This is a little late, but I like to defy convention (and/or am lazy). Like everything I write, this introspection is mostly for my own benefit.
I posted 14 times in 2022, down from 16 the previous year. I wrote about CAPEX, cloud repatriation, energy and the cloud, the unwind of the startup bubble, virtual real estate, and posed a few questions about “supercloud” (not least whether it takes an article or not).
Word count was up to 17,263, from 10,177 (I really am trying to write shorter and leave more parenthetical comments on the editing room floor (but not this one…)).
Traffic has flipped so the majority of distribution is now via email (YOU TOO CAN SUBSCRIBE VIA EMAIL IN THE RIGHT COLUMN). Top sources of inbound traffic were search, Twitter, LinkedIn and then a very long tail.
The five most viewed posts were:
Should I Stay or Should I Go? (cloud repatriation)
Cloud Power Up (energy and the cloud)
Supercloudifragilisticexpialidocious II: Questions Remain (“supercloud”)
My favorite posts were:
My favorite line of the year:
This has led to skyrocketing valuations and, admittedly, a few excesses (e.g. SoftBank, drunken weaving across the fine line between a tech company and an IT department, “due diligence, shmoo diligence…”).
I’m staying on Twitter, but like to think of this as my long form Twitter. And I have noble if very abstract goals of spending more time here than on Twitter.
Some possible future posts in various stages of completion include a new antitrust doctrine, edge networks (CloudFlare), the scope of AWS, perfidious private equity, the financialization of software, the metaverse’s arms race in legs, mercantilism in everything, winning the meme war with China, a look at a very special Oracle PR operation, and a compendium of jokes about technology and the EU (slogan: you can’t spell Butlerian Jihad without E and U).
Topics I’m interested in, from a startup investing and/or pure intellectual curiosity standpoint: the next wave of XR startups, what new crypto narrative emerges (there will be one), AI personal agents built on LLMs (please manage and summarize all my feeds), brain-computer interfaces, abundant energy, orbital computing, and the great SaaS consolidation.
Comments on Specific Posts
This was published just about two months after what is now recognized as the market top, and holds up well. The august venture capital community leaned hard into the same message about six months later (and haven’t stopped). There is still a ton of excess to work out of the system, but I’m cautiously optimistic as a very early angel investor. I’m seeing better quality and more focused founders now, and less crazy valuations relative to the three previous bubblicious years.
My greatest regret of this bubble is not to have been tracking a time series of people on LinkedIn identifying themselves as venture capitalists. It will be parabolic.
CAPEX has long ceased to be an insightful cloud tell but maybe it will be more interesting in a belt-tightening economic environment. I will labor on in summarizing cloud CAPEX for old time’s sake. The most interesting CAPEX spend now is around energy, both for the cloud and more generally. There are also some other areas of technology making real CAPEX investments, if not at the scale of public cloud, that I keep threatening to dive into.
Oracle is pulling away from the IBM in the race between the traditional CAPEX clowns (but not enough to get off the x-axis relative to the hyperclouds). But we have some new clowns in town to amuse us: the EU and the “European cloud industry”, Bank of America (slogan: the worldwide leader in financial crime), and the Trump Media & Technology Group (slogan: incompetence is functionally indistinguishable from censorship).
Where oh where have the boosters of virtual real estate gone now that the market has imploded? Hopefully the A16Z media empire will update its bull case soon! And not surprisingly, the EU managed to beclown themselves in the metaverse.
Cloud repatriation is another A16Z thesis I struggle with. Hopefully this post will beget further Clash-themed posts.
Like a lot of people, I focused on energy after the Russian invasion of Ukraine. The CAPEX lens naturally applies. It is scary to understand the path we’re on. Conflating opposition to carbon with opposition to all forms of energy generation and infrastructure is a looming disaster and if not corrected will leave us cold, hungry and in the dark. There is no excuse for our civilization not to have cheap, abundant and clean energy. I urge the cloud companies to take the lead on a new generation of nuclear power.
This simple heuristic gives me an excuse to look for actual evidence of cloud repatriation being a thing. My bet is the tsunami of workloads moving to the cloud will continue to swamp the anecdotal examples moving off the cloud.
The Supercloudifragilisticexpialidocious Saga
This “super” niche series of posts dating back to last December (here, here, here, here, here, here, here, here and here) stemmed from my absolute and unabating bafflement with the proposed concept of “supercloud”.
You can think of this saga as kind of the Lincoln-Douglas debates of our time, except where the topic is middleware and the other side doesn’t actually respond to the arguments presented.
Team “supercloud” are threatening to belabor this topic into year two, so I will continue to monitor any progress towards a precise and useful definition.
From the Back Catalog
I’m glad to see “loser” is finally catching on as the tag of choice for Donald Trump. More than a little surprised there is not a crowd outside Maga-a-Lago chanting “loser” 24×7.