"I'm sorry...he said what now?"
06.03.2026 04:03
👍 3
🔁 0
💬 0
📌 0
"I'm sorry...he said what now?"
Pls bring back Background Costa
I would watch an alternate-history version of The Hunt for Red October with Klaus Maria Brandauer as Marko Ramius
Quite the chart from Stripe's annual letter assets.stripeassets.com/fzn2n1nzq965...
China’s solar power generation overtook wind for the first time last year as a boom in cheap panels continues to reshape the country’s grid and climate trajectory
Analysts now expect AI hyperscaler capex to total $667 billion in 2026.
None of the things that we were talking about in 2006 eventuated in 2008; none of the things we were talking about in 2014 eventuated in 2016. Decent precedent for the vibe you're looking for.
I like the comps in the paper; a challenging element, to me, is the fact that the capex series stops at 3Q 2025 (fair enough) but the 2026 forward guidance is for ~$700B in total capex from the five companies mentioned (to say nothing of the REITs, neoclouds, foundation labs).
Warm banana, a bold strategy for keeping warm in snowy wind, let's see if it pays off
I now want a compilation of every script mention of smog in any movie or show set in LA
Indeed
And now the air in Beijing is nice too!
Musing on two great 90s LA movies set up in the hills and how smog is a story feature:
Get Shorty: “they say the smog’s the reason we have such beautiful sunsets”
The Limey: “you could see the sea out there, if you could see it”
LA smog: not what it used to be!
Happy lunar year, everyone. Obligatory linguistic IYKYK meme
Groundhog Day is Pareto Optimal. Cannot add more anguish without diminishing the comedy, and vice versa
Five tech companies are on track to spend $700B on AI infrastructure in 2026 — a one-year commitment bigger (relative to U.S. GDP) than anything except the Louisiana Purchase.
In this week’s blog, @nathanielbullard.com steps back to ask what endures after the boom.
Read more 👇
hubs.la/Q042JF640
Final EIA-860 is out, and 2025 really did it: 54 GW of new U.S. electric capacity, 96% clean. Solar again carries the offense, and we built more storage in 2025 than the cumulative total through 2023. More and more and more additions... but also record low retirements.
I like this one so much I went in to triple-check to hyperscaler capex. It checks out! But something often forgotten in here: North American oil & gas exploration & production was 1.6% of GDP in 2013!
But with T+1 settlement, it may not be enough
With the added parallax of watching it from Singapore too (not a lineman to be seen...everything is underground)
Had to screencap this @shastingssimon.bsky.social @raviam.bsky.social @laughlax.bsky.social @costasamaras.com @shannonosaka.bsky.social Super Bowl moment
Singapore new car sales by country of origin. Japan now sells about 1/4 the number of new cars that it did in 2016; China sells almost the same number as Japan (up from 5 in 2015).
Singapore new car sales by country of origin. Japan now sells about 1/4 the number of new cars that it did in 2016; China sells almost the same number as Japan (up from 5 in 2015).
Asia just puked Silver. Yikes
That is the right spirit
It isn't worth it in the first instance. Having data-templated sheet-chart + populated-templated-slide/no-chart is 95% of the work, and the chart would always need fiddling anyways (if you're me). So, I stuck with what works, for now, and it works great. Very 'jagged technological frontier' stuff
I've got separate .md for data extraction, for chart templating, and slides. But (familiar Moravec's Paradox stuff here) the step that doesn't work well and which I binned for now, is porting a formatted / automated chart into a formatted / populated slide template. 🤷
