraClusters and serve as a powerful supercomputer to rapidly train the most complex deep learning models.
This type of ML training may otherwise be done on Nvidia GPUs. It’ll be interesting to see some real world reviews of the Trainium chips, and how flexible and easy-to-use they are in real-world use.
Oh, and Trainium is kind of a cheesy name, sounds like something out of Avatar…
This means that AWS now has three main lines of custom-silicon: Graviton, Inferentia, and Trainium.
Amazon Web Services on Wednesday unveiled SageMaker Studio Lab, a free version of Amazon SageMaker -- the AWS service that helps customers build, train and deploy machine learning models. Designed for machine learning novices, users can try SageMaker Studio Lab without an AWS account, credit card or any cloud configuration knowledge.
This is cool. Making this tech more accessible, and creating more “citizen scientists” and “citizen ML users” or whatever is a worthy goal that can only be a positive, since there’s such a human talent bottleneck in these fields.
The service is based on open-source JupyterLab and provides free access to AWS compute resources. To begin, a user creates an account (separate from an AWS account) and selects whether they need a CPU or GPU instance for their project. The service offers 12 hours of CPU or four hours of GPU per user session, with an unlimited number of user sessions available.
It remains to be seen how good it is in practice, and whether they keep iterating on it, but it looks like a good start.
I also liked the private 5G announcements. Making it as turnkey as possible to deploy is a good way to increase adoption, and there’s no doubt a bunch of use cases for large industrial players (lots of IoT), healthcare, etc.
‘FTC Sues to Block $40 Billion Semiconductor Chip Merger [Between Nvidia and ARM]’ (plus Jensen 🔥)
1/ Since we are talking about the Nvidia-Arm deal today - a couple weeks back, Nvidia CEO Jensen Huang received the chip industry's highest honor at an event hosted by an industry trade group. That same night, Qualcomm's CEO Cristiano Amon was named chair of the group's board.
2/ As chair of the group, Amon's main task is to help advocate for the industry as a whole. After Amon had said a few words, Huang took the stage to receive his award and made a biting joke about Amon and Qualcomm, which has had - and won - its own extensive fight w/ the FTC.
3/ Here's what Huang said:
"I want to say that Cristiano is the perfect chair for the SIA. [Applause) He's the perfect person to advocate for our industry. And let me tell you why. I connected some dots tonight. I was trying to figure out, how is it possible that Cristiano knew every single regulator on the planet, and by the time I got there to tell them about my story on Arm, he was already there advocating against it?" [Stunned laughter, groans, applause] "Am I right? Am I right? Based on my experience, and this is a personal one - based on my firsthand experience, we are gonna be well served."
Follow-up to ‘Yesterday's luxuries become today's commodities’
Back in edition #209 I posted about things that used to be expensive, exclusive, scarce, etc, becoming cheap and commoditized. Here’s anther interesting example:
If you want your own lab-created “precious” gems, check out this store.
Liberty’s Highlights is reader-supported. To support my work, consider becoming a paid subscriber. 💙
Science & Technology
New magnet tech helping bring fusion power closer
Continuing the focus on nuclear power from edition #210, here’s a cool update on a new small fusion reactor design made possible thanks to a new kind of extremely powerful magnet, with a working test reactor being built in the next few years (much smaller and less expensive than ITER, yet still with a theoretical Q ratio — that’s how much energy it produces vs how much you need to put in — in the small ballpark).
📡 🛰 Geeky Review of SpaceX Starlink Internet (Still Beta)
This is my kind of reviewer, not doing things half-assed, but going full geek-mode 🤓 for this review (I love it).
You can combine the review above with this great explanation of how Starklink works for an even better overview (note the video is from last year, so the numbers aren’t up to date, but the general principles have stayed the same):
One of the cool things about Starlink that I don’t seem mentioned enough is that it provides a new option for people in repressive countries to access the “real” internet without filtering/censorship.
Sure, it’ll be possible to try to prevent people from having the ground hardware to connect to Starlink, but I suspect that over time this will become increasingly inexpensive, small, and hard to fully banish (insert joke about “shaking your fist at the sky” here).
Liberty’s Highlights is reader-supported. To support my work, consider becoming a paid subscriber. 💛 🤫
The Arts & History
Forget about LIDAR for Self-Driving Cars — How About for Music Videos?
The video for the 2007 song "House of Cards" by Radiohead was believed to be the first use of real-time 3-D laser scanning to record a music video. The range data in the video is not completely from a lidar, as structured light scanning is also used. [...]
The filmmakers passed sheets of acrylic glass and mirrors through the lasers to create scenes in which the image appears distorted, partially disappears, or disintegrate as if being carried by wind. The data used to make the video was released as open source under the Creative Commons Attribution-Noncommercial-Share Alike 3.0 license and is available at Google Code as both CSV raw data and Processing code