98: How Cloudflare Grew a Zscaler Inside Itself, Twitter vs Clubhouse (again), Grimy NFTs, Coinbase, Superconducting CPUs, Shrek's Law, Brain RAM, and Zoom Seances

"Aren’t you tired of hearing my words in your brain"

No computer has ever been designed that is ever aware of what it's doing; but most of the time, we aren't either.

—Marvin Minsky

Thanks to the Gambler’s Fallacy — which doesn’t just have to do with gambling — most of us are probably thinking “wow, the past few years have been really weird and rough, the past year especially… so we’re due for a pendulum swing in the other direction, for some good times, for some calmer seas”.

But it’s not how things work.

You can have a run that goes something like WWI -> 1918 Influenza -> *party* -> Great Depression -> WWII.

Recent negative events tend to increase the chances of more bad events in the near future, because they leave behind all kinds of landmines and stuff that hasn’t been fully dealt with (so years are not independent like coin tosses, and the state of things isn’t reverting to the mean on any predictable timeline).

So I’m hoping that we have good times ahead, but try to steel myself for anything.

🐦 Follow-up on Spaces vs Clubhouse: One idea I had about the situation, but not fully developed until I had a DM conversation with someone on Twitter (who prefers to stay anonymous — I always like to give people credit, so I asked), is that distribution isn’t just some abstract advantage that Twitter has. For a lot of bigger organizations, ‘influencers’, and brands, it’s a very real $ consideration:

If you’ve spend years investing in building your Twitter audience and followers, you don’t want to start from zero again on Clubhouse if you can avoid it, it’s very expensive.

At first, CH may be the wild west, but over time, the power law curve will probably re-form and a lot of big names and brands will be central to a lot of the most popular content (maybe not corporate brands, but personal brands, show/room brands, etc.. Kind of like with podcasts — you don’t just pick podcasts at random, you go back to the same few). If these people already have lots of followers on Twitter, they’ll probably get a much better ROI by doing their social audio stuff on Spaces than on Clubhouse.

The main benefit of CH will be that it’ll be the wild west for a while, so new entrants have a once-in-a-few-years chance to try to cement themselves as real players on the new platform. Kind of like how the big Tiktok accounts mostly aren’t just Youtube people that went over, but new people that grew with the platform and “got” what made it different early.

So some are no doubt hoping that CH will grow up to be a new major star in the social media constellation, and that as early-adopters they can ride that wave and grow big with it.

But if Twitter can instead just graft this new way to communicate onto its existing network, then it’ll probably be the existing Twitter big accounts that have the upper hand…

It’ll certainly be interesting to see how that plays out over the next couple years.

📚 The idea of differentiation through having an unusual skill stack isn’t new, but as usual, re-learning old ideas and obvious things is where most of the value is.

So rather than think about what skill you have/want, think about what skillS you have/want, and how these work as a group.

It can take you pretty far in life to just add to your stack one or two skills that are rare among your peers.

So if you’re a research scientist who happens to be much better at making Youtube videos and building a Twitter audience than most research scientists, it could have a bigger impact on your career than if you had just focused on becoming 10% better in the lab (the last few percents at any skill usually take as much work, or more, than the rest combined).

An investor who’s better at writing, an athlete who has a better understanding of psychology, a teacher who can draw and do CGI animation pretty well, a small restaurant owner who understands design and marketing better than the competition, etc.

Unusual success seems to come from either being super-specialized at one thing, and being on the peaky part of the power law, or when you can combine some stuff in ways that make you scarce.

In between, there’s lots of people who aren’t good enough to stand out purely for one thing, and aren’t exotic enough in other ways to stand out for that, so they have a more difficult road if you want to separate themselves from the pack (of course, not everybody has the ambition to do that, and there’s nothing wrong with that).

💚 🥃 Wow, thank you everyone, your support means the world to me.

I’m just a guy in pyjamas banging away on the keyword about random thoughts I had while walking in the woods, or making stupid dad jokes, so the support since I announced the ability to become a paid supporter (💚) has been truly touching.

Do I need to go in Photoshop (I actually use Pixelmator Pro, but when you say that, fewer people know what you mean — Adobe really owns the mindspace in that vertical) and make “founding members” badges for the early group of supporters (💚)? Maybe I need NFT badges to be with the times?

Doing this was a leap of faith, and I was very nervous and excited. It’s still a leap of faith, because it’s too early to know how it’ll go — but whatever happens;

Seriously, thank you.

Investing & Business

How Cloudflare Grew a Zscaler Inside Itself

I like how Muji describes this in his post about Edge Networks (check it out):

Cloudflare has their Argo Tunnel product to bring their edge network all the way into your data center (protecting the edge-to-origin above). Their Magic Transit product goes a step further, and brings Cloudflare’s platform directly onto your enterprise network.  They then added another product, Cloudflare Access, to protect incoming traffic, allowing users to safely access a company’s origin server, regardless of where it lives (in the cloud or an on-prem data center). This is accomplished through Zero Trust capabilities they have built, which are possible due to the software-defined networking architecture they've adopted.

Using those software tunnels, the only unprotected part left was the Last Mile between the endpoint (requesting device) to the nearest edge server. So Cloudflare then created a Secure Web Gateway (SWG) product called Cloudflare Gateway, to protect the traffic from the endpoint making requests to the SaaS services that an enterprise uses. Between outgoing traffic production in Gateway and incoming traffic protection in Access, Cloudflare now offers end-to-end protection of network traffic for an enterprise's users.

If this combo sounds familiar, it is because this is entirely what Zscaler (ZS) has built. They protect outgoing traffic from endpoints to external servers (ZIA, their Secure Web Gateway) and incoming traffic from endpoints to origin servers (ZPA, their Zero Trust secure access method). Cloudflare evolved their Gateway and Access products into a new competiting platform called Cloudflare for Teams.

This shows what kind of architecture leverage these edge networks (aka edge cloud platforms) truly have. Cloudflare may have just doubled its TAM, with an ideal product for today’s stay-at-home world.

Of course, it’s not because you’ve built the capabilities of someone else that you automatically are as good at it or have their economics, but it’s a prerequisite to compete, and if you can graft this unto an existing effective distribution system, and you are already known as being good in the space (security), your chances of being able to get somewhere are pretty good.

For reference, Zscaler has an enterprise value of $27.6bn and CloudFlare has an EV of $22.8bn.

Speaking of Cloudflare (this is what we, in the business, call a smooth transition), I gotta admit that it increases the feeling that they have ears and eyes everywhere online when the CEO of the company pops in a Twitter thread where he wasn’t tagged by anyone — kind of like Beetlejuice, but for Cloudflare:

Coinbase S-1

So that’s happening. I have nothing to add, I just think it’s kind of fascinating that this is turning out to be a pretty centralized part of this decentralized revolution (“[Coinbase] currently hold almost 12% of the world's crypto assets.”):

‘Grimes made $5.8 million in under 20 minutes selling crypto-based artwork’

I like Grimes. Or at least, I like most of her music that I’ve heard (primarily the albums ‘Visions’ (2012) and ‘Art Angels’ (2015)). So nothing against her figuring out how to make a few millions from jumping on the latest hot thing…

The singer launched her digital collection, "WarNymph," on Sunday and plans to sell the crypto art for 48 hours. [...]

Grimes' 10 NFTs brought in over $5.8 million in under 20 minutes. [...]

One piece, an image that depicts a baby guarding Mars, has garnered significant attention, trading at over $300,000 in under 10 minutes. The "Newborn 2" NFT has already been relisted at $2.5 million. [...]

The NFT collection is one of many that Grimes plans to sell. She said that she was creating a mythical universe called Oth3rkin and that WarNymph would evolve as a part of it. (Source)

You can see the various pieces and auctions here.

So, I’m still processing what I think of the whole NFT thing.

It’s too easy to say that it’s either the 21st century way to collect art and for artists to monetize their work and the most brilliant thing ever, or the dumbed thing ever and you’re being scammed into buying nothing.

Reality’s probably somewhere in between, but it’s not because it’s *somewhere* in between that it’s in the *middle*. Which way will it lean when the dust settles?


It does kind of feel like a mirror image of the fact that it’s usually not a good idea to think that physical things can have the same kind of attributes as digital ones.

Now we’re trying to act like digital assets have the same attributes as physical ones, but do they really? Just because it’s written in a distributed database somewhere that Charlie Utter owns Grimes’ ‘Space Baby’ digital painting doesn’t mean that it’s the same, or even similar, as physically having a painting.

When you own the vast majority of art, what you own is inherently different from reproductions. You can get a print of a painting, but the brushstrokes and the 3D nature of the dried paint won’t be the same… You don’t have something that the artist has shaped directly, that he/she touched, smiled while looking at, cried with a bottle of wine in the studio where it was made, etc.

With digital assets, you have exactly the same thing as everybody else. So even though it says that you “own” it in a database (or own one of the dozens or hundreds of exact copies, because NFT sellers have realized they can make more by selling a bunch rather than just one), this ownership is a lot more abstract than having the actual piece of canvas that Rembrandt spent weeks caressing with his brushes.

According to Tanay Jaipuria:

In the last 30 days, NFT collectibles have seen over $350M worth of transactions, dominated by NBA Top Shot.

I think what would really answer my questions about NFTs is if you could somehow inject a truth serum to those who buy them, and know their real motivations.

I think the biggest tell would be: what would the price be if you couldn’t resell it later. What part of the price is “wanting ownership of something” and what part is “financial speculation”.

There’s always financial speculation with physical art too, but I wouldn’t be surprised if the ratio was much more skewed with digital NFTs. A lot of wealthy people enjoy having paintings and sculptures in their living room and would buy them even if they could never resell them — I’m not so sure about Nyan Cat GIFs.

🍞🍞 🦆

I mean, if you zoom out, I kind of don’t mind if the wealthy finance/business people end up transferring a bunch of their money to artists and performers through NFTs. That’s not the worst outcome. ¯\_(ツ)_/¯

But hey, I don’t claim to really understand the space, I’m just starting to think about it, so consider this my thinking-out-loud, not some final declaration.

*Wipe Brain RAM*

(maybe it’s more like the brain’s L1 cache, but that’s a geekier, more niche headlines)

Ask economists, and few things are more important to solve the humanity’s problems than productivity growth. Whoever’s not living in abject poverty today can thank the massive growth in productivity per capita since the industrial revolution.

And yet, there are still so many practices that are just terrible for human productivity, low-hanging fruits caused by bad incentives, misconceptions, and neglect.

How many neurons are working on important problems at any time on this planet, and how often are they interrupted for BS reasons that cause real damage. We’re not computers that can just switch back and forth between tasks. Our working memory can hold something like 7-9 numbers and it can take hours to get in a flow-state, if it happens at all…

So it’s cool that we’re building robots and software and faster computers to increase productivity, but how about we really focus on getting the most out of the wetware between our ears too?

This is a pretty broad mandate that covers education and working environments and the tools that we use and such, but it’s a very high-leverage area of focus that more people should try to tackle, IMO.

Science & Technology

Superconducting CPUs?

What if the real industry of the future isn’t “semiconductors” but “superconductors”?

One group of researchers in Japan have taken this idea to the limit, creating a superconducting microprocessor—one with zero electrical resistance. [...]

Superconductor microprocessors could offer a potential solution for more energy efficient computing power—but for the fact that, at present, these designs require ultra-cold temperatures below 10 kelvin (or -263 degrees Celsius). The research group in Japan sought to create a superconductor microprocessor that’s adiabatic, meaning that, in principle, energy is not gained or lost from the system during the computing process. [...]

“the data processing part of the microprocessor can operate up to a clock frequency of 2.5 GHz making this on par with today’s computing technologies. We even expect this to increase to 5-10 GHz as we make improvements in our design methodology and our experimental setup.” [...]

The price of entry for the niobium-based microprocessor is of course the cryogenics and the energy cost for cooling the system down to superconducting temperatures.

But even when taking this cooling overhead into account,” says Ayala, “The AQFP is still about 80 times more energy-efficient when compared to the state-of-the-art semiconductor electronic device, [such as] 7-nm FinFET, available today.” (Source)

Cryonics isn’t the kind of stuff you want to have at home, but in a massive high-performance data-center in 2060 or whatever, it’s not out of the question if it means that you can reach very high clock speeds while using a fraction of the energy.

But it’s still all pretty sci-fi at this point.

‘Microsoft Power Fx, a new low-code language that takes its cues from Excel formulas’

If you’ve been spending way too much time in Excel, good news, there may be some new horizontal transfer for those skills:

Power Fx, a new low-code language that takes its cues from Excel formulas [...] Power Fx will become the standard for writing logic customization across Microsoft’s own low-code Power Platform, but since the company is open-sourcing the language, Microsoft also hopes that others will implement it as well and that it will become the de facto standard for these kinds of use cases. [...]

Beyond Excel, the team also took inspiration from tools and languages like Pascal, Mathematica and Miranda, a functional programming language developed in the 1980s. [...]

Microsoft plans to bring Power Fx to all of its low-code platforms (Source)


The Arts & History

‘Shrek’s Law’: Computer Animation Rendering CPU-Hours

Despite the advances in computing power over the 2000s decade, the increasing usage of novel techniques like global illumination, physics simulation, and 3D demanded ever more CPU hours to render the films.

DreamWorks Animation noticed that every Shrek film took roughly twice the CPU hours than the previous film and thus labeled this trend as "Shrek's law". Similar to "Moore's law" the Shrek's law says, "The CPU render hours needed to complete production on a theatrical sequel will double compared to the amount of time needed on the previous film."

In 2001, Shrek required approximately 5 million CPU render hours.

In 2004, Shrek 2 required over 10 million CPU render hours.

In 2007, Shrek the Third required over 20 million CPU render hours [...]

2010 3D release of Shrek Forever After demanded more than 50 million CPU render hours on account of rendering an increased amount of frames.

Puss in Boots, which was released only one year after the previous Shrek film, utilized 63 million render hours (Source)

Kora Jazz Trio

I write and I write and I write. Aren’t you tired of hearing my words in your brain, as mediated by photons coming off your screen? Here’s one to just listen to.

This band is so consistently good that I can basically recommend that you listen to anything from any of their 4 albums, and whatever you pick will give you an idea of the heart-enlarging (not in a medically dangerous way) goodness that they produce.

But if you need more words to seal the deal, here’s some:

Kora Jazz Trio is a three piece African musical group, founded in 2002 by Djeli Moussa Diawara, Guinean Korafola, with Abdoulaye Diabate and Moussa Sissokho, best known for producing a music that is a mix of American jazz with traditional African music.

Described as "the encounter between mandinga musical tradition and the freedom of jazz, between West African percussion and Afro-American swing", they have been recognized for their focus on sharing their cultural heritage, without doing so for the sake of mainstream success or in an effort to create a movement.

Here’s some convenient links (don’t say I never did anything for you):