It's getting a little tougher for the youngs
Good thing we don't have that many of them, amiright? Plus, some AI tidbits
Begin with a little pleasant tiding, and then on to the (accidentally) less pleasant stuff
Grokking Grok and the price of tokens
New England was a YIMBY paradise
Wither the teenagers
“Get a job!” “We’re trying, but we can’t.”
👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. Grokking Grok
First, the good news.
X.ai released a new Grok model and apparently it’s awesome.
Like, it’s the best at everything (for at least a few days):
No model has ever been the best at everything all at once, until now.
And then there’s the benchmarks:
I couldn’t tell you what all those mean, but Grok is apparently the best at them.
Here’s some actually useful information. And yes, I used Grok-3 to make a mini-version of astroids, and it’s really that easy. Say what you will about Elon’s flaws (and he’s got ‘em), but he routinely pulls off what would be career-making accomplishments, it feels like every few weeks.1
This stuff is developing very quickly, and it’s getting noticeably better (especially for someone who stopped using ChatGPT for a while because the mistakes were too many).
Plus, costs are going way down:
Median token costs are a tenth of what they were two years ago.
Good times.
New England, the YIMBY Mecca
And now on to the other stuff.
If you’re a housing shortage bro, the Northeast and New England are poster-children for NIMBY-excess.
All the housing shortages are because those lousy NIMBY boomers throttled supply with their zoning, and stuff. BUilD MoAr HoUSIng, bro. It’s simple.
You would also be surprised to learn that New England consistently permitted more housing per increase in population than the nation, as a whole—at least, until the GFC.
As per a research paper from the Boston Fed:
From 1980-2010, permits per 1,000-person changes in population ran ~30% higher in New England than the national average.
It was only following the Global Financial Crisis (and the “housing bubble”) that New England began permitting less than the rest of the country.
Compared with the nation, New England issued more permits per 1,000 change in population from the 1980s through the 2000s, but this differential reversed in the 2010s.
New England’s decline in permits per 1,000 population from the 2000–2009 decade to the 2010–2019 decade was driven by Massachusetts and Rhode Island, where the rates declined sharply and fell well below the US average. Over this same period, changes in the other New England states varied, including a sharp increase in Connecticut and a substantial decline in Vermont.
More accurately, Massachusetts and Rhode Island fell way behind, while CT stormed ahead.
So, why did New England stop building as much?
If you’re a shortage bro, presumably you blame the Post-GFC slowdown in permitting on the outbreak of NIMBYism that definitely happened, albeit secretly, and without a trace. NIMBYism is sneaky and evil, like that. Before 2010, by contrast, New England was a YIMBY paradise, which explains why it permitted above the national average.
Alternatively, if you’re a sane and reasonable person, you might conclude that supply and demand is a delicate dance, but builders build where people are, but also where they are going to be, and it was right around the GFC that population growth began to roll over, and it became apparent that perhaps there were more new houses than there were people to buy them. Permitting slowed because people-growth slowed (and it would get slower still).
Rather than a zoning-driven-shortage, you might further conclude that zoning has little to do with it, and in fact there is no shortage, because supply appears quite responsive to demand (even if the former is somewhat less volatile than the latter).2
Idk. It’s a tough one.
No Country for Young Teenagers and Adults
Fortunately, the fair-weather YIMBYs have not yet had their way.
Because on the subject of people growth (and the Northeast), here is one of the more depressing things you’ll see, hopefully in a while:
The expected change in high school graduates is largely negative for everyone but the sunbelt, and DC(?)
New York and CA will shrink their graduating high school classes by 27% and 29%, respectively.
Now, to be fair, 2041 is still some time away, and projections like these are pretty squishy, but . . . yeah, that’s pretty grim.
No jobs for young teenagers and adults, either
As a mostly unrelated aside, Guy Berger reiterated a prior observation about young folks in the labor market: recent grads are having an unusually difficult time entering and re-entering the job market.
Here are unemployment rates for BA-only, cut by age:
Unemployment for recent college grads is well-above the pre-pandemic trend.
And again, for the most part, these are new-entrants and re-entrants, not people getting laid off. If you have your job, you’re not losing it, but getting a job, is getting awfully hard.
The duration of unemployment (for new entrants, especially) is escalating quickly:
The median duration of unemployment for new entrants is almost at the pandemic peak.
Note that these are not just college grads, so this would include “weaker” job candidates, like non-college and illegal immigrants, but you get the idea.
Why is this happening? I’d chalk it up to White Collar Stagnation and generally tepid demand for new workers (because remember, the job market has been tight, not strong), but who knows. Maybe recent grads aren’t up to snuff? Maybe it’s something(s) else entirely? More investigation needed.
It’s worth watching, if only because the last times hiring declined without much firing were during the “pre-recessionary” phases of 2000 and 2006.
Have a nice weekend!
Other links
Tech in 2025: The Cycle Rolls On. Goldman’s bullish take on demand for tech and IT services in 2025
ETSY tumbles on revenue miss, pullback on discretionary spending, and competition from China. As expected. Etsy is selling niche nice-to-haves, while competing with Chinese super-discounters. Double headwind. No contest.
Labor shortages straining employers. Minneapolis Fed on the secularly tight labor market, boomer retirees and the the lack of young folks to replace them. Migrants, though, were the primary source of replacement workers (in Michigan especially).
Previously, on Random Walk
Random Walk is an idea company dedicated to the discovery of idea alpha. Find differentiated data, perspectives and people, and keep your information mix lively. A foolish consistency is the hobgoblin of small minds. Fight the Great Idea Stagnation. Join Random Walk. Follow me on twitter. Follow me on substack:
Best AI model? Best EVs? Best satellites and remote wifi? Best rockets? An actual telepathy gadget that lets paralyzed people move things with their brains? All the same guy? I mean, come on.
You might also observe that rather than any shortage, there was a sudden and unexpected surge in demand for low-density housing, a preference shift that moved far more quickly than anyone can reasonably build houses.
Before the pandemic, net-migration favored higher density places. After the pandemic, net-migration favored low density places.
Except Rhode Island, which no one liked before or after the pandemic, apparently.
Massachusetts has a zoning override bill, Chapter 40B. The majority of market rate housing in Massachusetts is built under this provision.
High school graduate projections by number aren't nearly as concerning when you consider birth rates have fallen cumulatively ~30% from 2005 to 2022. Given the projection period is 2023 to 2041, it makes sense that total graduates over this period would fall relatively in line with the birth rates T minus 18.