vector_spaces

As someone who sometimes helps companies figure out content strategy, I am sorry to say that it is not true that linking in comments is safe -- this tends to penalize your post nowadays.

LinkedIn isn't alone here -- virtually every platform with an algorthmic feed penalizes both links and the obvious workarounds.

Not to defend the practice -- it's unfortunate because it trains younger users that attribution isn't important, it enables and incentivizes outright theft of content (good luck chasing down original sources after convincing yourself that it wasn't AI in the first place)

Judgments are not generally impossible to automate -- judgements are typically binary or quantifiable interpretations, so in some sense are perfect targets for automation, but the sheer volume of judgements needed to build something coherent is overly cumbersome to specify to the point of being intractable. There are also many hidden judgements, ones where the thresholds may not be well understood, and interactions between them.

But humans still manage to wrangle these, sometimes seemingly effortlessly, through a process which we call by shorthand "taste". This is a largely vibes-based heuristic that combines expertise with life experience and cultural training -- intuition, more or less.

This is likely not possible to automate either -- aspects of it may be automatable for a given expert, in small pieces in narrow subsets of their particular domains of interest, but even those likely will require some manual intervention.

This is in part because it is, to a large degree, a black box, even to the expert deploying it. With some self-awareness and strong language skills we can articulate approximations of the judgements that go into taste. But even those will fall short, as even the most self-aware individual will fail to notice certain judgements and dependencies.

In practice many of these are not even explicitly articulable. Humans are idiosyncratic and messy and dynamic, and the suggestion that we can build a machine that approximates this in a way that pleases our sensibilities and doesn't require supervision is kind of ludicrous, even in light of recent developments.

The certitude with which you condemn your peer without the slightest consideration for subtlety or alternate explanations for what you've understood to be true -- this comment is such a beautiful encapsulation of the state of discourse today. In some ways it is the modal comment of the moment. Tastelessness and mediocrity and inability to grasp nuance or feign at humility are on full display, not to mention a fantasy for violence

It's a work of art

Do you "know", in any meaningful sense, any of OpenAI's recently publicized proofs? Do you suppose that there is any large community of non-academics that does?

One of the points the parent makes, along with the TFA, is that academia -- or more specifically, the "mathematical community"-- is a setting primarily for creating and ingesting mathematical knowledge, and disseminating it to the next generation and to other fields. Humans absorb this material slowly, through lots of discussion and collaboration -- it is necessarily a slow process. Facilitating this is one of the important functions of academia. Your usage of academic as a slur here is a bit silly for this exact reason.

I don't claim it is perfect, and we can argue about pedagogy in elementary courses till the cows come home. That's not really material. But this is one of the only settings in which such knowledge is broadly valued for its own sake, and in which there is a semblance of incentive to help others "know" this stuff as well, be they future generations of mathematicians, science and math educators and communicators, practitioners in other fields, or genuinely curious amateurs.

"What if I made X?" is basically Product Development Antipattern #1. Most feedback you get here is incredibly low signal. The stakeholder you're talking to isn't a technical person in all likelihood, so they probably aren't visualizing what you're describing in the same way you are. There is a high probability that they are telling you it sounds great as social lubricant or because they don't really understand what you are talking about and don't want to look dumb. Not to mention that positive response is the path of least resistance for most people -- the hard work of thinking critically about an idea so that you can locate potential flaws requires caring in the first place, and most people assume positive feedback is what you want to hear anyway.

There have been a few papers recently suggesting that ChatGPT responds differently to different demographics. Specifically, depending on your gender, education level, socioeconomic status, race, and other characteristics, or how it reads those, it might give less accurate responses to the same prompts. These unfavorable outcomes are generally unfavorable in the ways that one would expect of course

https://www.sciencedirect.com/science/article/pii/S187705092...

I have never been in the situation of knowingly taking salary from any fascist and would never tolerate it. I've left jobs I needed over this. My "walking away" was more about wanting to work on stuff in my day job that felt somewhat more worthwhile and tangibly useful to humanity

I completely agree with your critique here but I will say that I have had difficulty understanding pure math texts because of 'colorful' language -- e.g. "chattier" treatments of mathematical analysis. But I will also say that I'm glad such treatments of the material exist, partly because I know that what appeals to me doesn't appeal to others and these might be a valued resource for them, and because often pushing through and understanding texts that were difficult at first helps us to see the material from a different perspective. I think this optimization perspective squashes this possibility a bit when taken too far, as it usually is

For the record I am with you and OP in the sense that I personally would struggle with such proximity. But the reality is that even if you do what I did and give up quite a bit of salary in exchange for something so mundane as the same work in a public university-run health system, you can still often follow the money to some ugly places that are not as far away as you'd like. Regarding startups, the ones like yours are probably the best one can realistically hope for, ideological inclinations of the founders aside

> the majority of the public

This requires some citation or justification.

Diagnosis is not a criminal conviction in a North American court of law. In particular, reasonable doubt -- particularly among casual, non-clinical, self-deputized jurists who don't generally have any insight into the patient's chart, let alone medical qualification whatsoever -- is not a sufficient justification to withhold diagnosis.

In your post, you are leaning a bit too heavily on vibes to do the heavy lifting in making your points. For example, you claim there exist people with epilepsy who prefer to be labeled as autistic, but you don't say how many people are doing this, why they are doing this in a way that doesn't strawman, nor why this is even topical.

My interpretation is that you are hoping the reader will have specific visceral (vibes based) reactions to these facts and the drama of the spectacle you've described -- e.g. "yikes, some morons are conflating correlation with causation/identification and using that faulty reasoning to claim a diagnosis they don't deserve". Please correct me if I'm misrepresenting what you are trying to communicate, it is not my intention

My suspicion is that this is a tiny number of people to begin with, and not an indictment of autism diagnosis in the way you believe it is. In fact, most of the people you seem to take issue with are self-diagnosing. What specific "methodology, diagnosis, and relevance to every day healthcare" are you taking issue with then?

Buying a company is very rarely about buying the tech, at least principally. It is true, as others have pointed out, that Nvidia could have "rolled their own" Hugging Face far more cheaply and probably in about the same amount of time as it will take to execute the deal end to end. That should give you a clue that the tech isn't really the target here

More than anything, they are buying a position within the ecosystem. Many of us could get a version of an app deployed in short order that implements most core HF functionality, but I am sorry to say that this is very far from the "hard part" of building a business, and this was the case even before coding agents were a thing. It's no small thing to become the first tool pretty much everyone reaches for in a particular niche, and therefore the incumbent in that niche

And there's lots of inertia keeping companies and developers on these incumbents like HF and GitHub -- see also Microsoft's play here dumping billions into a company that is basically peanuts and probably nowhere near as profitable as Microsoft normally would like. But they bought it at a time when they were trying to make Microsoft synonymous with open source, and it was a great play to that end. Likewise Nvidia wants to be synonymous with AI. Even if HF is a loss leader, the positioning is probably worth it

To a lesser degree, Nvidia probably also might like to reinvigorate its flagging cloud compute business and this could be a natural way to do that. At least TechCrunch thinks this, but I, an overconfident layperson, feel that the positioning is where most of the value is for Nvidia, and consolidating that will act as a force multiplier for that + anything else they may want to do with this space

RE AI writing and slang, the particularities of the examples aren't so important. My point wasn't to comprehensively catalog of every instance where we're losing something by not having timestamps on textlog specifically. The point was that time -- even relative time if not timestamps -- can be used to understand a piece of content -- often they are even critical, which you get.

I like the idea of playing with granularity. I'll also suggest: one way to make timestamps available but less of a focus is to add friction. Require an extra click into a post metadata page that loads with a full page refresh. Any friction introduced reduces a feature's usage, and even as little as just adding an extra page refresh tends to be enough to reduce its usage by an order of magnitude or more. If you wanted to be particularly sadistic you could require the user download a pdf with that data lol

By "mints" I meant the family lamiaceae, so pretty broadly which contains sages, rosemaries, basils, teak, verbena, and in hindsight of course contains examples like this which straddle medicine (depending on dosage and circumstances -- I'm not advocating it be used as such) and poison. Even if just restricting to the genus mentha apparently. Anyway good call

There has been some research on spearmint suggesting anti-androgen effects in PCOS, which is a bit unexpected to me as I wouldn't expect mint-family herbs to have much hormone modulating effect at all. I would imagine there to be similarities in profile between spearmint and peppermint so perhaps a related mechanism is involved, although I would probably expect the opposite outcome as a naive layperson. But I haven't closely evaluated the paper nor do I know whether there's been much else suggesting this

It is important to remember that herbs are difficult to study in this setting compared with isolated compounds -- likely because their pharmacological and chemical compositions are sensitive to many factors: hour and season of harvest, form consumed (raw powder, alcohol extract, cold water infusion, fat emulsion, hot water infusion, combustion, pyrolysis, essential oil), part consumed (leaf, root, flower, seed), among staggeringly many others. Anyone with experience consuming cannabis or coffee can probably testify to the above

Not to mention that the pharmacologically active compounds in many plants are far from being comprehensively understood, let alone their interactions and dynamics

This isn't to say that every claim about herbs is to be taken seriously. My point here is more that there is no good reason to dismiss the herb itself as a cause yet simply because it's an herb and a fairly mundane one at that. There are easy examples of herbs which have pronounced physiological effects, used for recreation, performance enhancement, as literal poison, to induce labor, to thin the blood -- among others.

https://pubmed.ncbi.nlm.nih.gov/19585478/

The irony here is that modern social algorithms actively conceal these for exactly the same reason. Social feeds want to show you a steady stream of content that you are likely to engage with. They are incentivized to frame content as literally timeless and impose a continuous, never-ending present on the user, since they don't want recency to get in the way of engagement. But the time-context in which a piece of media was produced is a fairly important detail in critically evaluating it and situating it in an actual, human conversation.

Did this piece of writing with common AI smells get posted before or after 2021? Did the author write this piece about someone they admire and aspire to before or after the murders became public knowledge? What was the expert concensus on the topic when the author wrote about it? What meaning did this slang have back then?

What technologies were commonplace? He said he needs help with something I'm an expert in -- can I help him, or was it over a decade ago?

What was happening culturally when this was written? What was the weather like? Did this collection of posts with a dour tone get posted during a global pandemic when mostly everyone was self-isolating? Did he write this post about his loved one before or after her death?

I mean hopefully I don't need to sell you on why this information is useful and actually antithetical to addictive, low nuance, engagement-driven social media. We need more media literacy, not less, and de-emphasizing critical details like this seems problematic

I do sympathize with the challenge and know that there aren't really easy answers here, and every possible change is likely to ruffle someone's feathers, but I have a feeling there is another way to achieve what you're after.

One tip: I suggest that you stop cluttering your writing with parentheticals.

If it's worth saying at all then it's likely not actually a parenthetical. If it's not worth saying, then you should cull it. Yours seem to mostly fall in this latter category, in my assessment anyway

Decide what you are trying to communicate and be sure that every word is in service to that message and that message alone. For example, you are discussing the practice of writing. That's why I'm in this thread, as a reader. In the middle of a sentence about finishing your novel, you mention via parenthetical that you're looking for an agent. Maybe the intent here was self-promotion -- you hope an agent or someone who knows one will read this -- fine. But if you can't decide and plan to use a parenthetical instead, know that the reader will be paying a dear cost: the sentence containing the parenthetical will be substantially more painful to read. Subsequent uses will be worse.

By that second paragraph, you blew your parenthetical budget, and yet you went on to deploy three more. Parentheticals need to be vanishingly rare as they are an eyesore. Yes, if you'd like to die on that particular hill, you can use them with stylistic intent. Go for it. But for the reader it will probably grate.

Likewise online you run into claims that full remote or hybrid can't and doesn't work for anyone, and we all interpret observed dynamics at work to construct theories supportive of whichever configuration is our favorite (WFH/Hybrid/In Office) or as evidence the one we dislike is worse. In reality the situation is messy and complicated and heavily circumstantial, depending on team/company culture and dynamics and working styles.

For instance, I've observed dynamics like what you describe at various orgs, and what they had in common was a lack of psychological safety. If people are responding defensively to benign messages then it tends to mean they have a reason to expect hostile messages. Hostile messages are not a given.

One common way this arises: a company ostensibly offers flexible remote work, but in practice strongly prefers in-office work, and enforces this via social pressure. Remote work has a small social cost that accumulates each full day that one works remotely. These remote-oriented workers then have to periodically show up at the office primarily to ingratiate themselves to management and their peers. This is incredibly common, and might even be the default configuration for an org that permits flexible WFH.

I don't think this is intentionally malicious, BTW. It's simply what happens when you make a policy shift without considering your team culture as well so that both working styles are supported as first-class citizens.

I expect that people who mention MUDs largely do not expect anyone to know what they are talking about, since these games have nearly always been tiny and nowhere near even nerd pop culture. At least this has been my experience -- so I will prefer to say "I am playing this game called Hesperia/Achaea/Discworld/whatever, which is like a text-based RPG" and avoid even the "MUD" terminology.

Further, I expect an even smaller subset of people who have played MUDs are aware of the 'distinct genealogical lineages' you describe. As someone who spent my teens browsing mudconnector and dabbling in lots of random MUDs listed there, I started hazily noticing patterns in structure, design, mechanics, and even networking to some extent, and suspecting that certain games existed within specific 'traditions' of MUD development or even used similar libraries or frameworks, but I didn't have the context or curiosity to dig deeper.

Like everything, this can be modeled as a funnel: of the relatively small number of people on planet Earth who know that MUDs exist, only a tiny fraction has even connected to one, an even tinier fraction of those spent meaningful time on one, and an even smaller number of these who have thought to contextualize and taxonomize what they were played

Anyway, I am curious what level of specificity would you find normal?

The comments on this HN thread are all very confusing to me and making me wonder if I read a different article altogether.

For instance, I am wondering what exactly you are responding to with this comment, since the game described in the article is not really about teaching the kid to code with AI at all? I mean, what you wrote is interesting, but it isn't really topical

The blog post that I read mentions a single generative AI feature baked into the game that allows the player to create a single item once per day using a text prompt. But it is one single feature with limited uses. The blog post devotes very little ink towards describing it, and it is very clear that the goal of the project is not to use AI to teach coding, but more to provide a constrained, hackable sandbox environment with a deliberately impaired scripting language -- so, lots of "basic building blocks" as you say.

Was there something else in the post that I missed?

ADHD is known to be able to affect speed and consistency of output, even of mastered material, and in a variety of ways -- hence "I have no idea..."

Standard diagnostic questionnaires commonly reference this. The last part of your comment reflects a lack of understanding of both human cognition and recall and ADHD in particular.

The site guidelines also say to assume good faith, to converse curiously, edit out swipes, not go fulminate, and to avoid grandstanding.

Before calling me dishonest, did you stop to consider that maybe you didn't understand my comment, or that you weren't clear with your meaning?

I genuinely don't think you understood what I wrote.

If I operate a business, you are not necessarily entitled to transact with me how you please. I can implement whatever rules I want so long as I am not outright committing discrimination against a legally protected category and so long as I am upholding other fiduciary obligations.

For all you know, the site operator is opposed to AI agents on moral grounds, or has beef with the developer of the specific LLM interface you are using.

More practically, an e-commerce website often won't make a profit if a purchase is too small. If they find that LLM users are easily bypassing their upselling tactics that usually result in higher basket sizes, then ideally they would adapt their approach to LLMs or implement a minimum order size, but they are also perfectly within their rights to simply ban you.

I am not saying that this is the likely reason -- just that there are reasons, many, and often they aren't very obvious unless you are in their shoes.

Every kind of business does this sort of thing! A software consultant or designer might only take money from healthcare companies, and not talk to anyone offering (apparently, to a naive outsider) equivalent work in a different industry. A restaurant might implement a dress code. A cab driver might insist you not wear strong perfume.

Some rules that externally seem flippant and unreasonable exist for reasons that are entirely sensible, but also entirely nonobvious externally. In any event, I don't agree that your usage is equivalent to a manual session, if it were, the operator wouldn't be able to detect it! Therefore they are within their rights to detect this and disallow it. There are plenty of reasons that they might do this, and unfortunately you probably don't have the information needed to interrogate those in a meaningful way

You are assuming that because you are giving the operator money that you are entitled to use their website however you please -- that's not how it works. If you enter a brick & mortar establishment and break their rules, even as a paying customer, you might risk e.g. being kicked out and banned, depending on the behavior. This is not unique to e-commerce

To be fair, I do not generally support wholesale banning of scrapers and non-browser clients, since this tends to have lots of ugly side effects.

But regardless, I'm interpreting your last comment as saying that you should be entitled to behave as you please because you are a paying customer. If that's the case, then you are in for a rude awakening one day

I had that one too. I remember I used to crunch on the rice from the bag it came with (I was like 8).

I will say that yes, sadly, I have browsed through code written by people, and sadly also have browsed through code written by LLMs. Who do you suppose wrote the code the LLM was trained on? Divinity?

I'm being tongue and cheek, but as someone who uses these things extensively in work and personal projects, I want to point out that it's new or novel that generating lots of plausible code is now vastly cheaper than it once was, so the scales are quite different.

And yes, slop is an apt moniker for something that requires lots of careful prompting and configuration and guardrails (which is extremely sensitive to small changes in these initial conditions in ways that can be fairly nonobvious) to produce something that will in most cases read at least a little bit worse than if a human had written it. And it will still go off the rails sometimes!

To be clear I don't really agree that typing out the code generated by LLMs is a reasonable counterweight, it largely defeats the purpose. I agree that at a certain point you have to treat parts of your code base like a black box and enforce modularity. But this is still nontrivial, especially in more specialized domains with lots of nuance

I'm genuinely a bit floored reading the comments here, but I guess my idea of the average HN commenter's ability to talk to non-technical folks about technical topics is biased because I work and socialize with lots of people who don't have technical expertise (thankfully along with other technical folks who also have lots of experience talking to the former group). Many of them don't even know what Claude is, let alone Claude Code.

No worries, I thought so, and agree with your point -- I'm a fan of your work and datasette historically was something I leaned on often at work and in personal projects. I have read some interpretations of recent developments that misunderstand what programmers and mathematicians do and have done, so was filling in the gaps in what you wrote to speak to some of those

I'm grateful for the work you've done!

> One of the arguments for open source software for end-users has always been the freedom to examine and modify how that software works.

A corollary is that more people will have eyes on the source code than the original developer(s), so the project can benefit from a broader skill set and set of orientations and priorities than if it were proprietary. Not just on the development front, but hopefully also in security, usability and accessibility

> The reality for most people - even expert programmers - has been that the freedom is more about being able to lean on other people to do that. Most people can't justify the time commitment needed to read and then modify the code for tools they use very often.

I imagine you are correct that most people historically didn't inspect or modify most tools they use, but most developers I know who have relied on open source software -- even if they didn't contribute directly upstream -- have had the experience of locally patching a small bug or adding a small enhancement to suit a use case, or studying a piece of an open source module as a starting place for their own implementation for some piece of functionality, or doing a cursory scan of a project to make sure it isn't e.g. phoning home, or simply taking a morbidly curious look to see how the sausage is made.

Sure, I'm not a representative sample and it would be great to have some actual numbers on this, but I have observed this among engineers at Fortune 500s and YC startups, but also non-tech businesses as well as at local public utilities and health systems.

Anyway, I agree that LLMs can provide a lot of value here and reduce the friction but I question the narrative that open source wasn't providing much value previously because most people didn't have time to read code. It provides value even without developers reading code in the first place, and reading code was historically one of the essential parts of the work

Beyond what was provided natively in ChromeOS, you could often get something like GalliumOS working on them. My old Chromebook was fairly painful to use as a Chromebook, but GalliumOS made it pretty snappy and fun. Unfortunately that project seems unmaintained now

What is the point of a sock puppet if you can't be explicit about what specifically you found worrying there? And I thought I was neurotic

Anyway recently I watched a video in which Alan Derschowitz compared Jews who voted for Mamdani to Jews who voted for the NSDAP. The ridiculousness of the comment aside, I was curious to what extent this was actually a thing. It did happen but it was a tiny number of fringe German nationalist Jews, <1% of voting age Jews -- but still a fair bit more than what one might expect given that the NDSAP was not mincing words when describing their intentions with Jewish people.

Those who supported them regardless included Max Naumann, who went on to collaborate with the Nazi government as leader of the VnJ movement, through which he pushed for Jewish pacification and assimilation as an answer to antisemitism. There were others -- though vanishingly few -- who collaborated in various ways, with enthusiasm seemingly beyond what was coerced.

Those who were publicly Jewish usually ended up being targeted by the Nazi regime in the end regardless -- VnJ was outlawed and Naumann was picked up by the Gestapo a few years before he died of cancer. The Nazis viewed Jews as politically irredeemable so extensive collaboration was not generally sufficient to save anyone

I recently googled the phrase "German Jews who voted for NSDAP" after seeing it mentioned in passing

The AI overview contained the construction "Consensus on Reddit is..."

What a time to be alive, I guess