👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

The Open v. Closed Debate Is About Something Else

Last week, the Chinese AI lab Moonshot released an open-weight model called Kimi K3 that, for the first time, scored higher than the latest and greatest frontier model.

It’s a big deal because, among other things, if open-weight models are just-as-good-and-maybe-better than closed US models—and served at a fraction of the cost—then it calls into question the entire sustainability of the frontier lab approach. Quite simply, the OAIs and Anthropics of the world cannot continue to spend billions of dollars training new models, if those models offer minimal competitive advantage. That’s a big “if,” however, but let’s just take it as an underlying premise of the zeitgeist.1

Even then, there were still a few open questions:

• is Kimi actually better? Is it actually cheaper? On which dimensions? What does “better” mean, really?

• How did Moonshot pull this off, while spending far less on compute? Did they actually train a model at a fraction of the cost, or did they distill a model from Fable, i.e. prompt Fable to divulge its reasoning secrets, and use that to release a knock-off?2

Since then, there’s a bit more clarity, but also still more questions:

• yes Kimi appears to be pretty good, and relatively cheap, but the reality of “cost per token,” is a bit more complicated than sticker price;

• yes, Kimi definitely appears to be a distilled version of Fable, plus some differences, of course, but consensus is that distillation carried the laboring oar.

What happened next was a thoughtful, well-reasoned discussion “bombast,” and then “counter-bombast,” shortly thereafter.

The initial bombast salvo came from the Treasury Secretary variously declaring that the US Gov’t was considering an effective ban on [Chinese] open-weight competition:

The counter-bombast came in the form of VC-posting, mostly accusing the big labs of seeking regulatory capture:

. . . and then eventually a joint letter, authored by Jensen himself, extolling the virtues of open-source competition. Since then, Zuck published an op-ed declaring AI for the Everyone (behind a paywall, no less), and the “yay, open source! boo labs!” discourse mostly continues apace. Mr. Market, too, appears to have some concerns about where this all may lead.

To my mind, these are all variously ships passing in the night.

ICYMI

No one has actually offered a real objection to a government intervention about distillation

The relevant debate isn’t really open v. closed.

It’s about whether something ought to be done about distillation, specifically the kind of distillation that most people seem to recognize as “a knock-off.”

And even there, rather than a sane discussion, what has transpired seems primarily anchored on whether one has a rooting interest in the possibility of mega-labs or not. Or rather, the vast majority of the criticism directed at team “do something about chinese distillation,” seems quite convinced that “doing something” is motivated by the labs’ desire to save their own skin, ergo, the critics seem to believe that “not doing something” spells doom for lab margins, which appears to be an outcome that they would rather prefer.

You, see what’s most important here is that I confirm my priors:

Yes, surely all that has transpired can be viewed through the lens of “I told you so.”

Nooo, Random Walk, you’re being unfair—this is an assault on the open source tradition! We are merely defenders of the open and free internet!

Really? Is Nvidia’s CUDA open-sourced? Lol, no. And Meta has been developing open models this whole time, and no kerfuffle ensued.

The issue is plainly not open v. closed.

It’s a specific kind of open, i.e. the kind that’s released by free-riding on the billions the labs spend on training. If Meta released a frontier open model tomorrow (or last year), trained on its own substantial dime, people might still be concerned about lab margins, but I’m fairly certain that the Treasury Secretary would not be involved.

If the Government does IP enforcement, that’s not unusual

OK, you’re right. It’s not really about open qua open. But, have you considered the horror of the federal government going out of its way to protect IP interests of individual companies?! And if there’s an IP interest, then why not articulate the specific cause of action?! IP is merely a facade for rote protectionism!

Wrong again. There are countless examples of the US Government leveraging its might to protect the IP interests of US companies, especially tech. Off the top of my head:

• every movie you used to watch opened with the FBI warning on copyright protection, including the threat of prosecution;

• counterfeit handbags are are routinely seized and destroyed at the border;

• back in 2017-18, the government undertook a massive investigation into Chinese tech transfer and IP practices, which led to the first wave of tariffs on China;

• back in 2007, the US initiated an entire WTO litigation around China’s incredibly lax IP enforcement regime (which it won);

• the FCC just added advanced robotic devices and power inverters to the Covered List banning import sales

OK, you’re right again. It’s not that unusual for the government to take an interest in IP enforcement. But then what IP has been violated? What’s the theory?

This is a slightly better objection, but one that I find slightly puzzling, and ultimately, not that compelling.

It’s true that no one appears to have articulated an actual theory of distillation infringement. And yet, everyone seems to “know it when I see it” understand the difference between “this kind of distillation is fine—because I’m not just planning on releasing a knock-off frontier model,” and “uhhh, we wouldn’t do that kind of distillation bc it’s just a path to a knock-off and pretty obviously against the lab T&Cs.”

The best evidence that everyone seems to know it (other than my own conversations with people) is that US companies are not releasing competitive open-models that are, in fact, close knock-offs of the frontier models. I strongly doubt that US companies are incapable of such a thing, it’s far more likely that they fear the legal ramifications of doing so.3

The point here is that if people seem to be able to readily distinguish between “ok distillation,” and “definitely-not-ok distillation,” then there’s no reason that that the law can’t make that distinction, as well. And if it doesn’t yet, then it could, quite easily. Will there be some uncertainty at the margins? Yes, of course. There always is.

Should it? My weakly held prior says “yes,” as I’ve quipped before, if ever there was a time for IP lawyers to shine, that time is now. But, IP is not my forte, and neither is distillation, so I could be moved in the other direction (just not by the arguments presented thus far).

Yeah, but we really want cheap AI

So, to recap up until this point: clearly, the labs aren’t categorically opposed to open models, and it’s neither unusual for the government to take an interest in private IP rights, nor does it seem beyond the ken of IP law to circumscribe “bad distillation,” specifically.

So, why shouldn’t the US government get involved, here?

Fine. The real reason is that cheap AI is good. Why should the entire ecosystem be beholden to the labs? Just two companies to rule the world? And besides, karma’s a bitch—the labs trained on the world’s IP, so what goes around comes around.

Yeah, I get the Karma point, but (a) the labs are now paying for a lot of that training data; and (b) there’s a big difference between hoovering up zillions of diffuse and mostly worthless (on their own) data points, and just ripping a highly concentrated and very expensive undertaking from a single source. Plus, it’s not like the labs have no competition—Google, Elon and Zuck are hardly what anyone would call trivial upstarts.

Mostly it seems like you just want cheap AI in the same way that many of us used to enjoy pirated music, which, however unsympathetic the record labels may have been (and however much other odious lobbying they have done), wasn’t the most persuasive argument.

Basically, I find the general objections around “distillation is bad and the gov’t ought to do something about it,” to be pretty make-weight. Here’s a pretty good example, courtesy of Benedict Evans, of the sort of make-weight reasoning I’m referring to:

• Tech “rolls their eyes” at the T&Cs? I mean, maybe they do, but they’re still not violating the T&Cs, and it’s sort of besides the point.

• That other forms of distillation also take place, and yeah man, anything is distillation when you really think about it, is again, neither here nor there.

What we’re talking about is a specific kind of distillation that everyone appears able to distinguish as different from the other kinds, and they’re perfectly capable of not doing it, especially when they fear some repercussions.

There may be other reasons why this isn’t an IP enforcement fight worth fighting (or some other regime), but these don’t seem like it.

Some people (like my friend Gil Dibner) think distillation just can’t be stopped. OK, that’s a possibility. Other people seem to think Anthropic could stop it easily . . . if they were willing to cut off big paying customers. Idk.

The bigger thing unsaid and under-considered

OK, so the outrage directed at the proposed ban on Chinese open-weight models is poorly considered, and perhaps motivated by other things. Fine.

Read more

• a chart from the future no one would predict what’s next

• bigger, longer, more profitable

• healthcare domestic product fraud strikes again

• China distills models that don’t even exist, yet

• Apollo’s Slok joins team Random Walk

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

1. An Epic Inversion

Sometimes it’s fun to make present-day observations and wonder what you might have said in the past had your future self gone back in time to give you a glimpse of the future.

This is one of those observations:

Long-dated rates are climbing a wall of worry, the world over.

If someone showed you this chart back in 2019, and asked you to predict the next 5 years, what do you think you would have said?

I’m pretty sure most people, myself included, would have sold everything and battened down the hatches for whatever catastrophe was driving borrowing costs (and presumably runaway inflation) to the moon.

As it happens, had I or anyone else gone “risk off,” I would have missed out on an epic run in public equities. It turns out that predicting the future is hard, even when you know the future.

ICYMI

2. Bigger, Longer, More Profitable

I don’t have any profound takeaways here, but these are both pretty striking examples of American Exceptionalism.

It turns out that the most profitable companies are staying most profitable for longer than before:

More than 60% of high ROE US stocks remain in the top ~20% of ROE five years later, a share that has doubled since 1990.

Part of the reason why the largest US techcos command such powerful multiples (even if less so, lately) is their epic ability to deliver historic profitability at unprecedented scale, and just keep on doing it.

Most valuation metrics in most markets are near the top of their 20-year range.

OK, so the US isn’t the only expensive market, but all those markets together are roughly equal the size of the 10 largest companies in the US, so it’s really no comparison. There just aren’t that many companies who can both absorb huge amounts of cash, and deliver that kind of enduring performance.

Of course, the AI buildout is going to test some of those ROE bona fides—and it’s going to be up to the credit markets to make a lot of that happen—but that doesn’t make past-performance any less impressive. 1

Here is the second one.

US business cycles are getting longer, or perhaps more accurately, the peaks are getting higher:

We’re currently in the 6th longest expansion cycle in our history, although query what the counter-factual would look like absent the non-economic pandemic recession.

Truthfully, it seems like the 80’s is really when the era of long cycles began. Maybe that’s a coincidence. Maybe it has something to do with the rising prominence of fiscal and then monetary (and fiscal) policy. Or maybe it has to do with speed and spread of information, keeping shocks and unwanted surprises to a relative minimum. Idk.

I just thought it was a notable observation.

3. More Healthcare Domestic Product Fraud

Random Walk was pretty early to the central role that home healthcare aids have played in the “private sector” job growth/catch-up story.

Random Walk probably wasn’t early enough in grasping the fraud implications, but with the benefit of hindsight, it seems pretty clear what happened: extremely weak eligibility and oversight controls unlocked an explosion of reimbursable services to create the illusion of “growth,” at the tax- and rate-payer’s expense. And yes, this exploit was especially popular in immigrant (legal or otherwise) communities, and yes, public officials were almost certainly in on the “Medicare fraud as a jobs program for migrants” plan, because how could they not be?

Anyways, it turns out that home health is probably not the only reimbursable exploit that flourished over the past few years.

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

K-Shape Goes Poof

The notion of a “k-shaped” economy just won’t quit the zeitgeist.

Whenever there’s the barest inkling of a k-shape, everyone seems very eager to opine on however which ways the middle- and lower-classes are falling behind the wealthy.

It’s hard to otherwise explain why the K-Shape is such a recurring theme, other than the obvious slopulist appeal. Random Walk has touched on this before, but the K-Shape is structurally implausible—the shape of the labor market and other cyclical factors favor the bottom, relatively. But, also, the data doesn’t support the K-Shape, either.

Did you know that lower-income households are the only post-pandemic real income winners? Well, it’s true (although other data sets vary a bit):

“Yay, the lower and middle classes are doing pretty great!” is apparently a dog that will not hunt [headlines].

Anyways, it’s bank reporting week, so it’s a good occasion to remember that the Almighty Consumer is A-OK, there is no K-Shape, and while there is at least one curiosity, there may be a contra-Random Walk explanation.

First, from the banks:

Everyone paints a pretty rosy picture of the Almighty Consumer. Only Wells Fargo hints at anything K-shaped, but (as per below), I strongly suspect that’s an artifact of higher gas prices.

In general, the banks are doing pretty well. Credit has grown, they’re making a ton on trading fees, an un-inversion has been good for the biz.

Also, I love a good bedtime story from the most consequential man in global finance, Papa Jamie Dimon:

“It’s great, but it could be terrible, any moment.”

Thank you, JD, for those illuminating comments.

ICYMI

Started from the bottom (reprise)

As for the data itself, well, whatever K we might have had, has become an E, which itself has become just a squished line.

It starts with hiring and wage growth.

Payrolls, looking pretty good:

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

Capital Intensity🤝Capital Light

Continuing on the theme of regime change, but shifting gears.

The rotation to tech continues to embigger, but within tech, the laggards have become the leaders. Low-capital software—the darling of the old regime—has given way to high-capital hardware.

The rotation from capital light to capital intensive is something Random Walk has written about before, but here’s a few more charts on that theme, plus a little wrinkle.

Delayed onset regime change

First, one “table-setting” observation on the current cycle.

While the inflection points for new market regimes historically tend to be recessions (reflecting the collapse of the prior regime), in this case the “regime change,” catalyst wasn’t the pandemic or the pandemic “recession,”—it was AI.1

Not that that’s very controversial (even if the pandemic was a bit oversold as a behavioral shift at the time), but the NY Fed researchers have a neat visualization to that effect:

If you aggregate the total post-recession change in sectoral contribution to output (i.e. the extent to which different sectors contribute more or less to output relative to pre-recession), then the “economic reallocation” post-pandemic was both less substantial, and less-enduring that the previous post-recession recoveries.

• After Dotcom, the economy became asset-heavy and financials-oriented (i.e. financing the labor-intensive housing boom).

• After the GFC, the economy became semis- and then software-oriented—the iphone, apps, cloud and saas booms.

• After the covid-recession, the economy didn’t actually change all that much . . . until year 4 (2023), when rates went up, and the AI Capex surge began shortly thereafter.

And boy did the Capex rotation change some things.

At the sector level, you might say “nothing has changed,” because tech itself is still running away from the field:

Fwd ROE for tech has continued to accelerate, while the world-ex-tech is basically stagnant.

Notice, however, the ZIRP-peak in ‘21, followed by the rate-hike valley in ‘22, followed by the ‘23-present reacceleration. Something definitely changed, but the change is coming from inside the Tech.

ICYMI

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

The Shift from Token-Maxxing to Token-Optimizing is Good, Actually

Random Walk has previously mused on “frictionful intelligence,” i.e. the emerging conflict between the rising cost of frontier models, relative to the returns on using the latest and greatest. That, plus the persistent threat of open-weight models—where Random Walk maintains a legal/political solution is necessary and inevitable—offering not-quite-as-good performance at a fraction of the cost, have combined to put some pressure on the perceived long-term sustainability of all this AI capex.

Simply put, if customers won’t pay the labs, then labs won’t pay the hyperscalers, which means the hyperscalers won’t pay the chip and memory makers (or data center builders), and the whole thing comes a tumbling down.

That’s true so far as it goes, and we’ll start with a few more charts to illustrate the point, but there’s more to the story, of course.

Non-frontier intelligence is getting cheap quickly

First, the price of intelligence is falling far more rapidly than the comparable fall of PC prices (back in the ‘80s):

The productivity gains of the PC-revolution were lagged, but definitely favored the early-adopters—and when it came to PCs, early-adoption was handicapped by cost, much more so than with AI.

Second, and relatedly, open-weight models are a good part of the reason why the cost of intelligence is dropping so quickly:

Note the log-scale on the x-axis: minimax open-weight models are nearly comparable to Sonnet 4.6, at a small fraction of the cost.

Here’s another look at the same thing:

The cost of tokens for SOTA models is exponentially higher than the next best thing—by the time you get to the leading Chinese open models, the cost difference blows out by a factor of ~10.

Little wonder, then, that Chinese open models have pulled away, recently:

By total token consumption (per OpenRouter), Chinese models have jumped ahead by ~5T tokens, after running more or less lockstep with US models since the beginning of the year.

Obviously, if frontier models cost billions to train, and then open-weight models come along a little while later with a nearly-as-good alternative, that cannot go on forever.

Again, while I can’t be confident on the technical details, it seems very likely that open-weight models are keeping pace through a form of “distillation,” that most people would recognize as “stealing.” The tricky thing is that the industry does not consider all distillation to be bad—there’s the sort of good faith distillation that everyone wants to preserve (i.e. some combo of “fair use,” and “transformation,”) and then there’s the rote copying that exists to cannibalize the frontier models (i.e. “piracy”).

If ever there was an opportunity for IP lawyers to shine, that time is now.

Or maybe that’s cope . . . but how else can it possibly make sense that the frontier labs need billions of dollars worth of compute to train the latest-and-greatest, and the Chinese can produce near replicates a few months later? Now, if a Chinese lab releases a true frontier model ahead of the labs, well, that would be something . . . but that hasn’t happened yet.

ICYMI

AI Demand Is In the Eye of the Beholder

In any event, if you want to venture a gander at measurable AI demand, here’s what I got for today.

The nightmare fuel is in the eye of the beholder:

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

Of Economic Regime Change, and Anti-Tech Sentiment

Some thoughts on the second-order effects of the outsized premium to doing more with less. It’s a somewhat speculative, long and winding tale.

Soft-hiring hits the new entrants hardest

Jobs day release last week was all over the place, so it’s really not worth trying to interpret it. This is as good a recap as any, but my instinct is just to say “noisy data.”

Elsewhere, however, there is some interesting data on one of the key labor market themes: the “broken” on-ramp, i.e. depressed entry-level hiring, especially for the college degreed.

It’s the one area of actual “softness” in an otherwise tight and supply-constrained labor market, where the existence of any slack whatsoever is worth considering. That’s by far the most significant aspect of the observation, but there’s some secondary import, as well. If you wanted to say “ai is a jobkiller,” then the decline of entry-level hiring is basically the only data you can point to and say “see, told you so.”

As for what it all means, to my mind, the pain for new entrants is downstream of general softness in demand, plus aging-in-place, combined with perhaps a small dose of misallocation. It’s also possible that remote work has created some optionality for firms, such that they can access talent all over the world, and therefore have less of a need for a steady pipeline of raw talent.

More simply, now that everyone is bottom-line oriented (and outside of AI capex and healthcare, secular tailwinds are hard to come by), the inclination to add payroll is muted. The negative effects of soft demand land on new entrants, first and foremost, and the fact that existing workers are working longer, only exacerbates the problem.

The St. Louis Fed recently published some research to that effect, i.e. job openings (as a proxy for demand), are most predictive of new-entrant participation in the labor force (for better and for worse):

Downshift in job openings makes the biggest contributions by far to changes in all of employment:population, U/E%, and LFPR.

Basically, if firms aren’t posting new jobs, it’s the youngest members of the workforce, that feel it most acutely. When demand softens, firms don’t fire birds-in-hand to replace them with young blood—they just stop hiring young blood.

That’s the theory, at least, and it seems plausible, enough.

Entry-level Thaw?

The good news, is that there is more evidence that things are turning the corner.

As per Revelio’s data, summer internships are tracking well-above the previous two years (with some pretty big gains for tech, so it’s not all low-paying work).

ICYMI

To be honest, I have no idea whether internships are a leading signal of anything, but more has got to be better than fewer, right?

And for those who would blame AI for the entry-level malaise, we’ve got some data for you too.

If AI was replacing the need for jr. hires, then how do you explain this:

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

Online shopping is all the shopping

Two big things that really haven’t changed in an awfully long time: the Almighty Consumer is (1) A-OK; and (2) a huge fan of online shopping.

Prime bonanza

Prime Day came and went and, as expected, it was a smashing success:

Adobe forecast the first $300B non-holiday quarter for online shopping, and calculated a $26.4B shopping bonanza for Prime Day(s) (June 23-26) alone. Hooray for buying cheap tchotchkes on the internet.

Payrolls look peachy

Whether it’s tax refunds, falling gas prices, and/or (most likely) rising employment and wages, the Almighty Consumer has some serious pep-in-its-step.

The JOLTS release was all over the place, so it’s hard to find too much signal there, but this, otoh, is always a good sign:

Withheld income and employment taxes grew at a very healthy clip back in May.

Yes, real incomes have been negative, and yes, the savings rate has dropped in-kind, but there’s no shortage of consumer confidence, at least when it comes to actual dollars and cents.

Prices and spending have run ahead of incomes, so it’s not entirely surprising to see the savings rate come down—and it’s been tracking lower ever since the growth of real incomes began to taper in ‘24. That said, deposit levels are still higher than the status-quo ante.

The point being is that all signs point to: the Almighty Consumer is just fine.

ICYMI

AI Shopping

But, back to the story at hand, the growth of ecomm is a thing to behold:

Read more

• ai agent usage, and the emerging top 1%

• bubbles? no bubbles…indecision, momentum, hunting for the triple-digit returns, high premium to growth, but bubbles?

• the K becomes an E

• mind the supercore (reprise, ad infinitum)

• no country for youngs, or the education bag-holder

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

1. The Token-Maxxer 1%

One of the more fun things about the “new media” turn, “go direct,” and the general disintermediation of tradmedia when it comes to company-storytelling is that a lot of high profile companies are hiring in-house economists, and those economists publish fun stuff with all the data that these companies have.

Anthropic, for example, just released a report on all kinds of things Claude-related, chock full of fun nuggets like this one:

It’s kind of amusing that “casual conversation” soon gives way to “sleep advice,” which peaks at ~7am, at which point it’s plainly too late to seek advice.

This, too, was amusing, in the sense of extremely on-the-nose stereotype reinforcement:

When it comes to using Claude, men tend to get work done, code, and build automations, while women . . . prefer to chat. But, it’s a signal of “collaborative engagement,” says Anthropic. Lol.

To be fair, women were only 12% of the sample, but still, it’s funny.

OAI, for its part, also released a report, focused on the “shift to agentic AI.” It’s also chock full of interesting stuff.

The gist of the thing is that more people across more functions are using Codex (rather than the web interface), and that token outputs are rising alongside (implying more agent usage). Naturally, OAI employees have way-above normal usage trends, but outside the org, similar patterns are visible.

The most interesting observation, however, is this one:

Putting aside the most recent June data, while everyone is running more agents (measured by cumulative daily active runtime), the top centile of users are really running more agents.1

There’s a step-change at the tail end of the distribution, where the gap is both substantial, and growing.

Here’s another cut of a similar observation:

Per Ramp’s spending data (so take it for what it is), the top 1% of token spend/employee is more than 10x the top 10%. Once again, the gap is big and growing.

All of which is to say there’s more fuel to the “AI Power User” fire.

The tech is real and transformative, but a small number of users appear to be way out in front, and they are compounding their lead. And that’s the thing about compounding exponentials: small differences become very big differences very quickly.

And therein lies some of the trickiness of finding “AI winners,” or the broader question of observing “AI ROI” in the 10,000ft picture. (1) If gains are concentrated then, by definition, diffuse benefits will be hard to observe (at first), widespread margin growth, notwithstanding; and (2) if you thought we had pareto distributions before, well, you ain’t seen nothing yet.

ICYMI

2. Not Bubbly. Indecisive, volatile, choosy, but not bubbly

Returning briefly for a moment to the totally-not-bubbly-but-rather-highly-discerning-and-somewhat-uncertain stock market, considering the following.

The Mag7 continued their underperformance relative to the Nasdaq100, but this week, they took it to a whole new level:

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

MBA Says “There Is No Housing Shortage”

As longer-term readers know, the wrongest thing that the most people believe to be true is that there is a “housing shortage.”

There is no housing shortage.1 If anything, there is a moderate housing surplus. That’s what the data shows, it’s what common sense would predict, and it’s what all the large homebuilders are screaming at the top of their lungs.

Nevertheless, defections from Team Shortage are rare, but they do happen, and the latest to declare “so, actually, about that shortage . . .” is probably the highest-profile defector to date: The Mortgage Bankers Association.

The MBA published a report on housing supply/demand dynamics, and they did not bury the lede, even if they soft-pedaled it, ever so slightly:

We argue that the post-financial crisis narrative of a persistent housing shortage may no longer accurately describe market conditions over the decades ahead.

Query whether it ever accurately described market conditions, except perhaps for a brief moment following the GFC, but the MBA is quite clear that the housing shortage has been cancelled.

Why? It’s a combination of demographics, immigration flows, and household formation, but the tl;dr is that the stork is no longer on our side, and the young ‘uns have abandoned any semblance of adult milestones, making demand soft-to-softer.

Household growth is slowing beyond the previous decades (even in the high-immigration scenario, eventually):

Fewer new households need fewer new roofs overhead.

All projections are hand-wavy, but compare the estimated net-new demand, with the estimated net-new supply:

• Demand:

  

• Supply:

  

11,340 new units needed is smaller than 12,651 units delivered, if I’m doing the math right.

We estimate a growth in supply of between 10.6 million and 14.6 million units from 2026 to 2035 . . . our estimated change in demand from 2026 to 2035 falls between the low estimate and the series average estimate.

This suggests that the growth in supply could exceed the growth in demand by 2035 depending on the trajectory of supply. This relationship holds through 2045, all else equal.

Yes, it appears I did the math right. “Supply could exceed the growth in demand by 2035.”

The MBA is an industry trade group with a lot of incentive to be strongly Team Shortage. Kudos to them for keeping it real.

ICYMI

Another builder says “definitely no shortage”

The truth is, supply likely exceeds demand, right now.

Read more

• binge-drinking is popular with some people

• a different theory of private for longer

• frictionful AI, and more tales from the ROI unlock (or not)

• ETFs to the moon

• China’s retail winter

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

1. Go Ahead and Binge-Drink

It’s father’s day this weekend, and while it’s true that drinking, in general, has been in steady decline, it’s not true for everyone.

Alcohol spending has dropped ~15% below the pre-pandemic baseline, but at least the decline appears to be stabilizing.

The number of binge-drinkers has also declined, but it’s the young-folks doing all the tee-totaling:

There are 3.2M more binge-drinkers aged 50+ than there were back in 2014.

So, maybe as a paean to your aged pa, tie one on for the weekend. Or something.

ICYMI

2. A theory about ‘private for longer’

Generally, there are two components to the observation that private companies are staying private longer.

The first reflects something of a structural break: private cos are ripening on the vine longer because the bid-ask on an exit price is too wide. Funds would like to sell, but not at prices that anyone is willing to pay.

The second reflects something of a structural shift in the comparative advantage (or not) of public markets v. private ones: companies are staying private longer because private capital has risen to meet the primary public value proposition (i.e. the sheer magnitude of capital needs), and all else being equal, being private is better because there’s more alignment with investors, less QoQ volatility, less regulatory tax, and less short-termism.

In other words, why go public if you don’t have to?

The net-result of all of this is that an increasing amount of enterprise value is accruing in the private markets:

~50% of the appreciation that used to happen in public markets is now happening on the private side (according to this analysis, at least).

Query how much of that shift is driven by the rate regime shift (that drove peak private valuations, only to be followed by a drawdown on the public side), but either way, the point stands.

There is, however, a somewhat different theory about private-for longer that I recently discovered, and it has to do with disclosure regimes.

The idea is that companies stay private to avoid sharing their intangible assets with the competition

Companies rich in intangible assets – proprietary technology (e.g. reusable rockets), production methods and know-how, advanced algorithms, etc. – want to avoid public disclosures that hand rivals (or would-be competitors) a template for expropriation . . .

Private markets resolve the tension through nondisclosure agreements. Investors can learn everything they need to underwrite the business, while preserving a degree of confidentiality unavailable in a public context.

The evidence for the theory is “life sciences,” where the exception proves the rule.

Life sciences enjoy a patent monopoly on their strategic goodies, and lo’ and behold, they go public a lot sooner than the rest of the market:

Life science companies have actually been getting younger since 2015, even as the “maturation gap” blows out ~4-6x wider than before.

So, where disclosure comes with IP protection, companies are happy to go public. Where it doesn’t, they do not. So the theory goes.

I’m not entirely persuaded by the theory because life sciences are different for a whole slew of other reasons, including having their own niche pocket of the capital markets, as well as a more binary set of outcomes . . . but, I like new theories, and this one was kind of fun.

ICYMI

3. Frictionful Intelligence (reprise)

Some quick follow-on observations regarding the propensity to spend on AI, what it means for AI to drive value, and expanding the surface area of demand via the “real world,” as a crucial unlock.

Mass market tapped?

First, on the mass consumer point, Random Walk posited that it seemed likely there was less pricing power and/or room to goose to demand.

As it happens, there’s some data that households “making AI payments” declined ever-so-slightly since January:

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

Reality Bites: Frictionful Intelligence

In what seems like ages ago, Citrini published a long report about how the cost of “frictionless” intelligence was going to collapse to zero, and then everything would go to hell in a handbasket.

At the time, Random Walk was polite, and said “Citrini Asks All the Wrong Questions,” but really the report was kinda dumb, especially on its core premise. The notion that compute was “frictionless,” is exactly wrong, and if anything, all the hard questions are around the impact of that friction.

Indeed, one of the biggest differences between AI and tech-splosions of yore, is that AI is not zero-marginal cost. Every push of the AI button costs money, such that selling AI-native software consumes gross margin.

Likewise, the real concern isn’t general stagnation once the cost of intelligence goes to zero, but rather that the exceedingly high cost of intelligence would grow more quickly than the quantifiable gains that would justify paying for it—or, perhaps more precisely, the incremental costs of model improvements are higher than the incremental gains of better models to enough customers.1

• In that case, if LLM prices went up commensurate with costs, that could negatively impact demand (and create an “overhang” of expensive model development without an end-market).

• Conversely, given competitive pressures to “race to the bottom” on price and maintain the leading model (and stave off open weight competition), frontier labs may have a hard time passing their development/compute costs on to customers, even if those customers could justify paying more.

Either way, customers (and/or investors) have to be willing to pay the labs, for the labs to pay the hyperscalers, for the hyperscalers to pay the chip- and datacenter-makers. If not, the music stops, or at least, slows.

The downside fear isn’t “dark fiber,” (where cables are laid, but unused, as per Dotcom), but under-priced fiber, that goes dark(er) once real costs (including energy) are passed-through.

Again, to analogize to cellphones, lots of people surely would have had uses for the first gen big clunky things, but not at prices that would have justified making big clunky things for a mass market. That’s why, in the beginning, only the stereotypical “rich a-hole from an 80’s movie” had one. Eventually, phones got less clunky, added a lot more utility, and became cheaper, and the mass market was, perhaps, bigger than people even imagined. But, all the while, investment scaled with market size to keep things relatively in check.

When investments run ahead of the actual realizable market, that’s what you might call a “bubble,” and obviously that’s concerning.

Things that are not concerning, on the other hand, are the possibilities that:

• AI isn’t useful (call it the “hater scenario”)—if anything, lower costs seem to increase demand; and/or that

• AI is so useful that nothing else is (i.e. the Citrini scenario).

So, why bring this up?

Well, in the first instance, recently there has been an emerging concern that token costs are, in fact, beginning to weigh on demand. Does it have legs? Let’s see.

ICYMI

Tokenmaxxing to token-rationing

Fairly recently, there was some anecdata of high profile companies (e.g. Uber, Meta, etc.) reversing their “token maxxing,” mantras, in favor of some more disciplined cost management.

Those companies do not appear to be alone:

In general, “token costs,” are becoming a point of interest in calls with analysts, management and experts, per AlphaSense.

Likewise, other soft data suggests that IT spending intentions, while still high, are cooling somewhat:

The majority of respondents still expect IT budgets to grow, but the level is coming closer to 60% (off the near record-highs from 2 years ago). Plus, estimates for the magnitude of growth have down-shifted somewhat from last year.

Token spend goes south (or does it)?

Perhaps more importantly, there’s some actual hard data to support the softer anecdata that token spending may be cooling:

The Silicon Data token spend index, which purports to measure total token spend, shows some fairly substantial deceleration-to-decline in June.

Of course, Silicon Data is just one data source, and the index has declined before (only to shoot back up), so it’s way too early to declare “game over.”

Other data shows a slight wow dip, followed by a full recovery:

Read more

• sticky car prices

• public markets: anything but bubbly

• for pe, the song remains the same . . . selection effects drive the rest

• vc is doing fine where it matters (sort of)

• private credit narrative v. fundies

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

1. Car prices are remarkably sticky

Random Walk has made this observation before, but when it comes to asset prices, there is perhaps no pithier example of how looks can deceive.

If you can believe it, the price of a new car rocketed up during the zero-interest rate period, and then just hit a plateau:

The average price of a new car has been remarkably stable at just under $50K.

Pretty wild how low interest rates can make the price go up, but high interest rates do not make the price go down, right? And all of that while disposable income ebbs and flows, and gas prices go up and down, and auto sales themselves have been pretty soft. It’s almost like the price is impervious to supply, demand, and most importantly, the customer’s ability to pay.

Well, it’s less wild when you appreciate that that’s not what happened, at all. Of course auto prices went down.

Cars are heavily financed assets, like houses and private equity, and when rates went up, prices most definitely tumbled . . . even if the auto dealers don’t want you to know:

The headline price stayed flat because rising discounts soaked up all the depreciation.

Autodealers surely know the psychology of auto-buyers better than Random Walk, so I’m sure they have very good reasons for staying anchored on a high watermark of headline prices. Maybe it has something to do with dealer financing lines, as well, I don’t really know.

Mostly, it’s just a brief intro into today’s theme: how are the asset prices, really? Is the price signal telling you what you need to know, or does something else lie beneath the surface?

ICYMI

2. Public markets are anything but bubbly

Say what you will about the state of the stock market, but there’s very little evidence of a “bubble,” at least not in the traditional sense.

Earnings led

If anything, the stock market has become substantially more cautious—multiples have contracted, and performance has been entirely profit-driven:

In comparison to the Dotcom boom/bust, when multiple-expansion accounted for 73% of the returns, it’s EPS doing all the work (especially for tech). And again, the prices paid for those earnings have declined pretty substantially, which is just a very weird thing for an earnings acceleration.

It’s somewhat less weird—and further evidence of the market’s “rationality,”—when you consider the narrowness of earnings growth. While it’s true that margin-expansion is expected to be pretty broad-based, the bulk of the earnings growth is concentrated in the AI buildout.

~60% of the S&P’s Q1 eps growth (15pp of 25pp total) is attributable to AI Capex mostly (plus some hyperscaler cloud acceleration).

In other words, profit growth is largely (but not entirely) a function of the largest companies ever spending north of a trillion dollars on traditionally cyclical hardware and services. But that’s still earnings growth—real profits, not just hoped-for profits.

Capexapalooza, but also caution

It’s actually a fun little exercise to think how big Capex could get.

Read more

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

Oil Shock or Nah?

Even more bullishness abounds, as on-again, off-again “progress” vis-a-vis the Strait hijackers switches back to “on.”1

It is pretty remarkable how bullish everything has been despite the clear deterioration in supply of a (supposedly) non-substitutable global commodity:

Supply buffers are what they are, but it does seem reasonable to suppose that there’s an outer limit to how much of a supply-side disruption the world can actually withstand without some real pain.

It’s bothered investors somewhat for sure, but not nearly as much as perhaps it should.

I mean, oil shocks definitely don’t matter as much as they used to . . . but they still matter:

In a less “oil intensive” economy, oil shocks just don’t hit as hard—there’s some inflation, yes, but employment effects are . . . positive?

Still, the world does need oil, and if supplies are shrinking every day, then do we run out? That seems bad, right? Unlike dollars, “you can’t borrow oil out of the ground,” as the saying goes.

On the other hand—and the reason for all the hedgy language—is that it’s also a massive mistake to bet too hard against the ingenuity and resourcefulness of the invisible hand, especially as it pertains to a global commodity. The picture looks insurmountably grim, but just because you and I can’t see a path, doesn’t mean there is no path(s). Random Walk has learned this the hard way before.

For one thing, less energy over there, means more energy here (which then gets sent over there):

2025 was (another) record-breaking year in energy production . . . 2026 will surely reflect yet another upward inflection. The US continues to grow its role as a net-exporter of energy.

What’s that? Can’t ship oil through Hormuz?

No problem—build some more pipelines, and in the meantime, load up the jumbo jets:

Seaborne net exports of oil, etc. are ~50% higher than they were last year.

The point being that when the going gets tough, the tough do, in fact, get going. And that’s all the more true when there’s lots of money to be made, which there certainly is.

Still, impressive adaptability notwithstanding, it’s hard to escape the lurking fear that (a per above) there must be some limits to what can be done and in how much time:

A 14% supply-loss (and counting) is an all-timer for oil disruptions, and eventually storage has to hit the bottom of the literal barrel . . . right?

It’s all a bit strange. For all the apparent data pointing to an incoming shortage, oil prices (forget stock prices) seem pretty complacent. I mean, they’ve surely gotten higher, but not that high in the big scheme of things:

WTI is ~$95/barrel . . . basically the same price as it was for the first-half of the last decade, until the shale boom kicked-in.

Now, there’s of course elasticity in demand, which means at a certain price, people just stop buying oil, and so maybe that’s keeping prices in check. But I do wonder if the physical reality is just a bit more optimistic than what the data seems to show.

Anyways, puzzlement aside, oil prices weren’t really the point. The point was just a brief look at Iran’s biggest customer: China. More specifically, China’s Global Emporium to Sell Everything to Everyone (aka the “Beggar Thy Neighbor Trade”).

ICYMI

China and the ‘Beggar Thy Neighbor’ Trade (reprise)

One quick thing on oil, but specifically as it pertains to China.

Read more

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

Why Isn’t the On-Ramp Working?

One of the things that makes me uneasy about the present state affairs is the “broken onramp” to the labor market.

Mostly, things seem just fine—not great, but fine—but some things are a bit off.

If the working age population isn’t growing much (and it’s not), and retirements are beginning to accelerate again (and they are), then why are young folks still having a relatively difficult time getting hired? If workers are riding off into the sunset, and there’s a relatively scarce pool of replacements, then you’d expect new entrants to be in hot demand . . . but they’re not.1

Unemployment for the 20-24yos got steadily worse, and then better, and then slightly worse again.

Perhaps, of course, that’s just noise, and youthful unemployment will revert to its recent positive trend . . . but even so, there’s still some trouble lurking beneath the surface of that recovery.

Take tech, for example. Employment rates did recover for the young ex-tech peeps, but the recovery wasn’t in tech:

Young folks became employed again, but their employment must have come from some other part of the economy . . . where, I’m not entirely sure, yet.

Why is the “broken on-ramp,” so troubling?

Well, partly because it speaks to the narrowness of labor-demand, but more broadly because it’s a bad sign for the best-case-scenario of our aging-shrinking-workforce. Mostly, though, things are just different now, and when it’s different, it’s hard to read the tea leaves.

ICYMI

The data that explains the most about most things

To recap, behold, a selection of the scariest charts in the world (that also explain more about the present than probably any other charts).

For Developed Markets, the demographic picture is grim:

The working-age ratio (i.e. the ratio of 65+ yos to the rest of the workforce, aka the “dependency ratio”) has been trending the wrong way since the GFC, and not coincidentally, working age population growth is approaching zero-soon-to-negative.

Man, the GFC really was a killer.

Or rather, whatever happened ~15 years before the GFC that led working-age-pop-growth to fall off a cliff was a real killer . . . big surprise we ended up with a little ‘ole housing surplus, right?

But more to the point, at no point this century is the working-age population expected to grow in developed markets. Let that sink in. It’s a remarkable shift from the century prior, and truly, uncharted waters.

And while it’s true that people are working longer, retirements are not slowing down:

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

Chart Dive on Private Credit, PE/VC

A two-part chart-dive on all things private credit, and PE/VC.

Before we begin in earnest, just one point that often gets lost in the discourse: what’s happening in these asset classes is not entirely unique (even if there are plenty of esoteric think-boy takes on the specific features of their own special strategies and cultures).

Reality is that it’s tough out there for pretty much everyone (with some exceptions, of course):

Fundraising has been tough-sledding across the board.

The exit-backlog is ubiquitous. Repricing, especially off of pandemania peaks, is ubiquitous. Concentration is ubiquitous. Zombie-funds (and zombie-assets) are ubiquitous. The new cycle is not like the old cycle, and that’s ok.

What worked before, just doesn’t work now:

Just using PE as an example, revenue (and earnings) growth is now ~2x the value driver that it was. Multiple expansion, by contrast, plays far less of a role—and this chart ends in 2025—in 2026, (contracting) multiples are almost certainly a headwind to enterprise value.

It’s not bad or good, but it is definitely different, and as per always, the only way forward is through!

ICYMI

Credit Check in the Discount BDC Aisle

Rightly or wrongly, private credit (and especially software-backed loans), continue to get hammered by investors:

~70%+ of Publicly traded BDCs are trading at a discount.

It’s mostly a software story, as software loans comprise the lion’s share of “distressed” loans, i.e. those trading at 70 cents or less on the dollar. It’s also mostly a ‘21-’22 era software loans story, as those were the ones made at peak pandemania, when revenue growth was ripping and credit was close to free.

Credit performance looks . . . fine?

The strange thing, however, is that credit performance is generally fine (if not improving). There also appears to be very little discernment within software loans (except perhaps on the degree of leverage):

Read more

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

“ICE has not improved labor markets(?)”

Tyler Cowen, the GOAT of generalist infovore internet-writing, surfaced a paper on the “the labor impacts of ICE,” with (his own) provocative header “ICE has not improved US labor markets.”

The gist of the paper is that deportations negatively impacted US workers, who had smaller employment and wage gains, relative to US workers in low-deportations areas.

From an econ 101 standpoint, it’s an expected result. Immigration is generally considered a positive-sum game: new arrivals get to work, which boosts productivity, and more productivity is good for native- and foreign-born alike.

There have been exceptions, of course. During the great labor-market catch-up, the trad 101 model was flipped on its head (at least in the near-term): pandemic retirees left a gaping hole in the labor market, which caused wages and prices to soar, and that hole was largely filled by migrants, which brought wages and inflation to heel. As Random Walk, and later Goldman Sachs, offered at the time: the standard 101 model of immigration’s impact on the labor force may not apply during periods of an acute supply shock. In that case, a migrant surge was deflationary.

True enough, but the supply shock is basically gone.1

That being the case, then it’s not unreasonable to think that the 101 expectation would hold, i.e. that deportations would hurt, rather than benefit, US workers. And a W for Econ 101 is what the paper purports to show

Obviously, from a MAGA politics standpoint, the result is disappointing: booting migrants is supposed to benefit American workers, not hurt them.

So does that mean that MAGA lost the day?

Well, maybe, but you wouldn’t know it from the research cited by the good Professor Cowen—at least, not when you dig under the hood (and I say that even though I’m inclined towards the 101 expectation).2

Non-college male construction workers effected

The primary finding of the research is that deportations had negative effects on native born employment and wages. But not everywhere and not for everyone.

It’s non-college male/construction that’s doing all the paper’s work:

Negative effects for predominantly male construction, whereas, effects on women and/or other sectors were null and/or positive.

The researchers (fairly) lean on the negative impacts to non-college male construction work to make their point because it’s a segment of the labor market that has a lot of illegal immigrants.

True enough.

ICYMI

Here’s the other thing, though, about construction, especially residential construction: there’s been a pretty substantial industry slowdown, having very little to do with immigration.

That’s a big-time confounding effect.3

In other words, the crown-jewel discovery of the research—that employment conditions for male construction workers declined—is pretty easily explained by cyclical demand (and not deportations).4 Somewhat surprisingly, that fairly obvious and substantial confounding factor does not appear to have played any role in the study (other than the general comparison to the “control,” i.e. zip-codes with fewer deportations).5

So, it’s not exactly 1 point for MAGA, but it’s not exactly 1 point for Econ 101, either.

But ICE is good for healthcare?

That cyclical demand (as opposed to deportations) is the prime mover is consistent with the paper’s other finding: null and/or positive effects for women and their sector of choice, i.e. healthcare.

Healthcare employment (and wages for men) improved following the surge in ICE arrests.

As Random Walk has offered countless times, there is no stopping the healthcare demand train, and the Rotation to the National Nursing Home, continues apace. As a result, it’s a been very much a “women’s world” out there, and has been for some time:

Read more

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

Subscribe now

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

1. A New Bull Rises

Once again with feeling, expectations for earnings growth are very bullish:

The upward revision to EPS estimates is an incredible thing to behold.

If we were coming out of a recession or something then maybe you could see expectations shift like that, but earnings were already pretty strong to begin with. It’s wild.

As for stock market prices themselves, well, it continues to be a tale of diverging fortunes, but semis (and memory) are having quite a run:

In terms of historic bullruns (unadjusted), semis are approaching Nikkei (in the 80s), Dotcom Nasdaq and post-WTO BRICs, in terms of scale.

As Random Walk has said many times in the past, eventually some of that spending on chips, etc. will have to flow through to the users of those chips, otherwise they will not keep spending quite so much money on chips.

But, for now, AI adoption is creating so much demand for chips, memory, and now CPUs, that the biggest spenders cannot get enough, and everyone knows it. And that’s how you get unprecedented upward revisions to earnings.

ICYMI

2. Bull market, but not for everyone (especially non-AI Tech)

Of course, if you’re not somehow in the AI infra supply-chain, you’re not doing bad—you might even be doing pretty well—but the market mostly doesn’t care.

Every sector, but Industrials and Staples, has underperformed the price action implied by its results.

Public software hammered

And no one has been having a harder time impressing the market than tech. For SaaSCos, especially, the bloodbath just continues apace.

• Hubspot beat earnings and raised guidance:

  

• And as a reward it got absolutely hammered:

  

Hubspot wasn’t alone.

Plenty of software companies met, exceeded and even raised guidance, but still got beat-up:

Read more