Artists and actors could point the LLMs at programmers by developing their own apps, doing their own engineering. Do they? Who's doing it already? Tilly Norwood's creator seems like a decent start.
I have no idea about the current article, and given that the author is the person with the first commit in the Kubernetes repo (https://joe.dev/about/), he obviously has a lot of credibility.
Just generally though: what we're seeing a ton of these days is people writing something and then passing it to an LLM with a request to improve it somehow, e.g. by fixing grammar, tightening the style, etc. In such cases, the answer to your question is that the "prompt" is (1) a first draft, and (2) an instruction to edit it.
It's clear, though, that the LLMs leave far more imprints on the text than most people realize, and that although they may have asked the LLM to restrict its edits to "just" X or Y, the actual changes to the text will often go beyond that.
How this will evolve over time is anyone's guess, of course.
I do this so many times. Type in a large amount of text and the only thing final in my mind is para breaks and the idea per paragraph. And then give it to AI saying "sending to director", "sending to friends on WhatsApp group", "sending to colleagues" and it does an awesome job of bringing the "AI polish" and then you edit or negotiate line by line or para by para on what you want to keep.
Yes, this is certainly common, and opinions and tastes differ about the outcomes—as they should, because we are all still in the early stage of sorting through the best way to use these tools. I think it's also already clear that "best way" means something different in different contexts.
Where some people are getting into trouble, at least in the HN context, is underestimating the impact that this has on their text. There's a big perception gap between the author's view ("fixed up the grammar a bit") and the reader's view ("this sounds entirely like an AI wrote it") in many cases. So many, in fact, that I feel I can say something about it. I'm no authority on any of this and don't want to sound like one, but this is such a common pattern at the moment that I feel confident reporting it. How it will change over time, I have no idea.
(I also don't want to sound anti-LLM - we rely on these tools heavily, they're amazing, they've already improved HN, and they show every sign of high potential to improve it further. The bottleneck isn't the LLMs, it's how quickly we can figure out how to use (and test) them. We just don't use them to process any text that we put on HN itself.)
And that is sad. I had speech impediment while growing up, but somehow I was able to focus on writing and used to write real well, first technical (sciency) stuff and then general thoughts too. Despite having never spoken english for conversation till the age of 16 or 17, someone decades later speculated that I was an english major because I could communicate really well (and I did use some heavy but very very context appropriate english words). Nowadays ... while I do use AI now and then to clean up my text, I have been accused of being an AI more than I can count :) You put in the effort and get accused of being chatGPT.
I have a steering .md file that instructs Opus how not to sound like an LLM when writing prose (I write prose in my IDE with Opus). The steering is specific to me, but I've found that giving Opus rules like eschewing punchy journalistic sentences ("Not because X. But because Y. And that matters."), varying sentence lengths and avoiding staccato sounding clauses go a long way in smoothing out LLM smells in writing (at least according to me).
Aside: different LLMs sound different too! ChatGPT is the worst offender for LLM-sounding writing and needs the most smoothing, but Claude (web) actually sounds like a humanities major from the get-go.
Attached is the writing.md I use to steer Opus 4.8. Prompt:
"use the writing.md steering on x.md and loop until all LLM traces are removed".
I ran it on TFA and Pangram flagged it as LLM generated but Claude Fable couldn't definitively tell.
-- writing.md ---
# Writing Rules (MANDATORY)
## Banned Words and Phrases
Never use: "incredibly", "extremely", "absolutely", "fundamentally", "dramatically", "crucial", "vital", "powerful", "robust", "elegant", "seamless", "cutting-edge", "game-changing", "groundbreaking", "It's worth noting", "Importantly,", "Interestingly,", "Let's dive in", "At its core,", "At the end of the day,".
No exclamation marks in technical writing. No contractions in formal writing.
## Banned Sentence Structures
1. *Semicolons joining independent clauses.* Do not write "X does A; Y does B." Use a comma + conjunction that names the relationship: ", while" (contrast), ", and" (addition), ", so" (consequence). Semicolons hide the logical link and sound artificially balanced.
2. *Label-colon-explanation.* Do not write "The key insight: ..." or "The limitation: ...". State the point directly or use "is that" phrasing.
3. *Colon after a bolded term.* Do not write "a *rollout engine*: a lightweight...". Use a comma.
4. *Sentence fragments as assertions.* Every claim needs a subject and verb. "No gap at any ρ." → "There is no gap at any ρ."
5. *Em-dashes joining independent clauses.* Do not write "X does A — Y does B." Use a comma + conjunction. Parenthetical em-dashes ("the policy — trained offline — cannot adapt") are fine.
6. *Tricolon lists of near-synonyms.* "It does not X, Y, or Z" is padding unless each item is genuinely distinct.
## Banned Rhythms
1. *Staccato sequences.* Two or more consecutive short declarative sentences of similar length. Join them with a conjunction or subordinate one. A single short sentence standing alone for emphasis is fine and often good. Do not eliminate it.
2. *Formulaic layout.* Do not produce: intro paragraph → three bullets → summary paragraph.
3. *Gratuitous parallelism.* Do not force list items into identical grammatical form if it makes them sound robotic.
4. *Saying it twice.* If you stated a fact, do not rephrase it from another angle in the same paragraph. One clear statement is enough.
5. *The negation-correction reversal.* This is the move where you deny one candidate and assert the real one. Surface forms to match: "not X, but Y"; "it isn't X, it's Y"; "X was never the point, Y was"; "for me X, for them Y"; the comma-tag "Y, not X" ("sanctioned, not stolen"); the "not so much X as Y" form; and the gapped version where a stranded verb delivers the pivot ("Hours aren't the bottleneck. Attention is."). One reversal at a genuine turning point is good writing. The structure is not the problem. The density is.
Detection is a whole-document pass, not a per-paragraph glance. Read the entire piece and mark every sentence or sentence pair that negates one thing to elevate another, including the comma-tag and gapped variants above. Count the marks. More than one per ~300 words, or more than three in a short piece, means the reversal has become the default sentence engine, which is the machine tell. A single dense paragraph with two stacked reversals also counts.
Fix by thinning, not by deleting all of them. Keep the two or three that land on the strongest turns. Rewrite the rest as plain declaratives that state the point with no foil ("The bottleneck is attention now." instead of "Hours aren't the bottleneck. Attention is."). Removing every instance flattens the voice, so the aim is to make the survivors rare enough to regain their force.
6. *Repeated hedge adverbs.* A softener like "almost", "somewhat", "rather", "a bit", or "fairly" used more than once in close range becomes a tic. Keep at most one, and only where it earns its place.
## Positive Rules
- Active voice. Use "we" and "our".
- Concrete nouns and verbs. "The model overfits after 50 epochs" not "exhibits suboptimal generalization characteristics."
- Plain English. Use technical terms only when they carry meaning plain English cannot.
- State consequences, not meta-commentary. "The policy has no lookahead" not "training compresses multi-period consequences into a single-step mapping."
- One sentence that advances to the next thought beats two sentences restating the current thought.
- State assumptions when uncertain. Do not hedge-stack ("it might be the case that perhaps...").
- Not every paragraph needs a topic sentence or a concluding sentence.
- Do not resolve the ending with a tidy bow. A piece may close on an open question, an admission, or an unresolved tension. Summary endings that restate the thesis read as machine-generated.
- Do not over-smooth. Removing every short sentence, every parallel, and every fragment flattens prose into uniform medium-length sentences, which is itself an LLM smell. Fixing a tell should not cost the voice.
- When editing existing text, match the density and register of surrounding paragraphs.
- Direct and conversational register, but no contractions in formal writing. Personal essays and conversational pieces keep their contractions; the no-contraction rule applies to formal and technical writing only.
"Turn this outline/lose idea for an article/4 paragraphs of text into a blog post similar to these previous blog posts, but make sure that this one has a table of contents and a bunch of references"
Swiss as racists. Amazing. People know Americans harbor racist feelings because they are surrounded by people of many races. But it's trivial to demonstrate racism among any population as soon as you introduce an "other" of virtually any type.
It wasn’t that long ago that American racists debated whether Italians and even Irish were truly “white.” The definition of white had expanded considerably over the years.
Eastern Europeans, Jews (of course), and Russians were also at times not “white.”
Maybe we should just keep expanding it. Declare everyone white. Black people are just white people with more melanin.
Then we can be racist about aliens from outer space.
There is a distinction between racism and xenophobia, no need to assign the same label to everything. i.e. the thing about Italians was cultural an educated immigrant from Northern Italy would have been considered as white as a French (not that there a significant number of those in the US) at least.
Technically yes, but usually I think there’s a racist undercurrent to xenophobia.
The US stopped accepting refugees recently… except white South Africans. I’d say these people share no more culture or values with the average American than a Central American refugee. Maybe less. I’d much rather party with a bunch of Central or South Americans than a bunch of Apartheid lost causers.
If ethnically white people were pouring across the US border I don’t think many of our immigration hawks would care much, even if some were committing crimes.
> It wasn’t that long ago that American racists debated whether Italians and even Irish were truly “white.” The definition of white had expanded considerably over the years.
Eastern Europeans, Jews (of course), and Russians were also at times not “white.”
This idea is mostly a modern fabrication. Various more granular ethnic biases were of course present throughout American history, but those were never conflated with racial categories: in times and places where the white vs. black racial division was relevant, the ethnic groups you're referencing were always considered "white".
And the types of discrimination that people in white ethnic groups sometimes experienced was of of a type and of a degree vastly different from that experienced by black people. They're really two very distinct phenomena, and weren't evenly distributed throughout the US -- black people in the South had the legacy of slavery, Jim Crow segregation, and other horrifying things to deal with, whereas white immigrant communities in the Northeast or Midwest never experienced anything remotely similar.
Yeah it's much better to make your money by enabling the worlds worst dictators to steal money from their populations. In fact it's all the other countries taxes that are paying for the swiss social security, because of all the aid money being funneled into swiss bank accounts.
Women want their own income stream because of the innumerable ways men get into trouble. If her man gets into trouble, she wants a plan B, for her and her children. I don't think anyone was thinking about how that would prop up the housing market 30 years later.
That's a nice story, but the truth is when women first entered the workforce in meaningful numbers, it was primarily unmarried women. Society strongly expected women to leave the labor force once married, and businesses actively banned married women from working
I have already forwarded this to a number of senior pathologists and industry leaders. People who are in charge of either developing these things or validating them for clinical use. We'll see what they say. But I suspect a great many of them will actually continue using the implicated products because they passed clinical validation at many institutions. TBD.
Google apps are the most downloaded apps in the Apple App Store already. This reminds me of the original Apple Maps, which was just a front end for Google Maps.
The flesh of the strawberry is very slightly translucent. By taking images from many angles, you can derive some information about the interior. This is not a separate step, it is simply the optimizer finding splats that contribute the right amount of color and light at various viewing angles. Of course those splats aren’t really going to look right to anyone who has ever eaten a strawberry, because they are only derived from the input images and not your past memories of a real strawberry.
You’ll also see the same type of problems in any incomplete part of any scene. Go poke around any of them and you’ll find places where there wasn’t enough information to make a satisfying reconstruction. For example, the back of the moon <https://superspl.at/scene/2ac8f423>, the fridge in the coffee shop <https://superspl.at/scene/6a0c3ccf>, underneath the footbridge in the forest <https://superspl.at/scene/23ebe85c> (or indeed anything in the distance), etc, etc.
The "enemy of my enemy" concept suggests that even if the people hate their government, their immediate pain is being caused by the United States and Israel, so I'm less confident about that.
> Iran does not think they are dealing with a madman.
Iran does think they are dealing with a mad man, or at least a government practicing a policy (as the US administration's apologists have termed it) "intentional volatility".
A far more interesting issue here is the oil supplies available in the Pacific. Brunei, Malaysia, Vietnam, Indonesia, Australia, and others are all ramping up production capacity. Non-OPEC oil production is increasing generally in response. This is likely to undermine the Middle East's ability recovery from current constraints as non-OPEC players gain clout in the markets.
Right now people are talking about China and California have limited supplies. But those are enormous, powerful entities that are deploying multi-pronged strategies to secure energy resources. Look at what they're doing and bet there. You also see developing countries retooling to support less oil-intensive economies, like increasing work-from-home options. Solar and wind are currently feeling weak without their subsidies but are exhibiting staying power as people look to move off more petroleum-dependent energy resources.
As for the tactical issue, the concept people seem to be trying to get at is "cost-per-kill". That needs to come down. Yes, we can kill drones with supersonic interceptors. But spending $6M to shoot down a $6K drone has terrible long-term economics.
> Iran does think they are dealing with a mad man, or at least a government practicing a policy (as the US administration's apologists have termed it) "intentional volatility".
We're going to agree to disagree. I know this is what "people" are saying about the US. But it's not what Iran thinks and it's not what the US is actually doing. This is what Iran wants you to think, as it weakens the US, and what it's going to say. Are you saying that the US will go to war with Iran if all the demands I listed were fully and transparently met? A by the way there is that Europe and Canada (e.g.) also don't think the US administration is "mad". Everyone is playing their little geopolitical and local political games.
I also doubt Iranians think their immediate pains are caused by the US and Israel. Some might but most don't.
I agree with you the energy crisis aspect is overblown (I think that's what you're saying). Supply increases in other places and alternative power sources can displace some usage- certainly over time. The other thing that's going to happen are more strait bypassing pipelines.
EDIT: So the problem isn't mad people or rationality. The problem now, as before, is simply that the Iranian regime is religiously and ideologically unable to give in. Giving in will likely result in their fall even if they were able to give in. This is what's driving the main dynamics here. It's not Iranian negotiation tactics or the US supposed not negotiating in good faith or being "mad". The "mad man" are those that believe that Iran is interested in giving in on its exporting the revolution and the destruction of Israel.
> But it's not what Iran thinks and it's not what the US is actually doing.
I think you need to provide some evidence for your claim. The US had a deal with Iran. A madman ripped up that deal, started a war with a decapitation strike, and is now attempting to negotiate a deal we already had before we spent billions of dollars killing school kids. The “People” you dismiss includes scholars, strategists, experts on international relations.
You could possibly explain trumps behavior as rational if you believe he is trying to avoid getting arrested for pedophilia, but that doesn’t build trust. In any case, the issue of competence comes up. Even if you could trust the person who renamed the Defense department to the War department, that person simply isn’t competent.
Many including Trump have long said the deal was a terrible deal. You can disagree with that (and you'd be wrong) but I'm not sure how we get from that to your statements.
Enough evidence? What sort of evidence are you looking for? Can you provide evidence for your claims?
EDIT: Also can you prove that we are looking to get the "same deal" we used to have?
The JCPOA was set to expire on 18 October 2025 after which Iran would not have any limits on pursuing their nuclear program. Are you suggesting the US is seeking a deal now that Iran would pause their nuclear program until 2025? https://en.wikipedia.org/wiki/Iran_nuclear_deal#Expiration
EDIT2: The JCPOA:
- Kept the Iranian regime in power with massive capital influx resulting in horrendous human rights abuse and 10's of thousands of deaths.
- Was being violate by the Iranians. Iran had nuclear sites at Turquzabad, Varamin, and Marivan, which they hid from the IAEA (something that was discovered after Israel stole documents about the Iranian nuclear program). Iran hasn't declared those sites and generally refused access to them for years after the fact. When the sites were eventually inspected years later (in 2020) there was evidence of undeclared nuclear material. https://news.un.org/en/story/2025/06/1164291#:~:text=Iran%20...
- Was time bound and didn't address many other issues.
- Trump said he would withdraw from the agreement. That was his election promise. Trump also said on multiple occasions (and in fact it had been US policy forever) that Iran would never be allowed to have nuclear weapons.
Any rational person adding would agree that the US attack on Iran is in line with its long standing policy. They would also agree that Iran had no other reason for the amount of highly enriched Uranium they amassed other than the manufacturing of nuclear weapons. So I'm not seeing the irrationality here. Ofcourse if your position is that Iran should have nuclear weapons, should oppress their people, and should use proxies to attack others then from your perspective this is an unwelcome development. It's still rational though.
If reasons are given, and it's an election promise, and you win the election, and you deliver on your promise... Most people who say this is good and how democracy should work. I understand that you disagree with the decision and disagree with the reasons but something doing something you disagree with is not the definition of mad.
You seem to have missed the reference to Madman Theory above, interpreting it as a literal commentary on someone's / some group's sanity.
Whether or not actual mental deficiency is involved here is irrelevant; the strategy is the same whether performed intentionally or otherwise. Unfortunately, its track record is dismal in both cases.
Yes I interpreted the parent as literally suggesting the US is indeed irrational not simply trying to appear as irrational to gain leverage. I think this perception is very common.
Is Trump really trying to appear to the Iranians as being irrational here as a negotiation strategy? Is he saying "don't mess with me I'm crazy"? I would counter that he is aggressive and rational. He is saying I am willing to use force and he uses force. His use of force feels fairly rational within the stated goals of forcing Iran to give up their nuclear weapon program (as a minimum short term objective) with maximalists goals of either replacing the regime or removing their offensive capabilities.
If this is irrational, what does a rational use of force look like? Do we agree that without the use of force Iran wouldn't give up their nuclear weapons program? What is true is that the use of force has consequences and of course it would be better to accomplish your objectives without the use of force. Why can't we all get along sort of thing. But Trump (and not just him, many people around him) believe there is no path that doesn't involve the use of force. They tried a bit before and they seem to be trying now before using even more force. I happen to agree there was no path that doesn't involve the use of force and likely more force is going to be needed. The Iranian "we can't agree because we don't trust the US" is an obvious lie aimed at those who oppose US policy (both in the US and out of the US".
All that said, there is still a question of whether the US is willing to use enough force over a long enough duration to accomplish its goals. If the end result is force use, price paid, and a nuclear Iran with long range ballistic missiles, then it's a pretty bad outcome. This is the outcome many are pushing for.
EDIT: In terms of the Iranian lies. If they are saying we can't agree to "suspend our nuclear program for 10 years" because the US and Israel will promise not to attack us and then they will - they are absolutely right. The US and Israel will attack them. That's not a "trust" or rationality issue. We know the Iranians won't keep this agreement because they haven't kept any agreement. We know they won't allow the IAEA random inspections of any site and even if they did they're capable of hiding their activities. The problem is they are not sincere. If they were sincere they would offer and keep the conditions that ensure that they will not be attacked. If they actually suspend their nuclear program (doubtful) and instead build 100k conventional ballistic missiles that's a threat to Israel that is equal to the nuclear threat. And so ofcourse they would be attacked. For them not be attacked they need to actually give up on the religious/ideological ideas and they can't. You might say this is a very opinionated take but it's backed by evidence and with what Iran says on these topics.
You're still missing the point; the strategy only works if no one is sure it's a strategy.
It also doesn't work very well then either.
FWIW, I'm not arguing for either. I think it's pretty obvious this is 100% market manipulation and consequences be damned. In that light, this approach is working very well, just not towards an end that benefits most of us.
It was not only the FDA that stopped snake oil. A lot of it was also stopped by state licensure and medical school accreditation (see the Flexner Report), all of which happened in roughly the same era.
What I take from this is that AI code, biology, etc, will not announce itself. We will have to announce the human-produced content. Introductions are about to be ritualistic again. Queue humorous (but not really all that valid) counterpoint: https://www.youtube.com/watch?v=hoe24aSvLtw
reply