Welcome to the first stand-alone issue of Off-Ramps! Today, and each Thursday going forward, I’ll highlight three-to-five interesting pieces that I think you will enjoy reading. For the most part, they’ll be on the same themes that our regular Tuesday issues explore: mobility, innovation, artificial intelligence, and progress.
Please enjoy these on your morning commute, or save them for your weekend.
, like me a Fellow of the Roots of Progress Institute, parses the debate going on right now between people who think A.I. progress is stalling versus those who think it proceeding as anticipated.
gives a good account of each side of the debate here.
Steve’s take is the right one, I think:
Who is right? Is AI advancing rapidly, soon to be able to do everything people can? Or is AGI still a long way off?
This is a false dichotomy. AI has not been plateauing; but AGI is not near at hand. Progress toward a destination is not just a question of how quickly you travel; it also depends on how far you have to go.
The things that the frontier A.I. models can do today are astounding. I use Anthropic’s Claude to help me draft e-mails, fill out forms, brainstorm, model other people’s thoughts and assumptions, and cook dinner; it’s already hard for me to remember how I did many things without Claude’s assistance. At the same time, the idea that by scaling up Claude’s capabilities a hundred-fold, or something like it, would permit the model to do everything a white-collar work can, at the same level of proficiency, still dramatically understates what a person can do; even if scaling laws continue to hold, an artificial general intelligence is still a long way off.
As Steve notes:
Here are some tasks you can’t yet hand to an AI: managing a room full of third-graders. Writing a tightly plotted novel. Troubleshooting an underperforming team. Planning a marketing campaign. Reminiscing with an old friend.
You might object that most of these are “cheating”. Of course an AI can’t manage a third-grade classroom, you need a physical presence for that. Of course it can’t reminisce with an old friend, it doesn’t know your history together. But that’s exactly my point: most of the time, when we evaluate AI capabilities, there are all sorts of tasks that we don’t even consider, because AIs aren’t yet eligible to attempt them.
Steve is an old tech hand, being one of the co-founders of the service Google bought and turned into Google Docs. He’s worth following if you are interested in A.I. and where it’s going.
Gothamist: New York’s rules on turbine transportation are blowing its climate goals out of reach. (Hat tip to Reilly Brennan)
New York state wants to fight climate change, in part, through wind power. But it feels no urgency on the subject, imposing strong restrictions on the movement of turbine parts by road. Because of the length and heaviness of turbine parts, deliveries must be escorted by state troopers, whom can only travel to a jurisdictional boundary; at which point the load must wait for a new trooper, and a fresh inspection. The delays are so great that one company, contracted to install 700 turbines, will need ten years just to carry out the deliveries.
This is the sort of story that makes progress-wonks, like me, want to tear our hair out. There are all sorts of things that New York state could do to facilitate deliveries: allowing more-frequent deliveries, allowing private firms to handle escort duties, or relaxing escort requirements. If the state actually believed that the green-energy transition is critical for fighting climate change and future prosperity, it would do so.
But the state doesn’t. Why not? As
has noted, “In America, you can do anything you want as long as 1) no one is inconvenienced, 2) no one can get hurt, and 3) no one can lose their job.” This seems to be an instance of 1). It’s frustrating.
Casinos make their profits not from casual players, but from regular gamblers who consistently bet large amounts. Such players are known as whales or high rollers. Another term for such people is addicts.
Casino gambling has been enabling addiction to their products for decades. Using the glamour of their settings and the thrill that casual players get from gaming, they attract, identify, and exploit addicts, blighting their lives in the process. None of this is new; here’s a grim article from 2016 that makes the case.
What is new is that now, everyone has a casino in their pocket, which has made problem gambling much worse.
Zvi Mowshowitz is perhaps best known today as an observer and prolific writer on A.I., but he came up in the world as a competive player of Magic: the Gathering, and has benefited from conventional gambling, both as a participant and on the business side. He reviews the results of the recent (post-2018) spread of legal, on-line sports gambling. He finds the evidence to be conclusive, and damning:
…ubiquitous sports gambling on mobile phones, and media aggressively pushing wagering, is mostly predation on people who suffer from addictive behaviors.
That predation, due to the costs of customer acquisition and retention and the regulations involved, involves pushing upon them terrible products offered at terrible prices, pushed throughout the sports ecosystem and via smartphones onto highly vulnerable people.
This is not a minor issue. This is so bad that you can pick up the impacts in overall economic distress data [emphasis mine]. The price, on so many levels, is too damn high.
But don’t people make bad financial decisions all the time, like buying tickets to a Taylor Swift concert that cost more than a new car? Certainly they do; but if a Swift concert tour correlated with a 28% increase in bankruptcies, as the introduction of on-line sports gambling seems to, I’d think such tours should be banned as well.
As a liberal, I want people to be able to pursue their own vision of the good life, and I have a great deal of respect for revealed preferences. But I have my limits. This is one of them.
There was an election in the United States earlier this month, about which I have nothing to say. There’s already ample discussion of the subject on the Internet, and I think that the marginal value of commentary from me is low.
So instead of offering my own take, I’ll simply highlight this piece. Joe Heath at the University of Toronto is one of my favourite thinkers and writers, so his comment on the election is worth reading, but it was this aside in the argument that really struck me:
Allow me to preface this by noting that the most common mistake Canadians make, when thinking about the U.S., is to imagine that we understand the place. Living there for a few years was enough to cure me of this delusion. America is a vast, sprawling, fragmented, deeply mysterious nation. The most important thing that I learned, living in America, is that I do not understand America, and probably never will [emphasis mine]. As Canadians we are, in a sense, almost handicapped in our efforts to understand America, because of the superficial similarities between our two countries, which generates a false sense of familiarity.
Like Heath, I have also lived in the USA for a few years; I also worked at Sidewalk Labs, a very American company that tried to do business in Canada, giving me a close-up view of the differences between my American and Canadian colleagues.
I think Heath is absolutely right. Most Canadians speak English and consume a great deal of American media; many visit the USA often for brief trips. They therefore assume that the USA is a known quantity to them. They are mistaken about this. Toronto is the most American city in Canada, and yet life here is dramatically different than life in Chicago or New York, Toronto’s closest Stateside analogues. Failure to understand this will end in tears; it certainly did for Sidewalk Labs.