January 10, 2025 | Read Online

"The age of AI Agentics is here."

— Jensen Huang

Friday superintelligent charts

There's a saying in college football: "If you have two quarterbacks, you have no quarterbacks."

Like a hunter who chases two rabbits, a football coach who can't fully commit to one of his quarterbacks will end up with muddled results.

I find myself in the same predicament after receiving your highly opinioned but evenly split feedback on whether to stick with macro charts or pivot to crypto on Fridays — I guess there's a reason football coaches don't poll the fan base to decide who should play.

This is a nice problem to have — you guys seem to like charts as much as I do — but it leaves me undecided on which direction to go in.

For now, then, I think we'll go in both directions: Friday charts will be a game-time decision, depending on what's been happening that week.

It's an easy decision this week, as seemingly all markets have been fretting about bond yields and this morning's US labor data — stocks, bonds and even crypto are lower on fears that inflation and interest rates are headed higher.

But all of this week's fun, interesting and long-term important news has been technology-related (mostly because the world's largest tech conference, CES, was happening).

Sam Altman kicked things off with a Monday morning blog post announcing that OpenAI is confident it "know[s] how to build AGI" — so much so that he and his colleagues are now aiming even higher.

The promise of AGI has always been computers with human-like intelligence, but OpenAI is now working on "superintelligence" — agents that far surpass both humans and AGI, perhaps solving problems that we don't even know exist.

This remains speculative, however, and probably a long way off (a cynic might say Altman is invoking superintelligence as a sales pitch to raise money for a company that lost $5 billion on $3.7 billion of revenue last year).

More immediately, Jensen Huang announced this week that the world's next major technological shift — "agentic AI" — is already here, and that "the ChatGPT moment for general robotics is just around the corner."

If so, Jetsons-like robot servants could soon become ubiquitous: In a recent report on the coming age of "physical AI," analysts at Citigroup estimate that 1.3 billion AI-powered robots will be among us by the year 2035 — mostly vacuum cleaners and drones, I think. But by 2050, Citi expects 650 million of these will resemble humans.

We've heard this before, of course — how long have we been promised flying cars?

But we also learned this week that flying cars are finally happening!

Sort of — it's more accurate to call them mini-airplanes inside of a van, but this is still pretty amazing (especially for the low, low price of "below $300,000").

More affordably, we'll soon be getting $3,000 personal AI supercomputers and $3,500 "rollable" laptops with screens that roll up like a piece of paper.

As consumers and investors, I think we should be excited for all these things, but as people living in the age of AI agentics we should also prepare for a bumpy ride: Alex Tabarrok predicts the intelligence revolution will be "more wrenching" than the industrial one.

As an economist, Tabarrok is mostly thinking about the impact of AI on employment, but the technology itself could also be a bumpy ride— as Mark Zuckerberg noted this week, "the problem with complex systems is that they make mistakes."

As an example of one such mistake, here's a video of a self-driving car hitting a delivery robot.

Financial markets are also complex systems — and they too make mistakes.

So let's check the macro charts to see if any collisions are coming investors' way.

Good news is bad news again:

This unassuming chart was enough to strike fear into investors' hearts this morning. US non-farm payrolls were up a robust 227,000 in December, with revisions adding another 56,000 to previous months' data. The good jobs news is bad news for short-term interest rates — the economists at Bank of America no longer expect any further cuts from the Fed this cycle.

Feeling uncertain:

This morning's University of Michigan survey reported five-year inflation expectations at 3.3% — the highest level since June 2008. Uncertainty about inflation (above) rose to the highest level in decades. I find it kind of weird that either of these numbers would be higher than they were in 2021 when inflation hit 9% — but it's a survey of people, and people are weird. (The superintelligent robots that replace us will do much better.)

Everything is politics now:

This chart from Bloomberg's @M_McDonough illustrates just how weird we are: Our perception of inflation is almost entirely a function of our politics.

We've never been less likely to lose our jobs:

Whatever our politics, we should all feel good about staying employed: The number of employed workers losing their jobs in the US remains near historic lows.

Tip fatigue:

Data from Toast shows that the average gratuity at full-service restaurants has trended down to 19.3%. (The red circle at the bottom is my average, but don't tell anyone.)

Flipping the script:

You know we're living in an all-new world when interest rates in Japan are above interest rates in China.

Historically top heavy:

"First time since the last time" is usually a low form of analysis, but this is a good one from Jim Bianco: The top five stocks in the S&P 500 account for 29% of the index's market cap, the highest level since the legendary Nifty Fifty of the 1960s (which was followed by the all-time worst markets of the 1970s).

It wasn't just the Mag 7, though:

The other 493 components of the S&P 500 accounted for a respectable 11.7 percentage points of the S&P 500's 25% gain in 2024.

It's mostly about earnings:

Zoom out and it becomes clear that stocks mostly go up because company earnings mostly go up.

Something to keep in mind if things get rocky in the age of AI agentics.

Have a great weekend, superintelligent readers.

— Byron Gilliam

Crypto's Annoying MEV Problem

Dan Smith joins the show to discuss avoiding noise in crypto data, the issue with FDV and stablecoin supply across chains. Join them to unpack the concept of conditional liquidity and Solana's sandwich attack problem.

Listen to Lightspeed on Spotify, Apple Podcasts or YouTube.

James Seyffart @JSeyff

Just how big was the first year for Bitcoin ETFs?
MASSIVE

Here's a list of the largest ETFs 1 year after they launched. Even if you inflation adjust the assets, 4 of the Bitcoin ETFs are in the top 20 US ETF launches of all time. $IBIT, $FBTC, $ARKB, and $BITB