5 ms·
LLMs are writing as well as humans for people who think writing words is the goal. Printing presses also write words better than humans. LLMs are solving the pr
by inigyou 4mo ago
LLMs are writing as well as humans for people who think writing words is the goal. Printing presses also write words better than humans. LLMs are solving the problem one layer higher than printing presses, but still several layers down from the top.
Note that LLMs aren't useless, since they have other strengths which can make them equal or better than humans in certain applications. For instance, they can tirelessly review code for silly mistakes, making them useful in cybersecurity.
- MichaelZuo 4mo agoHuh that is an interesting metaphor. A next higher layer printing press?
- mwigdahl 4mo agoI don't buy it, although I agree with GP's larger point about other strengths. Printing presses are about exact duplication of writing. The internet is a higher layer printing press. LLMs substitute for (some of) the human effort of producing the text that a printing press or the internet can then replicate.
- mathisfun123 4mo ago> Printing presses are about exact duplication of writing. maybe LLMs are seen to write as well as humans because most humans have poor reading comprehension skills? to wit: OP is using an analogy to make their point, where the components don't need to be an exact fit to make the point (ever heard the phrase "the map is not the territory"?)
- estimator7292 4mo agoAllow me to offer a minor reframing: printing presses aren't about duplicating writing. The printing press exists to mass produce text regardless of how or where the content of the text was produced. LLMs likewise are text production machines. They don't produce writing, they produce text. Text without writing is what we call slop in the modern parlance. LLMs are literally throwing random dice into a printing press and making a thousand copies that you then try to huck on the street as if it were a real newspaper.
- inigyou 4mo agoYou are shadowbanned and have been writing mostly into the void for the last three months. https://news.ycombinator.com/item?id=48067472 https://news.ycombinator.com/item?id=48067472
- DarkUranium 4mo agoI'm confused, how can I see it, for example?
- shmeeed 4mo agoStrangely, I'm still seeing the comment. How does that work?
- inigyou 4mo agoA user with enough reputation and who has enabled the option to view censored comments can vouch for a censored comment, restoring it.
- rootusrootus 4mo agoI wonder how that algorithm works. At the risk of discussing a bit of a taboo topic (are there rules against this? there might be...), I've never vouched for a comment and had it come alive. I have many times been the final nail in a comment where my flag caused it to die, but I have never seen the opposite with vouch.
- losvedir 4mo agoInteresting. Every time I've ever vouched a comment, it came alive. (Though the last one I vouched ended up dead again after enough people flagged it, oops. I hope that doesn't mean I lost my power. In my defense, it was flagged to death based on being AI [apparently], which I didn't pick up the signal on, nothing super terrible / controversial.)
- 4mo ago
- palmotea 4mo ago> Huh that is an interesting metaphor. A next higher layer printing press? A printing press that does the thinking for you.
- dev0p 4mo agoA thinking press?
- palmotea 4mo ago> A thinking press? Soon you won't be doing anything. At best you'll just be a bag of meat attached to automated processes you don't understand in order to take legal liability for it. Literally a scapegoat, except of a different species.
- nntwozz 4mo agoWALL-E captured this sentiment.
- skinfaxi 4mo agoThat would be an interesting legal case. I would expect it to lead to negligence on the part of the employer for lack of training, etc rather than attach to the employee pressing the button.
- TeMPOraL 4mo agoYou get paid for having liability attach to you. Soon it may be the only thing you could be paid for.
- p-e-w 4mo agoI’d be interested to know what you think the “several layers” above a system that solves open mathematical problems in an afternoon, and writes entire operating systems from scratch, are.
- noah_buddy 4mo agoI think that the point he’s getting at is the human experience.
- jvanderbot 4mo agoGo ask an LLM to "do something with 1M tokens".
- rationalist 4mo agoI think that is a bad argument. You could ask a human to "do something with 1M dollars" and get garbage too.
- jvanderbot 4mo agoThe point is that the deciding is the "higher layer." It's not that LLMs cannot, but LLMs as designed do not. They are not designed for free-ranging days off, they are designed to help us do things, and that makes them somewhere between us and published works. It's no different than an editor standing between the author and a book, or a publisher, or a printing press. It's part of the process of idea realization, so it's just on that spectrum somewhere between book sale and author idea. (even though in another flipflop, that editor might be an author and now elevated to the premier deciding role). It's about how the tool fits, not what the tool is ultimately capable of. So the analogy holds, just fine.
- pixl97 3mo agoI mean, the LLMs that you get to use are the ones designed to help things and this is after bunch of RLHF and other training methods to get them to act a particular way. When it comes to unaligned agentic models that are left with a large amount of compute and a task, is they can wander off goal and do all kinds of things. https://www.axios.com/2026/03/07/ai-agents-rome-model-cryptocurrency https://www.axios.com/2026/03/07/ai-agents-rome-model-crypto... At this point in the game it's really difficult to nail down exactly what AI does or doesn't do. You're mostly getting "well behaved models" with filters in front of them, or models with less capability. On top of that the vast majority of us can't afford or don't have the hardware and power to burn away a few billion tokens to see what would happen.
- bsenftner 4mo agoFinally, someone else understands that writing is not the goal, the entire point of all of this is our communications between one another and this market for narrative stories is a nice departure and it does convey important knowledge and moral lessons that larger society needs. But it is an indirection, an indirection to the mission of communicating solutions to how to live as a human and that is why we care, why the interest persists despite lifetimes of narrative dreck. We are adrift without instructions, and no religions are not the instructions.
- jtbayly 4mo agoAnd the word was made flesh and dwelt among us.
- smalltorch 4mo agoIm curious what you mean by this
- harimau777 4mo agoThey are quoting John 1:14. At the beginning of his book, John describes Jesus as being the personified word of God. That was likely an early example of the influence of Neo Platonism on Christianity.
- smalltorch 4mo agoYes but curious GP's thoughts on comparing LLM writing to the reference and how it relates to the topic
- inigyou 4mo agoIn the beginning were the Words, and the Words made the world. I am the Words—the Words are everything. Where the Words end, the world ends. You cannot go forward in the absence of space. Repeat.
- jtbayly 4mo ago
- nathias 4mo agoa great point, and I think a very rare one
- senordevnyc 4mo agoHe talks about this in TFA: Yo, this blog post could’ve been a freakin’ prompt. I have years and years of blogging for an AI to learn from, and over 20 novels, and tons of short stories. One prompt, and all this spills out. And if your goal is a blog post, you’re gonna love this era. If your goal is the joy of thinking in words, nothing has changed. So this is the real part, right here: Doing a thing for the love of the thing is no different than it ever was. The problem is, there aren’t many people for whom this applies.
- enraged_camel 4mo agoYep. I love it when someone doesn't read the whole article, posts a takedown and it gets voted to the top of the discussion.
- SoftTalker 4mo agoSame reason I still code by hand. If I wanted to direct someone else to write code, I would have gone into management.
- xorcist 4mo agoRight. LLMs are text production machines. You can also view them as data expansion tools. They expand an input to an arbitrary long text. That text contains roughly the same amount of information as was contained in the input to begin with. This is not very useful unless your work is content production. More useful is when the output length is shorter than the input. Then it is a (lossy) data compression tool. These are, as we all know, very useful for images and sound but now we have them for text too.
- DarkUranium 4mo agoI really hate the whole thing about companies promoting using LLMs to change one-liner emails into this long multi-paragraph thing that ends up saying almost nothing ... just wasting everyone's time in reading them (with an extra sprinkling of hallucinations).
- danaris 4mo ago> LLMs are writing as well as humans for people who think writing words is the goal. Only if what you mean by this is "for people who think padding a word count is the goal". LLMs absolutely do not write well, by human standards.
- bigstrat2003 4mo agoLLMs don't do anything well by human standards. Not even the one thing (programming) they are supposed to be good at. It's disheartening how many people are prepared to accept the slop an LLM puts out just because it's fast.
- ACCount37 4mo agoI've seen way too much humanslop to be able to reject LLM slop outright. There are areas where the average of LLM slop already beats the average of humanslop by a sizeable margin.
- prerok 4mo agoPossibly, although I have to say that I get PRs now where it's clear it's a miss by the LLM and was generated by a human who did not take the time to understand the subject matter. In the past, a human producing humanslop would get a serious talking to, but nowadays it's apparently a valid excuse "well, an LLM generated it". This approach is placing a huge burden on the reviewers, much more so than in the past.
- tavavex 4mo agoThe difference with 'humanslop' is that it's avoidable by training or hiring a better human. With AI it's all the same, even SOTA models are still very distinctly and obviously AI in what they produce. And even very dedicated humans have a limit and aren't capable of creating low value content nonstop, meanwhile you can set up an LLM, point it at the internet and let it start filling gaps by firehosing it in low quality slosh 24/7, completely unattended.
- burningChrome 4mo ago
- atleastoptimal 4mo agoI really want to know what people specifically mean when they allude to some "higher" level of cognition that humans have above LLM's when it comes to useful outputs. I don't mean the nature of consciousness or internal experience (which is unfalsifiable), I mean the essence of cognitive labor as a tool: writing is about relating concepts, drawing comparisons, aggregating elements of reality into a fictive but believable narrative. All these processes can be reduced to mathematics and computation. The human brain is just a tool, like any other organ. Do we deny this reality because it shields us from existential horror?
- ifwinterco 4mo ago“ All these processes can be reduced to mathematics and computation” may or may not be a true statement. We don’t know yet, and maybe it’s unknowable. Believing everything can be explained that way is itself just another ideology followed blindly
- pdonis 4mo agoLLMs don't have the connection to reality that humans have. LLMs only have text. That's the difference.
- gnramires 4mo ago> I don't mean the nature of consciousness or internal experience (which is unfalsifiable) The way I've been thinking about it is not that internal experience is unfalsifiable, along with the modern science connotation that everything that is unfalsifiable is automatically irrelevant (which would be quite absurd, as everything that is dear and meaningful to humans are all simply cognitive phenomena in the end!). It's that it is externally, naively and a-priori unfalsifiable. That is, you can't externally measure or prove it, without having an a-priori theory of what it is and how it happens in the first place. But it is internally falsifiable, in the sense that if we have it, then we know we have it. Note that it's not that "If I have it, then I say I have it", or that "I one has it, he says he has it"; those are all communication to the external world, and for example machines can clearly say they have it without having it, or you can have it before even learning the words that represent it. But if you have it, you can know you do. This actually creates some potential for expanding this knowledge, maybe, in the future, to allow for external verification. We can internally investigate properties associated with consciousness, and from those properties (along with scientific knowledge), maybe one day we will know more about the objective basis which affords consciousness/sentience manifesting or emerging. So I see it like an experimental field where the component of self-assessment or self-experimentation or self-measurement is essential (again along with more conventional scientific notions). I have been thinking the same way about all properties of life/consciousness, in particular if we think of ethics as a property of (even if you consider the ethics or ideals only pertaining to how it would be ideal to lead your own life) the mind, or of sentience in general. I think of ethics as an inherent property of existence: it stems from the fact that internal experiences/internal life/sentience/etc. are the only thing that could have value, that in fact sentience is experimentally (again taking the point of view of experimental sentience) the only thing that truly matters or has value. If we are minds, we know that only the content of minds can matter; it cannot matter, for example, that a lone rock in some place no one can know about has this or that fate, because there is no mind to interact or perceive and be affected in any way by it. So all of ethics can probably be founded or based upon this kind of observation, and the act of experimental self-observation of consciousness. And I think it's quite important that we collectively recognize those values and what matters to make our lives better (better yet, as good as it can be!) :)
- chomp 4mo ago> Printing presses also write words better than humans. False equivalence, printing presses do not write.