gojomo

The word "government" doesn't magically erase all the same individual & institutional incentives, ambitions, biases, & flaws that exist elsewhere.

And sometimes, the extant magical belief that "government" is different & immune lets those same human factors be ignored until they feed bigger, slower disasters that everyone is afraid to admit, because (ostensibly) "we all did this together".

Which email client will stylize raw markdown itself, making the HTML step here superfluous?

You may be bored of AI, but because AI is not yet bored of us, turning away may be dangerous.

If you're relying on a serialized 'source of truth', against which everyone must independently ensure their changes sanely apply in isolation, the. you've already resigned yourself to a single-threaded process that's slower than what improved merges aim to enable.

Sure, that works – like having one (rare, expensive) savant engineer apply & review everything in a linear canonical order. But that's not as competitive & scalable as flows more tolerant of many independent coders/agents.

Should you be counting on confusion of an underpowered text-merge to catch such problems?

It'll fire on merge issues that aren't code problems under a smarter merge, while also missing all the things that merge OK but introduce deeper issues.

Post-merge syntax checks are better for that purpose.

And imminently: agent-based sanity-checks of preserved intent – operating on a logically-whole result file, without merge-tool cruft. Perhaps at higher intensity when line-overlaps – or even more-meaningful hints of cross-purposes – are present.

Can translate between 1600 languages.

Can't achieve subject-verb agreement in 1st sentence of their English abstract.

Advances made through No Language Left Behind (NLLB) have demonstrated that high-quality machine translation (MT) scale to 200 languages.

If you're Robert Reich, you can! You can make up anything, and someone will submit it to HN to waste everyone's time!

Competitor MapBlast's 'LineDrive' directions were my favorite printable option – both the abstract overview and the turn-by-turn abstract clips.

https://somethingaboutmaps.wordpress.com/2011/03/08/remember...

Like partial courses of antibiotics, this will only relatively-advantage thoae leading efforts best able to ignore this 'poison', accelerating what you aim to prevent.

"It had been a slow Tuesday night. A few hundred new products had run their course on the markets. There had been a score of dramatic hits, three-minute and five-minute capsule dramas, and several of the six-minute long-play affairs. Night Street Nine—a solidly sordid offering—seemed to be in as the drama of the night unless there should be a late hit."

– 'SLOW TUESDAY NIGHT', a 2600 word sci-fi short story about life in an incredibly accelerated world, by R.A. Lafferty in 1965

https://www.baen.com/Chapters/9781618249203/9781618249203___...

I look up old iMessages, emails, group chat comments, and so forth constantly, often finding valuable gems of wit, reference-material, recommendations, or media that I dimly remember from years or even decades ago.

Signal and other messaging apps offer a 'search' bar across all sessions & history, so I doubt I'm the only one.

It's hard for me to imagine being so present-focused such a history wouldn't be personally useful.

Or, so worried about "someone [using] it against [me] in court" that I'd need more than the occasional auto-expiration, and specifically my messenger "protecting" me with intermittently-enforced loss-of-histories (on just theft/loss/hard-failure of primary device).

Anecdotes of occasional problems, even at a low or unquantified rate, are valid & useful evidence that something negative is happening.

Anecdotes that sometimes those problems don't occur are nearly worthless. Of course that's true - the original anecdotal complaint already implicitly relies on, & grants, the idea that there's some default, "hoped for" ideal from which their experience has fallen short.

To chime in, "never had your problems" thus adds no info. Yes, people lucky enough not to hit those Signal limits that cause others to lose data exist, of course. But how does that testimony help those with problems? Should their frustration be considered less important or credible, because of your luck?

The as-if portrayal is one way your anecdote will be perceived, even if that wasn't your intent.

"I've been lucky not to have devices lost/stolen/bricked. Why can't other people just be lucky like me?"

If the device with an 18GB message history was truly 'dead', how did you transfer the history over?

The Wikipedia page is useful, and as you've identified the 2025 MN Representative Hortman murder as the "first assassination of a sitting legislator at the state or federal level in my lifetime" – not counting the 2015 murder of SC Senator Pinckney – is it safe to assume you're a precociously-posting 10-year-old?

I was born in 1970; per your reference, there've been a bunch of state & federal legislators (or recently-former legislators) killed for political (or pseudo-political deranged) motives "in my lifetime" – and far more in the 1970s than in the last 10 years.

In my lifetime, one sitting President was shot at & missed (Ford in 1976), and one was shot at & hit by a ricochet (Reagan in 1981) – again, more in the past than the shots that grazed candidate Trump in 2024.

The Wikipedia-listed murders of other officeholders, like mayors or judges, are also more frequent in the past than recently – especially going before either of our lifetimes.

So trend impressions are very subject to frames of reference & familiarity with history.

I suspect if people in general had a deeper & broader sense of how common political violence has been, both in US history & worldwide, they'd be, on the one hand, less prone to panic over recent events & rhetoric (even though it is concerning), but also on the other hand more appreciative of the relative peace of recent decades (even with the last few years' events).

Plot these against the rate of similar attempts (or successes) over the past century if you want to convince others of anything other than your own subjective presentist perspective.

Lessons are repeated until learned. And again, after those lessons are forgotten.

> I believe translation should be scale-invariant, and scale should not affect rank ordering

I don't believe this is true with regard to ending angles after addition steps between vectors of varying magnitudes.

Imagine just in 2D: vector A at 90° & magnitude 1.0, vector B at 0° & magnitude 0.5, and vector B' at 0° but normalized to magnitude 1.0.

The vectors (A+B) and (A+B') will be at both different magnitudes and different directions.

Thus, cossim(A,(A+B')) will be notably less than cossim(A,(A+B)), and more generally, if imagining the whole unit circles as filled with candidate nearest-neighbors, (A+B) and (A+B') may have notably different ranked lists of cosine-similarity nearest-neighbors.

Thanks!

So of those 3, despite the superficially "large" distances, 2 of the 3 are just as good at this particular analogy as Google's 2013 word2vec vectors, in that 'queen' is the closest word to the target, when query-words ('king', 'woman', 'man') are disqualified by rule.

But also: to really mimic the original vector-math and comparison using L2 distances, I believe you might need to leave the word-vectors unnormalized before the 'king'-'man'+'woman' calculation – to reflect that the word-vectors' varied unnormalized magnitudes may have relevant translational impact – but then ensure the comparison of the target-vector to all candidates is between unit-vectors (so that L2 distances match the rank ordering of cosine-distances). Or, just copy the original `word2vec.c` code's cosine-similarity-based calculations exactly.

Another wrinkle worth considering, for those who really care about this particular analogical-arithmetic exercise, is that some papers proposed simple changes that could make word2vec-era (shallow neural network) vectors better for that task, and the same tricks might give a lift to larger-model single-word vectors as well.

For example:

- Levy & Goldberg's "Linguistic Regularities in Sparse and Explicit Word Representations" (2014), suggesting a different vector-combination ("3CosMul")

- Mu, Bhat & Viswanath's "All-but-the-Top: Simple and Effective Postprocessing for Word Representations" (2017), suggesting recentering the space & removing some dominant components

Not sure you can judge whether these modern models do well on the 'arithmetic analogization' task based on absolute similarity values – & especially L2 distances.

That it ever worked was simply that, among the universe of candidate answers, the right answer was closer to the arithmetic-result-point than other candidates – not necessarily close on any absolute scale. Especially in higher dimensions, everything gets very angularly far from everything else - the "curse of dimensionality".

But the relative differences may still be just as useful/effective. So the real evaluation of effectiveness can't be done with the raw value diff(king-man+woman, queen) alone. It needs to check if that value is less than that for every other alternative to 'queen'.

(Also: canonically these exercises were done as cosine-similarities, not Euclidean/L2 distance. Rank orders will be roughly the same if all vectors normalized to the unit sphere before arithmetic & comparisons, but if you didn't do that, it would also make these raw 'distance' values less meaningful for evaluating this particular effect. The L2 distance could be arbitrarily high for two vectors with 0.0 cosine-difference!)

If by 'doc2vec' you mean the word2vec-like 'Paragraph Vectors' technique: even though that's a far simpler approach than the transformer embeddings, it usually works pretty well for coarse document similarity. Even the famous word2vec vector-addition operations kinda worked, as illustrated by some examples in the followup 'Paragraph Vector' paper in 2015: https://arxiv.org/abs/1507.07998

So if for you the resulting doc-to-doc similarities seemed nonsensical, there was likely some process error in model training or application.

When you describe a tax that is "paid annually and only if you win", that's plain generic income tax.

That's not the gambling-activity-specific taxes that Stoller's article discusses - typically applied to gambling businesses' revenues, not bet winners specifically.

Huh? What is Loeb's 'captive audience'?

No, I can't, because all I've seen from Loeb is pretty clear about what is fact, and what is speculation.

I don't know, and there's no way to find via "Google search", what HN user ~dylan604 is specifically alleging has been improperly "pushed as fact".

If it's clear to you, can you share a representative quote from Loeb? He's got a lot of writing to choose from!

Why would someone "with his standing at Harvard" be expected to avoid "wild, and public, hypothetical[s]"?

Does everyone at any prestigious institution have some duty to remain conventionally mundane in all their musings?

Is there any reason to think such hypotheticals are, on net, more harmful than helpful?

Isn't tenure (like Loeb's) designed to encourage a fearlessness around topics & speech?

What is Loeb pushing as fact with no evidence? Can you provide a representative quote?

WTF is 'SFP'?

Yes, though it's possible a more-general core model, further enhanced with some other ways to bring those texts-of-interest into the working context, might perform better.

Those other ways to integrate the texts might be some form of RAG or other ideas like Apple's recent 'hierarchical memories' (https://arxiv.org/abs/2510.02375).

Imminent Death of the Archive Averted, Film at 11.

The LLMs need the embedding function, benefit from growth, do the training – and then other uses get that embedding "for free".

So an old down-pressure on sizes – internal training costs & resource limits – now weaker. And as long as LLMs are seeing benefits from larger embeddings, they'll become more common and available. (Of course via truncation/etc, no one is forced to use larger than works for them... but larger may keep becoming more common & available.)