I write this letter to share some of the things I’m working on and thinking about. A little about me:
I own and manage a digital agency in Toronto called August, where we design and build websites and apps. We spend about 80% of our time on client work and the balance on internal projects. Most recently, we’ve been working on a directory of real estate development vendors called Buildstack (think: Builtwith for buildings). I also invest in and develop real estate with my two brothers.
some thoughts
I think that this was my most viral tweet ever:
I got that 2-10x figure from Nat Friedman’s website, and I think that it’s mostly right. Elon’s certainly proved that it was right at Twitter, which has decreased its headcount from ~7,500 to ~1,800 over the last six months with no obvious degradation in service.
So what if government is similarly overstaffed? The Ontario Public Service employs some 60,000 people. If we can assume an all-in cost per head of $120,000 per year, including non-wage benefits and overhead, cutting that number in half could save us $3.6-billion per year.
And again, we’re working under the assumption that this staffing reduction can be achieved with no obvious degradation in service.
Can you think of a better way for the Province to spend $3.6-billion per year?
And it’s not just the cost. Overstaffing is generally a consequence of resisting efficiency improvements in all facets of an organization. It comes about, over time, by doing less with more.
Technology, by contrast, can be defined as the ability to do more with less.
What would it take to reduce staffing levels in government by 2-10x? More and better technology would surely be a big part of the answer.
Ian Vanagas, who’s stubstack has become one of my favourites, wrote a recent piece called Outmaneuvering Bureaucracy with AI. He starts by describing how more and better technology has not reduced the size of government to date—just the opposite.
The problem with bureaucracy is that it is the opposite of productivity. Time and resources spent on bureaucracy makes the process less efficient. You need more inputs for the same output.
The world continues to innovate, but bureaucracy seems to grow at a faster rate than non-bureaucratic positions. Among other reasons, this is a failure of innovations, most importantly, computing.
Computing theoretically automates and streamlines many bureaucratic functions. For example, email is faster than mail, an online form is faster than an office queue, and internet archives are faster than physical ones.
In reality, computing hasn’t made an impact, it’s maybe made bureaucracy worse. There isn’t a solid explanation for why this is the case. As time goes on, this grows increasingly concerning. Will continued innovation in computing only lead to further bureaucracy?
He then suggests that recent AI innovation might help us break from this trend. That this time, more and better technology might actually lead to greater efficiency.
The benefit AI can provide to solving bureaucracy is simplifying the interface. Nick Arner, in LLM Powered Assistants for Complex Interfaces, writes about how people can use large language models (LLMs) to interact with complex interfaces, such as 3D modeling software. It’s not a stretch to apply the concept to bureaucracies when thinking of them as complex interfaces.
An AI can be the layer between a motivated individual or team and the bureaucracy they are interacting with. It can help automate, summarize, and translate in ways that take massive amounts of manual work. A lot of dealing with bureaucracy isn’t doing a “perfect” job, but doing a “good enough” job which AI is perfect at supporting.
In this way, AI can help individuals outmaneuver bureaucracy. Because bureaucracy is slow, a motivated individual or small team can leverage AI to “solve” bureaucracy and get things done. It provides motivated groups the leverage they need to move quickly and capitalize on opportunities. This moves power back to individuals, and away from bureaucracies and large organizations.
That sounds right to me. This time, and this technology, might be different.
The August team and I are launching an experiment to test that hypothesis out. We’re calling it 311AI. I've described the general concept in a previous newsletter.
As the name would suggest, it’s an AI-powered 311 mobile app. You can take a look at our marketing website for it here.
The City of Toronto spends $18-million per year and employs 178 people for its 311 service. Can you imagine a near-future in which a highly accurate and reliable automated 311 interface performs the same function as, say, 90 call centre employees, saving the City something like $10-million per year? We can. And can you imagine that this new interface would provide a better user experience to Torontonians? We think it would.
That’s doing more with less.
Can you think of a better way for the City to spend $10-million per year?
business stuff
August: Back in 2015, I joined an agency called Plastic Mobile as a project manager. Its founders, Mel Adhami and Sep Seyedi, launched the agency in late 2007 just a few months before Apple’s launch of the App Store. They very quickly positioned Plastic as a mobile-first agency helping established companies develop their mobile strategies and applications. It was a big success, developing apps for companies like Rogers Media, RBC, Shoppers Drug Mart, and many others, and was acquired by Havas Worldwide Canada just over 7 years later. Mel and Sep made a bunch of money and are now on to even bigger and cooler things.
I think that the ChatGPT Plugins launch could be to August what the Apple App Store launch was to Plastic. That is, I think that we finally found something worth focusing on and specializing in as an agency. (Longtime readers of the newsletter will know that this question of niching down has been a consistent source of internal debate.)
I think that ChatGPT Plugins can be as big, or bigger, than the Apple App Store, and that most large organizations, whether companies or governments, will want to get on board with a plugin of their own sooner rather than later. And I think that we might be uniquely situated to do that for them given our initial forays into the space with the 311AI app described above and some other projects we have in development.Buildstack: Two big developments this past month.
First, we’ve shipped a pretty comprehensive design update to the site. It’s more white and more pink, and generally cleaner and more minimalist, which I think works well.
Second, we’ve found someone to join the team and take on the tasks I mentioned in last month’s newsletter. And that person is my wife, Madison. With her help, we’re going to demo and pitch our way to a $1-million annual run rate by the end of the year. Our momentum is slowly building.
real estate stuff
FH1*: We’ve begun foundation work for this project, including excavation and underpinning of the existing footings. I’ll take some pictures of our progress over the next few weeks to share in next month’s newsletter.
MR1**: We have a lot happening on this project, including, potentially, a move that would significantly increase its size. Unfortunately, I can’t say much more about that until we have some key details finalized.
What I can say is that property management continues to be an interesting challenge. Between City Bylaw enforcement notices for graffiti removal, disappearing tenants who just stop paying rents, relatively high turnover in commercial tenants, I won’t miss any of it when the time comes to start building.Other stuff: I’ve been digging into some potential land assemblies lately and am planning to ramp those efforts up over the next few months. I have a well-developed thesis around why now might be a particularly good time to pursue this strategy, but the thing about well-developed money-making theses is that they’re most useful when not shared.
That said, I am willing to share some of it with some of you, if you’re interested.
A land assembly, for those of you who don’t know, is the process of making offers to purchase contiguous properties that, when assembled, form a larger development site. For example, you might make offers to purchase seven contiguous properties featuring detached bungalows, knowing that when assembled, the land use planning context would support a highrise building.
Value is created through this process as the combined purchase prices of those seven properties, even if above-market, would still come in lower than the value of the highrise development site.
The first step in a land assembly is to identify a promising assembly target. I’m getting pretty good at that with the help of a few friends who think along similar lines.
The second step is to underwrite the assembly and make offers that are 1) sufficiently above-market so as to be enticing enough to accept, and 2) not so high that the combined purchase prices are greater than the value of the assembled development site.
The third step is to start getting those offers accepted and put up the required deposits.
Land assembly purchase offers are generally made with long conditional periods, to provide enough time to both complete the assembly and any required due diligence (including environmental site assessments and geotechnical and hydrological investigations). Within those conditional periods, all deposits are refundable. You can walk away from the deal and get your money back. Once conditions are waived, however, the conditional period comes to a close and the deposits are no longer refundable. That money is now at risk.
You then close on the properties and entitle the newly formed development site and build the new building and everybody’s happy.
But let’s go back to the deposits, as this is where you might fit in.
My brothers and I are raising a small fund to fund these deposits as we scale our efforts up. I skimmed over the last steps in the process but that is where much of the magic happens and where the money is made.
If that sounds at all interesting to you, I’d love to share more, including details regarding terms and that sort of thing for this deposits fund. We’ll be raising about $1-million with a minimum check size of $50,000. Holler.
*Forever Hold 1. A proposed four-storey multiunit rental building in Toronto’s west end that we plan on holding forever.
**Midrise 1. A proposed midrise multiunit residential building (tenure to be determined) in Toronto’s west end. We’re starting with a rezoning and will see where we take it from there.
stuff I’ve enjoyed
This is some of the content I’ve come across over the past month that’s worth sharing.
Article: This article on the economics of data businesses sounds like it might not be very exciting, but it is if you’re trying to build a data business (like Buildstack). It’s the most comprehensive breakdown of this business model I’ve read yet. Barring OpenAI eventually eating all of these businesses, I think they will continue to be the silent money-printing machines of the internet they have been for the past 20+ years. Shout out to Brett Chang for sharing it with me.
Book: I’ve started reading a bit more science fiction lately, including Isaac Asimov’s The Complete Stories. A recent favourite is The Last Question. The last question, first asked in the year 2061 to a self-adjusting, self-correcting computer called Multivac—which sounds a lot like a future version of ChatGPT—is: “How can the net amount of entropy of the universe be massively increased?'“ The computer replies that there is insufficient data for a meaningful answer. The question is repeated a few times across a few eras as humanity conquers the stars and eventually even death itself, but the heat death of the universe looms increasingly large. At each point, progressively more advanced versions of Multivac answer that there is yet insufficient data for a meaningful answer. Until, eventually there is. You’ll have to read the story for a great ending.
Podcast: I don’t love Pomp’s podcast, but Sam Parr on Pomp was good. It was the sort of discussion that gets you excited (pomped?) about hustling. On a related note, motivational self-help-like content remains underrated.
Video: I was on TV this past month for an episode of TVO’s, talking about the controversial topic of urban and suburban sprawl into Ontario’s protected Greenbelt. My position can be roughly summed up as: tall > sprawl > wall or nothing at all. Get it? You can check the episode out here.
***
And that’s all for now. Here’s to a good and productive April.
Feel free to reply to this email with any comments or questions. I love chatting about everything mentioned above.