senderista

And China scaled up solar production to the point that it's now truly practical.

Amazon is not "meaningfully in inference"? Bedrock seems to have a ton of enterprise customers, some of which would never trust the AI labs themselves with their data but will trust Amazon.

Now that is some spicy autocomplete.

Why are you implicating ghostty? Have you compared its CPU usage to any other terminal?

call me back when it's on Poob

But SO had already been acquired by PE when LLMs went mainstream, so I doubt they would have made any strategic growth bets rather than trying to squeeze all the monetization they could out of the site.

I tried to answer an SO question exactly once, around 15 years ago. Someone asked a question which was covered perfectly by some OSS code I had already published, so I linked to that in my answer (while explaining how the linked code was relevant). My answer was promptly deleted, of course, for linking to code. After that I lost all desire to contribute.

I think a big part is the decline in organized religion.

“Classically trained” in the musical context usually means “I took a year of piano lessons.”

Could you elaborate on "GCs are not used there [high-performance throughput-optimized systems]"? Are you referring to the cascading effects of tail latency on systems with high fanout?

Not to mention that in general GC simply requires more memory.

agreed, the de facto genocide bit is slightly more concerning to me

Telegram channels are about the only way to get the reality on the ground on both sides.

No, it's not the same as the German tank problem, because that assumes all sampled values are unique (and totally ordered). It's closer to the "species richness" estimation problem, where you estimate the total number of species from a sample (specimens) with many repeated values (species).

There are many different statistical estimators for this problem, many from the ecology literature (e.g. Chao1). This paper and a Python library based on it covers many of them:

https://www.vldb.org/conf/1995/P311.PDF

https://pydistinct.readthedocs.io/en/latest/docs/readme_link...

Just to save anyone the trouble of a web search:

West Frisian: Bûter, brea, en griene tsiis is goed Ingelsk en goed Frysk.

Modern English: Butter, bread, and green cheese is good English and good Frisian.

Really just a tribute to the syncretic capacity of Hinduism.

A famous example (similar to the ones you can find between West Frisian and modern English):

Vedic Sanskrit: Devas adadat datas; Devas dat dhanas.

Lithuanian: Dievas davė dantis; Dievas duos duonos.

English Translation: God gave teeth; God will give bread.

I worked on Java code at AWS for a few years and nobody tried to optimize allocations. Then I changed jobs and started working on a Java MPP database and my first code review was brutal. You were expected to avoid allocations as much as possible (mostly by using the SoA pattern everywhere). At that scale no GC could save you from excessive allocations.

Sadly Amazon has been known to enforce its 18-month NC in WA state (not sure if they’ve ever prevailed in court). They will absolutely not negotiate on this.

The Kalash are very similar genetically and culturally and are also facing cultural genocide by Islamists. Truly one of the most fascinating ethnic minorities extant.

Black Americans have on average 20-25% European ancestry but we don’t call them “Europeans”.

We call Arabic speakers “Arabs” but there wasn’t that much genetic replacement.

Try reading about Kafiristan, where primitive Hinduism (think Rig Veda) survived for 3500 years before being wiped out in the 1890s by a British-appointed Islamic despot.

There are many more cool posts on that blog.

Will this be the last generation to remember the night sky?

At the start of the war there was plenty of meat being thrown into the trenches on the Ukraine side. Have you not seen any videos of men being press-ganged off the streets or in their homes? Or of the protesting wives of TDF units thrown into the trenches with virtually no training and only token weapons and ammo? (You can find plenty of both on Ukrainian Telegram channels.)

If Ukraine is treating their soldiers better now, it’s at least partly because they’re running out of them.

That last point (fortunately) does not apply to this conflict. In fact, the fraction of civilian deaths might be the lowest for any conflict with comparable casualties. I haven’t verified this, but would be interested if someone did.

Uh, we know who the ancient Armenians were, from both linguistic and genetic evidence. They were Indo-European, not Semitic (unless you actually mean the Urartians or something).

Archeology has certainly been affected by unscientific intellectual fashions: take the post-WW2 marginalization of mass migration hypotheses, which happened as a reaction to Nazism, not on the basis of any scientific evidence. It took the ancient DNA revolution (i.e. being embarrassed by outsiders who could prove the archeology consensus was wrong) to correct this purely political scientific bias. (Note that historical linguists generally continued believing in mass migration all along though.)