Chris Auld

🥝 in 🇦🇺 & 🇯🇵 & 🇳🇿. Bike. Ski. Paddle. Law. DataSci. Quantish. Half Stack Dev. Researches before Reckons. Bayesian a.f. Stormy Normie. Per/Capita

[Chris Auld @cauld]

Sugoi desu!

[Jack468 @jack840A]

Here is the "Prison Mike" scene from The Office Japanese Dub.

[Chris Auld @cauld]

@bustedblonde @theplatform_nz Company colours?

[Chris Auld @cauld]

Putting the Out in Outrageous.

[Vincristine @vincristine]

Suggesting Hipkins should fire Kerrin Leoni for speaking at an event gang members were at is outrageous. Journalists suggesting this should happen are creating news not reporting it @rnz_news

[Chris Auld @cauld]

@EricCrampton I am vehemently opposed to things like this all up, but also pretty comfortable that if people were ok with a national digital ID like Singapore’s SingPass then zero-knowledge proofs for third party sites would be trivial to provide.

[Chris Auld @cauld]

Flailography (n.): the frantic, highly stylized hand gesticulation Gen Z and millennials perform in TikTok selfies—wild hand-flapping, emphatic jabbing, pouting through pointed fingers—as if every sentence requires its own interpretive dance.

[Chris Auld @cauld]

What happens to the odds on pdoom() if the AIs have to fund themselves at 1100bps over Treasuries like $CRWV does…

[Jesús Fernández-Villaverde @JesusFerna7026]

Everyone talks about what AI will do. But almost nobody asks who will pay for the machines. Taking advantage of AI will require an extraordinary amount of capital, from data centers to new power plants. Will that capital be around?

I try to address that question in the first draft of my new paper “Saving for the Machines.”

To show how important capital is to fuel the transition to an AI-based economy, I document how the same technology path that generates 11.7% annual growth when households choose their saving rate optimally delivers only 2.2% when the saving rate remains fixed. In other words, if we experience slow growth over the next few years despite fast improvements in AI, this slow growth may reflect too little capital accumulation instead of a technological bottleneck.

The key to this result is that AI raises the capital share by moving tasks from labor to machines, which requires that those machines are built. In comparison, ordinary technical progress makes labor and existing machines better at the tasks they already perform.

To inspect the mechanism, I start from an accounting identity: the saving rate is the capital share times the ratio of investment to capital income. If I take the evolution of the capital share as given, I can ask how much the latter ratio moves in response to that evolution. Crucially, with an optimizing household, this ratio barely moves. As machines take over tasks and the return to capital rises, the household responds by letting consumption grow faster, by the intertemporal elasticity of substitution times the rise in the return. To pay for that faster consumption, capital must also grow faster by the same amount. So, the household keeps investing about the same fraction of what the machines earn. And since machines earn a larger share of output, the saving rate also goes up. With a fixed saving rate, the same accounting identity tells me the opposite: the ratio falls in proportion as the capital share rises.

Will the U.S. and other advanced economies then generate enough capital? For many reasons, I doubt it. We might end up in a situation where many of the opportunities offered by AI are left on the table. This is particularly important for all who think that AI might get us out of many problems, such as fiscal troubles (like in France) or ultra-low fertility. We need to save for the machines, and it is not clear we can do that.

Anyway, this paper has been long in the making (many months of hard work), but I have learned much from it. In particular, working through the model changed my mind about many issues related to AI. Again and again, my intuition said X, but the math said Y. And math always won in the end. That is why we write formal models.

A nice example is Proposition 5, whose proof is the one I like the most in the paper. With improving machines and strong task complementarity, there is no fixed balanced growth path because the target itself keeps moving. The nice part of the proof is to realize that, after the right normalization, the moving target becomes a fixed point, and the saddle gets steeper as machines improve. So instead of falling behind, the economy tracks the moving BGP more and more closely. A fair amount of ODE work to get there, but a very simple economic idea underneath it once you get it.

The paper is here:

https://t.co/zthH4T6nS3

and the slides are here:

https://t.co/VAsvWd9EQn

[Chris Auld @cauld]

Today I found this. It is good because who doesn’t love a good ShitCo short…

[Dirty Texas Hedge @HedgeDirty]

A GENTLEMAN’S PRIMER ON SHORTING INSOLVENT SHITCOS

So you’ve found a ShitCo. It’s such an ugly zombie it couldn’t be an extra in season 19 of The Walking Dead
But it’s still got a $185m cap
And you want to be the one to mung the last rancid, putrefying drops out of that corpse

[Chris Auld @cauld]

@dadsweb67 🍋 💪

[Chris Auld @cauld]

@LewSOS So I “re-enrolled” today. I did it all on my phone… but both my Passport and my Drivers License didn’t work for validation so it ended up being manual.

It wasn’t hard. But still took maybe 10-15mins…

[Chris Auld @cauld]

@kaiviti_cam Mummy��s ideas are the best ideas…

[Arvel Demon @avelrmd]

Your name is Jean-Marc, you’re 46, and you run a video game and collectibles shop in Nice, wedged between an optician and a real estate agency selling 1-bed flats for €410,000 “with sea view” if you lean far enough off the balcony.

You know every GTA San Andreas cheat by heart. You tested all of them in 2005 on the PS2 in your bedroom at your mum’s place.

You opened in 2011 with a state-backed loan at 4.7%, and the naive belief that passion eventually pays.

It did pay, sort of. €1,900 a month in a good year, after the French tax office has eaten first, second and third. Your commercial rent is indexed on a formula nobody in the Republic actually understands.

You sell Funko Pops at €17.99, sealed Pokémon cards, refurbished controllers. You let little Enzo, 12, pay in installments because he’s saving for a Switch 2. You close at 7:30pm and walk home because the harbor parking is €3.40 an hour.

Saturday, there’s a protest “against the cost of living.”
You also think living costs too much. You almost went.
At 5:12pm, your metal shutter gets ripped open with a crowbar.

Forty teenagers in tracksuits and crossbody bags walk into your shop like it’s a bonus level. They take the PS5s, the Dragon Ball figures, the till with today’s €280 and even the box of paper bags that say “thank you for your visit.”
A guy in a floral shirt films it vertically for TikTok.

You step outside and say “stop, this is my shop.”
You get hit from behind.
Then rocks.
They laugh. Floral shirt guy is thrilled: good angle, good lighting.
Sunday morning, the police tell you there are “1,400 arrests nationwide” and they won’t send a squad for some Funko Pops. Filing the complaint takes 2 hours 40. The officer types with two fingers on Windows 7.

Monday noon, your insurer reminds you that your €1,140-a-year policy has a €3,000 deductible and a “riots and civil commotion” exclusion on page 31, font size 6.
Saturday night, an MP explains on TV that “the anger is legitimate” and we “shouldn’t cry over shop windows.” He’s never owned a shop window. He has a state-funded apartment.
Sunday, your limited edition Vegeta figures show up on Vinted. “Brand new, never opened, €39, pickup in North Nice.”
The seller has 4.9 stars.
Monday, the tax office emails you. They haven’t heard about the riots. They’ve heard about your payment deadline.

You did everything right, Jean-Marc. That was never the problem.
The problem is you were playing by the rules in a lobby where nobody else does.

So stop believing in the State. It’s not coming to save you. It’s coming to collect.

Stop voting. Every five years they hand you the controller, and it’s never plugged in.

Go to the gym. Not for revenge. For you. Because when the system folds, the only asset nobody can loot is the one you built under a barbell.
And make money they can’t see. Cash, side hustles, sats on a cold wallet in a sock drawer. Not your keys, not your coins. Not your shop, not your loss.

Fuck the system, Jean-Marc.
It already did it to you first.

[Chris Auld @cauld]

@aniobrien … unless you’re a moron…

[Chris Auld @cauld]

@TimHaines 12. Tops.

Microsoft has already removed the punch bowl*…

* Disclaimer: I do still like my employee all-you-can-eat but I suspect that’s not long for this world either…

[Chris Auld @cauld]

@judah47 @Nature Ok. Side quest will be do built a decent downloaded library with smart retries etc. because it’s both slow *and* flaky…

[Chris Auld @cauld]

@judah47 @Nature Yup. First cut is all built and ready for validation… not sure if it is just me but the ECMWF data service is slowest bulk data service I have touched in a long time… a few hours til I can do initial validation… 4 days until I have a full 10yr of data…

[Chris Auld @cauld]

@judah47 @Nature Ok. Running up mechanics to apply to the JMA monthly (CPS4) forecast … let’s see how we do on a backtest…

[Chris Auld @cauld]

@judah47 @Nature Cool. Ambitiously pulling paper and code right now and working out the practicality of applying to JMA MSM this winter ❄️ ⛷️ 🤓

[Chris Auld @cauld]

@FundamentEdge I have Hindsight as general purpose knowledge store and then Nocodb as a structured database that agents can use to create tables of shared data. Still refining, but yes, this approach is very useful. Use w/Claude + GitHub CoPilot.

[Chris Auld @cauld]

Sound like it will float my boat…

[AJ Asver @_aj]

@typesafeai And on github: https://t.co/vKfKe2shoe