Gen X and Millenials don't share Boomers' obsession with green lawns, so it's a race against time, whether Boomers or lightning bugs will go extinct first

On the other hand, I don't think I know any millenials that don't have an extremely overbearing HoA that forbids anything other than a grass lawn.

I looked it up and a couple of states have laws against HOAs from forcing your to have a grass lawn. Alternatives can include native plants, drought tolerant plants, xeriscaping, vegetable gardens depending on state. The states I've found are California, Colorado, Florida, Texas and Maryland, Nevada.

Maybe not directly, but they definitely care about property value which gets you municipal codes requiring you to mow your lawn or get fined.

Don't inference servers like vllm or sglang just translate these things to openai-compat API shapes?

They do, but that's kind of the article's point - someone still has to write and maintain the per-model chat template and tool call parsing inside vllm/sglang. Every time a new model ships with a slightly different format, the inference server needs an update. The M×N problem doesn't disappear, it just gets pushed one layer down.

I would like to see them include some analysis of the first recorded audio conversations such as on wax cylinders or the 1930s WPA/Federal Writers’ Project. Was sentence complexity the same as it is now?

The focus on clarity of thought is a modern shift, or at least the "modern preference". Hemingway was a big proponent on short, direct sentences and few or no adjectives.

In the 1800s and early 1900s, complex sentence structure signaled intelligence for both the author and reader. It was a form of entertainment, in a way, when books were few and nights were long. Try reading Henry James for an idea about what this looked like in practice. Shakespeare is another obvious example of "heightened language" besides archaic words the play are written in iambic pentameter and the spoken text is far from natural (yet incredibly precise).

> the play are written in iambic pentameter and the spoken text is far from natural (yet incredibly precise)

(Iambic pentameters are 10-syllable lines with alternate syllables unstressed and stressed, like "if MUSic BE the FOOD of LOVE ...", the so-called heartbeat rhythm)

Shakespeare actually used a variety of different styles to demarcate different characters, moods, etc. As a very rough rule-of-thumb in Shakespeare, posh characters speak in iambic pentameters, commoners and clowns speak in prose. In A Midsummer Night's Dream, for example, the Athenians speak in iambic pentameters and the clowns speak in prose. When the clowns put on a play for the Athenians, the clowns and the Athenians swap speaking styles, so the Athenians make snarky comments in prose (just like a badly behaved audience) on the badly rhymed acting of the clowns. The fairies, meanwhile, speak in trochaic verse, so their king and queen sound stylistically different from their Athenian equivalents, almost like Shakespeare has given them a foreign accent. When two characters are arguing, the ten syllables of a normal line are sometimes split between them to emphasize the back and forth nature. If a character is flustered or annoyed, their lines may be obviously different from the 10 syllable norm, again to emphasize their mood.

For actors learning their lines, the syllable counts almost act as stage direction hints: if they aren't 10 syllables, then some mood or other needs to be taken into account.

Because chromium III oxide is a very light-fast pigment

increasing the (social) pressure on maintainers to get PRs merged seems like the last thing you should be doing in light of preventing malicious code ending up in dependencies like this

i'd much rather see a million open PRs than a single malicious PR sneak through due to lack of thorough review.

Not really the time for that. There's also PRs being merged every hour of the day.

A lot of culture was lost in the Cultural Revolution

I proxy all of my llm completion subscriptions. In a typical 7d span-

model completions read write cached_read cache_write

claude-opus-4-6 11000 16900000 5840000 1312000000 66120000

17M uncached reads (input) and 6M of uncached writes (output) is

  $5x17+$25x6=$235 for Opus 4.6

  $2x17+$12x6=$106 for Gemini 3 Pro

  $0.60x17+$3.6x6=$31.80 for Qwen3.5 397B-A17B via Huggingface API

You did not add up cache writes, which are $6.25 / MTok, which is another ~$400

It's safe to say that 100% of the Northmen who invaded England in 1066 shared that same "job description", however.

By 1066, not quite. That was an invading army led by the King of Norway to press his claim on the throne of England. I’m sure many of the soldiers in that army had been Vikings but at that time they were soldiers of a Christian king, which would have been considered much more legitimate than being a heathen raider.

I guess the Normans were also of Nordic descent but they had given up the Viking way of life a century before.

> Nordic descent but they had given up the Viking way

So they were indeed Vikings as a matter of heredity?

“Viking” isn’t a matter of heredity though. If your grandfather was a bricklayer, you’re not a bricklayer as a matter of heredity, you have to actually lay bricks. Likewise, if your grandfather was a viking, you have to actually go raiding and pillaging to be one yourself. Which is not something you’re likely to do if the king of France gave your grandfather an entire duchy in exchange for a promise to stop doing that sort of thing.

99%.

Poor Sven just liked maps but they kept making him come along on their “excursions”.

What's gonna bake your noodle is, Viking raids were the VC-funded startups of medieval northern Europe. Norse kings were very generous with their kingdom's treasure, to the raiders with the most fearsome reputations.

sqlite_vec is already the sqlite for AI memory

Any chance you went beyond the surface comparison and have thoughts on how the libraries compare in functionality?

I scrolled through the entire readme and didn't see any mention of sqlite_vec. My feedback for the readme would be to optimize for signal- if it is a layer on top of sqlite_vec say what it does on top of that etc

It uses GRDB which is an SQLite wrapper https://github.com/christopherkarani/Wax/blob/786ad6cc541392...

Nothing is very clear here.. the benchmarks might just be comparing WAL mode on vs off, or something else entirely, SQLite does not have 150ms latency on such a small database.

The original commenter wasn't making statements about sqlite being involved, they were saying that a specific library should be mentioned if it was involved, which it wasn't. Unless you are saying sqlite_vec is part of the dependency chain through GRDB?

It would be like commenting "If any other developers were involved in this project you should mention them."

It is Not a layer over sqlite_vec

sqlite-vec is a great vector index — Wax actually uses SQLite under the hood too.

The difference is the layer. sqlite-vec gives you vec_distance_cosine() in SQL. Wax gives you: hand it a .mov file, get back token-budgeted, LLM-ready context from keyframes and transcripts, with EXIF-accurate timestamps and hybrid BM25+vector search via RRF fusion — all on-device.

It's the difference between a B-tree and an ORM. You'd still need to write the entire ingestion pipeline, media parsing, frame hierarchy, token counting, and context assembly on top of sqlite-vec. That's what Wax is.

Thanks for clarifying. If mv2s is a sqlite3 db file under the hood that is something I would like to see in the readme as it would make me more likely to use.

Article doesn't say the "trick" - it was a technique now known as inlay grafting.