Casey published a wonderful straight-to-camera analysis of Eric Schmidt’s now-notorious 2026 Commencement Speech and the reactions to it - especially by the audience vociferously booing & jeering the former CEO of Alphabet.
Stop right now and go watch it if you haven’t seen Casey’s video yet:
OK, did you watch it? Now go watch it again.
OK, back? Here’s the theme that jumped out at me: Architectural Accountability.
Casey takes us on a wonderful journey reflecting Eric Schmidt’s total lack of awareness, and the dis-associative mental gymnastics the former CEO must be doing to say that that “It Just Happened,” when remarks upon the complicated, regressive, problematic, and dangerous technological world came to be as it is today.
It.Just.Happened.
No. We Built It.
One of my grandfathers was an engineer, an architect, and led teams, and supervised complex projects, including ones with mixed social + technological implications, like the B2 bomber.
Another grandfather of mine also worked on construction projects, supervising complex builds, establishing grandiose buildings, and very much swimming in the cultural realms of traditional architects.
I suppose a certain amount of their way of thinking rubbed off on me, but what I was struck by growing up, learning from them, was the intentionality, the care, and the focus on posterity, as an artifact they were helping to craft, that they always seemed to aim towards in their engineering work.
Not always perfectly, mind you, as humans, but it was a core value that - to me - very much seemed in-line with the mores of the domains of engineering and architecture they worked inside of and amongst.
That was embodied by the care, devotion, and paternalistic love my most cherished grandfather (the architect) showed me when, while I was the age of 8, he helped me build a treehouse in my backyard.
I was fabulously lucky to grow up in a wonderful home with two giant trees in the backyard in Southern California.
When my parents were looking at homes, to evaluate where to choose to buy, my father, his grandfather a painter, knew to go around the neighborhood to meet the people living in these homes. He found people in the trades: electricians, carpenters, plumbers, and so forth - that told my parents the bones of the home were strong and it was a safe investment.
It was, and that was part of why we were lucky enough to be able to afford the lumber + parts for what this young imaginative kid wanted, a very specific vision for his treehouse.
What I distinctly remember about my grandfather is what he taught me about the architectural discipline.
I was the beloved program manager, who could do & say no wrong, and who he could only very deftly, with a clever, persuasive and expertly-charming manner redirect my wild ideas into a - how shall we say - actually achievable direction.
He had to - artfully - find a way to correct me, teach me, and disappoint me, tactfully, when my original ideas for triangular spars had to be reconsidered and replaced with upright 4x4s at each corner quadrant of my treehouse.
I still remember - and it STILL pains me a little to this day (we sold the house, the person who bought it CUT DOWN THE TREES, ugh) to think of it - how IRRITATED and DISAPPOINTED I was when I learned that no, it would not be build according to MY SPECIFICATIONS.
I had dreamed it up - I was a very precocious and imaginative boy who had a vision of a wide, flat, broad deck of my treehouse, with railing (I remember this was my idea, whether or not it was) but I distinctly remember that it made sense to me - at my age then - for the floorboards of the treehouse to obviously be supported by corner angle joists angled toward the trunk of the tree from underneath the treehouse.
I distinctly remember my simplistic triangular designs had a sort of beautiful simplicity in profile as I imagined the treehouse I wanted.
But no, my grandfather had to teach me, that would not work.
The triangular joist would have put enormous pressure on the trunk, where it met the bark, and would have progressively killed the tree all the way up as the wood pressed into the circulatory system of the tree, leading to the tree dying, and my whole treehouse would have fallen down after killing the tree.
But it looked so PLAIN and so BORING to have straight uprights at each corner quadrant!
And why should we have an ugly 2x4 scaffold underneath the floorboards I remember thinking to myself, vs. clean lines of smooth plywood?? What if we did ANOTHER layer of plywood underneath that scaffold so that - to the naked eye - we’d see just the clean lines of the plywood?
Well, those areas could collect water, you wouldn’t be able to visually inspect it if any wood had begun to rot, etc.
Wait why must each 4x4 sit inside of an ugly ring of cement? That doesn’t LOOK clean! Well, that is what anchors the uprights and prevents them from rotting when they meet the lawn.
Ugh why does the wood look so UGLY - well it has to be treated so it doesn’t rot, we’ll paint over it (I chose dark brown, so it would look more like natural wood).
But the swinging ladder should NOT be tied down at the bottom with metal rods buried deep into my lawn, it ought to SWING back and forth! Well, if it did that, you’d probably break your neck as it swayed back and forth.
You see what I’m getting at here?
That’s a discipline practiced every day by so many builders of large, medium to gargantuan physical projects all over the world.
Get the requirements. Negotiate with the benefactors / investors / sponsors of the project on those goals / needs / aspirations. Figure out what CAN be done, and work with the people building it, and as you do so, find out where you have to compromise, and bring the whole team, especially the most aspirational drivers of the project, along for the ride to what is truly safe, practical, and achievable, and hopefully - if you get it right - you all build something you can all be proud of.
I was proud of that treehouse.
Was I spoiled? ABSOLUTELY, but Godamn I loved that treehouse. Did I spend ENOUGH time up in it to justify all of the pain and anguish I put my grandfather (and my father) through as I project managed them to death in the building of it? No.
But it still puts a huge smile on my face every time I think of it. I’m so grateful for it now for what it taught me about how to BUILD things.
I’m so grateful to my grandfather (and father) for what it taught me about Architectural Accountability.
They helped me build something that didn’t kill me, and they corrected and adjusted my irresponsible designs when / if it would have put me - the child they were responsible (and the builders themselves, lol) at risk.
As we approach Father’s Day that sort of parental instinct my Grandfather and Father both showed me in letting me “order,” them a round a bit while they extended incredible patience to me in hearing me out, looking at my sketches (I remember being so frustrated trying to EXPLAIN and DRAW my vision and trying to work with them to ensure they UNDERSTOOD it and in understanding what limits / requirements we had around safety my grandfather and father both had to teach me), and getting the lumber, letting me put some nails into the wood, etc.
But the best gift was what it taught me about the responsibility we have to others for what we build and HOW we build it. We have to have care for how it will age, how will it work, how will it function, and we OWE people who use our things a baseline of safety, reliability, and we need to build it intentionally.
I bring that whole childish project and anecdote up because it feels to me like in the past 10-15 years especially - that sort of journey has been less and less likely to come to pass in the things technology companies build.
If you’ve worked in FAANG, or worked WITH technology companies, or interfaced with the major technology companies, you know this.
It’s why Cory Doctorow’s term Enshittification has become a mainstay in understanding how Silicon Valley operates in a lot of ways these days - we seem to be able to count on things diminishing, degrading, destructing themselves and the things around them, and - increasingly - the people who use them.
Our software, our hardware, and our technological world has suffered for it as a result, so much so that so many people don’t trust tech companies to produce something good, something stable, durable, scalable, and something fun.
So, as Casey points out, it’s no wonder the young people in that crowd vociferously boo and jeer Eric Schmidt for his role in our systems, the very architecture of our world in the software systems he helped architect, as he had the gall to speak to them as though he had simply seen the traffic gotten jammed up & “oh well.”
No, there’s been a diminishing of the values that I saw my grandfathers espouse in the work they did, and what I’ve seen in the trades, in architecture.
Sure, lots of mistakes, lots of incompetence.
And people complain it’s expensive to build buildings in America.
For my part, I’m glad of the expense, it’s how I feel a certain amount of universal confidence my building won’t randomly collapse in on itself while I’m sleeping.
My software for the last 10-15 years? Not so much.
In the software world, if you’re not familiar with it, building things often looks something like this:
- You have an existing product or service, an app, a website, or some such other
- Leaders steer the company outlook / goals / objectives in “serving the market,” or “supporting our clients,” - reaching new customers and giving them value
- Middle-management help translate what’s working & what’s not up + down, feedback goes up & direction/instruction + feedback ideally goes down as well.
- Individual contributors, hopefully, at their level, get a clear sense of what to do first, second, and last, and step toward their goals.
- That usually manifests, at the individual contributor level, in discrete sprints of adding to, refactoring, or updating code, or shipping new “features,” - an integration, an interface, a set of buttons, or a revamped look / layout, that altogether hopefully provide greater value, save someone some amount of time or money, and - ideally - that (much like the Toyota manufacturing discipline) continuously improves the systems in aggregate.
What Casey so beautifully calls out here in his video is the stark extent to which Eric Schmidt declines to take an active role of any sort of accountability for his part in the above process that - clearly to his audience - has led to such a REGRESSION in technology for everyone.
It’s all over the place.
People feel like they are just renting their things, their data, the stuff they own.
People feel like every time they want some kind of technology to do something for themselves, it’s going to come with a financial hook in their monthly budget.
People see the ways in which technology could have obviously, clearly, and - to them (whether true or not) - could have EASILY been solved had enough care and attention been paid to it. Their phone glitches out when they just want to call a family member. They can’t find where this file was that they took great care to store in a specific place in their computer or their accounts. They ask an AI tool a simple question and get…the dumbest response anyone could have come up with.
People are conspicuously foisted up with exemplars of all the ways technology is TAKING from them, and they either miss or overlook (or take for granted) all the ways in which it makes a difference for them - which is another dark-pattern Casey is talking about here, some shame on all of us for failing to make it obvious and communicate to people when/if we improve their lives materially.
Part of what a lot of software engineers are worried about with AI, with the less attention individual code will get, is the potential degradation, en-masse, to said code, with more bugs, errors, glitches, and worse experiences for users as a result.
This was a trend well on it’s way before the latest hype-fest about LLMs & so forth.
I think we’ve lost something drifting away from Architectural values, and I’d love to see us go back to that.
Not long ago, the older book, “A Pattern Language,” had a sort of mini-renaissance in the Silicon Valley world. One of the architects of GraphQL (which I use to build my personal websites, and we used at Project N95) talked about it in this AI Engineer video, and he linked it to disciplines + practices of having your AI Agents articulate, document, and record their choices ABOUT the code, about the frameworks, about the systems you are building, with what Addy Osmani also emphasized: Architectural Decision Records.
One of Casey’s frequent co-creators, known as “The Primagen,” often loves to deride people super excited by AI as having never been true software engineers - never having been the glorious hand-code-rollers, who now are walking around pretending to be glorious software engineers just because they can generate code that looks credible.
I don’t know who these people are, I’ve never heard of them.
Do I often feel like such an imposter? ABSOLUTELY, but have I ever met any of these fools pretending to create code but wholly incompetent? No not yet.
Are there people fabulously incompetent putting up terrible code in all kinds of places? I’m sure.
But here’s what I keep coming back to: Architectural Accountability.
Eric Schmidt refuses to take any, but we KNOW he’s both Responsible AND Accountable.
So what’s the takeaway here? Where am I going with this?
We can ALL be Architects.
There’s nothing wrong with being an Architect, if you do it intentionally, ethically, and responsibly.
I find it to be such a fascinating discipline myself, especially architecture for massive projects.
Take the Golden Gate Bridge, or major skyscrapers. Architects for such projects have to - literally - take into consideration that the entire thing will be built by humans, tasked with putting hot rivets in to steel girders with a certain amount of human error, or human negligence. Supervisors, inspections, and everyone up will make a certain amount of mistakes. Humans are fallible, so the literal design of the buildings and structures had to be conceived in a way that anticipated these small imperfections.
The architects had to plan for such machinations of small, iterative mistakes.
Lots of projects don’t anticipate this - the Golden Gate Bridge’s recent massive renovation of the railing famously created a loud harmonica due to the strange resonance of the arrays of thick girders for the railings, a system that is being actively redesigned and redone to reduce the ominous humming that persists throughout the bay on blustery days.
But I continue to see - perhaps only because I am not in civil engineering / architecture - exemplars of - even when there are mistakes made - where humans, steered by architects supported by sponsors / creators / investors, all worked together to build giant buildings, tunnels, bridges, and even small homes that - with exceptions - we humans can generally rely upon.
And while America’s infrastructure of roads, freeways, bridges, tunnels, railways, and even some skyscrapers certainly has it’s warts, we - collectively more or less TRUST these buildings (again, partly because the ones we can’t fall down or are taken down).
That comes from a long, long healthy and distinguished tradition of architects checking and re-checking their “code,” - returning to old designs, and even some beautiful examples of architectural students studying designs and discovering extreme design failure points that could have led to a total collapse of a skyscraper, only to help avert it by reaching out to the architect, who then spun up a fabulous story of how they avoided disaster.
But that infallible reputation, where we just trust that the things around us will stand up and generally stand the test of time is just … not true more and more of our software, of our technology. We’re just…stuck with it.
Of course I’m generalizing here, but intentionally so, for effect, to make a point. Are there exceptions to that rule? Absolutely, but it’s something that on the margins, generally, to me, seems to be more and more true every day. It’s harder to trust the technology in your laptop, in your phone, on your computer, and in our internet, and in our AI.
And I think that says a lot, we don’t see a lot of the technological infrastructure around us as thriving, as something that we can trust.
And sure, I trusted the iPad I wrote this on and the internet that helped me save it + serve it to you.
But on the other hand, it wouldn’t have surprised me much if my Notion or WiFi had gone a little haywire somewhere as I typed it out.
Wouldn’t have shocked me at all if my connection got interrupted or the Notion system didn’t save my essay perfectly, etc.
The whole time I typed I was fighting stupid auto-correct typos even with a keyboard (still so much less efficient than a ink typewriter even with an iPad Pro Apple Keyboard.
What I’m trying to get at here is I think, if we can use AI ethically, responsibly, and intentionally, we have more opportunities to become architects here, but - it’s my hope our discipline + practices can transition MORE toward the strong traditions of Architectural Accountability that force us all to be intentional, sustainable, reliable, and certifiable about what we build and HOW we build it.
I’m likely biased here because ADRs (Architectural Decision Records) are a framework I’ve been leaning on to ensure I’m accountable for what I’m delegating to agents.
Is that system perfect? No. Does it get me far closer to the code I’m shipping vs. just casually scanning diffs / MRs? I’d like to think so.
Is it as laudable as say - hand-rolling all the code myself? I don’t feel like it is, but I can’t do that at the scale I want to build all the things I want to.
I think - as we all scale to build more - we all have an opportunity to actually build BETTER as we do, if we target the high-leverage points we have: the architecture of HOW we build far more software these days.
For me, right now, that’s why I put so much thought and time into my agent harnesses, my enforcement systems around how the agents I delegate to build + use the software I’m shipping.
I am trying to take that intentionality into what I’m building because I want to be proud of what I build but also because I think lots of other people will have to rely - to some extent - on what I ship.
Am I building the world’s telecommunications network infrastructure? Absolutely not.
But am I writing, building things at least a handful of other humans will rely on to some extent and hopefully find a lot of use for? I hope so, and even if not, that’s HOW I want to build because that’s HOW my Father and Grandfather taught me to build things.
I hope more of us in our fields take the opportunity to lean MORE into that sort of intentionality as we reflect on why the kids at that commencement booed and jeered their elder who I too look down upon, not up to, Eric Schmidt.
Something to think about as we head into Father’s Day.
