> EpiFusion uses particle filtering to simulate epidemic trajectories over time and weight their likelihood according to both case incidence data and a phylogenetic tree using separate observation models, resulting in the inference of trajectories in agreement with both sets of data.
> This content is from 2014-15; perhaps they AI-rehashed the Yahoo! News story about them?
I went into the article with the HN submissions’s “(2016)” in my head and was shocked that it kept triggering my AI detector senses. It even checks all of the boxes for the phrases like “One honest footnote”. Had me questioning my ability to detect AI writing.
Then I scrolled up and saw that the article linked to was published in 2026. So yes, I’d agree this is an AI rehash and republish of an old story. An interesting story, but the older articles you linked to are so much better to read and more informative than this AI rehash.
"Every published piece is editorially directed and reviewed before it goes live.
We use AI tools in research, drafting, and editing. Editors decide what claims, framing, and pieces make it to publish."
Aurornis 13 hours ago [-]
That’s a weasely way of saying it’s AI written but a human kind of looked at it.
nullpoint420 10 hours ago [-]
Kinda like code reviews these days
tclancy 15 hours ago [-]
Thank you so much for this. Turned an article I was annoyed about (for AI speak) into an amazing story again.
altairprime 14 hours ago [-]
I may not be able to write a proper post for the blue here at the orange, but I’m glad to see it’s enjoyed all the same :)
DivingForGold 14 hours ago [-]
Rhizobacteria are not known to colonize the roots of cereal crops, typically only only legumes. You would think that scientists would spend more effort trying to create "super" rhizobacteria that can colonize even more aggressively and also fix many times more nitrogen than ordinary ones. I cannot say this one below has independent reviews by well published agroscientists, but it's a step in the right direction:
Do a Google search: Methods for Breeding "Super Strains"
Chemical Mutagenesis: Scientists expose the rhizobacteria to alkylating agents like Ethyl methanesulfonate (EMS) to force point mutations.
Physical Radiation: Exposing bacterial cultures to UV light or gamma rays breaks DNA strands, inducing random repair variations.
Experimental Evolution: Google it.
pbmonster 14 hours ago [-]
> You would think that scientists would spend more effort trying to create "super" rhizobacteria that can colonize even more aggressively and also fix many times more nitrogen than ordinary ones
I think the problem with that idea is that the rhizobacteria needs energy to break the nitrogen tripple bond. A lot of energy. It gets that energy by getting carbohydrates from the host plant - which in turn are not available to the plant to grow.
I'm sure there's plants where not fertilizing with nitrogen at all would make the trade-off worth it, but the vast majority of our optimized crops can't spare double-digit percentage points of their total yield to feed bacteria. You're much better of just using natural gas (or green hydrogen) to fix the nitrogen.
IAmBroom 12 hours ago [-]
Or cover crops. They require a season of no yield, but are widely used after harvest, betting on a long pre-frost growing season.
Modified3019 8 hours ago [-]
There are also “free living” bacteria that are capable of nitrogen fixing (diazotrophic), but are not as strictly root dwelling like Rhizobia. https://en.wikipedia.org/wiki/Paenibacillus
Compared to the more well known legume colonizing Rhizobia bacteria, they can host on a wide variety of plant families. Our Paenibacillus Azotofixans can be applied to soil even before a crop is planted and does not need to be directly inoculate seed (though you can do so).
They also don’t seem to have their nitrogen fixing ability suppressed by nitrogen presence, so you can apply both them and fertilizer.
I’m told that under ideal conditions, their presence can behave like adding up to 0.5 pounds of nitrogen an acre per day. While they can’t certainly can’t replace fertilizer applications, that is an impressive amount, especially since it’s a constant source rather than whenever a grower can get around to putting something out.
Tade0 18 hours ago [-]
Using nitrogen-fixing bacteria is not a new concept and it's the life's work of Dr Mariangela Hungria, who recently received the 2025 World Food Prize for her achievements in the field:
The new key detail of this story is using them for cereal crops. The piece is about discovering that they can be used on some crops that were previously not thought to work with them.
chrisgarand 14 hours ago [-]
There are some plants with symbiotic relationships with nitrogen fixing bacterial.
"Hungria" cant have a better name for food research.
Nitrogen fixing is one of the more frustrating things to work on as there is such a massive ready supply of unfixed nitrogen to work with everywhere on our planet and all in all we need such a tiny proportion to dramiticly improve agriculture.The whole thing is so finly balanced that herbivors can likely smell the diference in plants that have been given nitrogen fertiliser and raid our crops and gardens for the fixed nitrogen that then facilitates there own internal bacteria and thereby achive better nutrition for themselves. But to my frustration and anger, nothing eats alder bushes, which are a never ending problem in taking over fields, but when after employing machinery to uproot them,( it bieng the only way to kill them), reveals there bright orange knobly roots, that host the nitrogen fixing bacteria that they contain.
There is enough nitrogen and other nutrients in alder that the area's where they are piled and burnt will remain significantly more fertile for many years after.
One of my(many) projects is to experiment with "terra preta" and build a charcoal kiln, primarily for alder and hopefully be able to retain a greater portion of the
"good stuff™" that is otherwise lost to rotting or combustion.
> There is enough nitrogen and other nutrients in alder that the area's where they are piled and burnt will remain significantly more fertile for many years after.
My experience with burn areas is that the potash alone will do this. Are you sure that the nitrogen is what's surviving? It could be - sodium nitrate has a fairly high combustion temp of ~300C - but even then it's being boosted by the concentrated potassium left over.
metalman 10 hours ago [-]
That gives me a target temp for building a regenerative bio char kiln * , Thanks!
* in a regenerative kiln, volotiles are captured and fed to a burner to provide the bulk of the heat needed to char organic materials, the chared materials can do double duty as soil amendments and long term carbon storage or as fuel for a variety of uses.
Also the volotiles can be condenced as liquids that have a wide variety of other uses.
What would be interesting is if there might be a way to use an enhanced combustion process to capture extra atmospheric nitrogen in the bio char materials.
Ecstatify 15 hours ago [-]
The Stripe Young Scientist & Technology Exhibition (formerly the BT Young Scientist) is Ireland's biggest school science competition, held every year in Dublin.
The winners are often kids whose parents happen to be experts in some domain.
There are certainly some genuinely bright kids in the mix, but it feels a bit unfair on those who don't have parents "helping" them.
Leaving aside the issues with this version of the story, it does demonstrate the value of teaching scientific method properly. Its very easy to teach the discoveries of science but neglect the method (hint for British people, get your kids to do GCSE astronomy, its better than the usual trio for this).
It also shows the importance of questioning expert consensus. Its when you find the current consensus is wrong that you find something new.
GaryBluto 17 hours ago [-]
Is the letter 'g' designed to appear as if the bottom loop is missing, or is that a rendering error on my end?
Haha, feels indeed like the bottom of the G is overlapped by something. A bit weird.
birdsongs 15 hours ago [-]
Oh I already posted elsewhere here about this! Hah. I absolutely hate the font.
fuzzfactor 14 hours ago [-]
Looks defective even under ideal conditions :\
badc0ffee 8 hours ago [-]
Not only is the g terrible, but the lower case a has these thin lines where the circular part meets the vertical bar, that are at odds with the way other letters look.
gignico 11 hours ago [-]
Unbelievably I read this news while visiting Cork. Went to Kinsale yesterday. Never been to Ireland before. What a coincidence!
pluc 16 hours ago [-]
Refreshing to see there are other outcomes than asking AI what it is these days.
RetroTechie 15 hours ago [-]
> Refreshing to see there are other outcomes than asking AI what it is these days.
In science, that's the only way to obtain results: do experiments, gather data.
It's also what sets scientists apart from other people: curiosity. "Hey that's odd, I wonder...". Non-scientists (read: most people) just care about quick answers & how to apply knowledge to their situation.
Note that knowledge in a field, intelligence, 'professional qualification' or whatever, is not essential. These can all be learned or 'sourced' from 3rd parties. But (imho) the curiosity part is essential to make a good scientist.
AI helps to trawl through the mountain of data & knowledge already out there. Re-hash & make it more easily accessible. Discover connections (like proving a math conjecture) that were overlooked before. Which is a good thing. But it can't produce raw data itself. Unless given control of equipment to run experiments on its own. Which might happen btw - if not done already.
The fun thing about research like this is that biology is a 'dirty' science (pun intented): results obtained in experiment A may differ from results in experiment B, due to differences in the local environment: presence/lack of other nutrients, bacteria occuring in A but rare/absent in B, specific varieties of crop used, pollinators, soil pH, etc (yes even local climate).
It's fine if people use AI to lower the bar to entry & make sense of things they don't (yet) understand. But it's sad if people think AI has all answers when it does not, and cannot possibly ever do so. Or even hallucinates at times. Rely on AI as The One Oracle(tm) at your own peril.
I thought this was going to be an eats, shoots, and leaves style discussion.
A wart-covered pea plant is what they refer to (near Cork, Ireland)
daRealDodo 16 hours ago [-]
I hate those g's
wood_spirit 15 hours ago [-]
Can we stop every comment thread being about why is this written by or with the help of AI? Right now as I comment that is pretty much all the comments here are about!!
Isn’t the actual underlying science - which I don’t recall seeing on HN before - more interesting?
I would like to know how it turned out - does it work in practice, is it practical to deploy, is it already or soon going to make any impact on real farming?
b112 15 hours ago [-]
I think the problem is, when someone rewrites a 12 year old story for clicks, it's a little deceptive, AI or not. And AI makes it even easier to be deceptive like this.
spiderfarmer 18 hours ago [-]
Cool a science blog! But it's a rehashed article, followed by an infinite amount of random Taboola scams, arbitrage posts, clickbait articles with jarring content.
Why does shit like that work better than quality contextual advertising?
classified 16 hours ago [-]
Agreed, it's an article with an interesting topic for once, and it's completely ruined by the way it's presented.
mrguyorama 10 hours ago [-]
Why did our fellow HN readers upvote this story at all?
stuaxo 16 hours ago [-]
". Émer, then in her early teens, did the thing that separates a science story from a compost story: " this reads so AI / LLMish.. it's like the same feeling if you had to eat the same food again and again, not quite sickness but something about the repetition.
15 hours ago [-]
14 hours ago [-]
telesilla 17 hours ago [-]
Definitely AI written, though it got the point across and it wasn't terrible to read and it was new information to me so I was glad to read it.
At this point what news site isn't AI written? Except for the few remaining human blogs (which are a delight and a pleasure). Why I detected AI: it said "The honest footnote.." at the end. That's how Claude always ends our long-rambling discussions.
Wowfunhappy 14 hours ago [-]
> At this point what news site isn't AI written?
The New York Times, The Wall Street Journal, The Guardian, The Atlantic, NPR, BBC, The Verge, The Economist. That's just off the top of my head. Basically none of the reputable papers will touch AI.
(I guess I have to add an "as far as we know" footnote but I'd be truly shocked if The New York Times was secretly using AI.)
tuukkah 13 hours ago [-]
The New York Times is publicly using generative AI, but responsibly: "We should tell readers how our work was created and, if we make substantial use of generative A.I., explain how we mitigate risks, such as bias or inaccuracy, with human oversight." https://www.nytco.com/press/principles-for-using-generative-...
alephnerd 11 hours ago [-]
The Beeb [0], WSJ [1], Guardian [2], NPR [3], The Economist [4][5], and others are using generative AI as well.
The reality is that everyone is using it, and has been using forms of AI and ML for news writing and copywriting for decades. For example, I remember hearing how Reuters used to use statistical models to auto-generate articles from notes taken by journalists 20 years ago that would then be copy edited.
The reality is that Generative AI is here to stay just like the Internet, but there are ways to use it responsibly (also like the Internet).
But they are not writing with AI. That's a big difference!
alephnerd 9 hours ago [-]
The absolutely are copy editing, writing, and translating with AI but keeping a human in the loop to steer, edit, interview sources, and do research.
These were entire divisons that made up the majority of a newsroom. Most members of a newsroom aren't reporters - they're editors, copywriters, translators, graphic designers, narrators, etc.
If the whole point of a good journalist is to provide and synthesize context from sources, validate with experts, and do research then most of those other roles are tangential and can be reduced with human gating.
This is how newsrooms and other similar organizations are approaching and leveraging AI now.
Aurornis 13 hours ago [-]
> At this point what news site isn't AI written?
The mainstream ones that pay journalists, for the most part.
These little internet blogs love AI because they’re usually content farms. They want maximum traffic for minimum effort invested.
albumen 17 hours ago [-]
Lots of news sites aren't AI written; or if AI is involved, it's not as blatant as in this article. The 'papers of record' largely remain AI-free. You increasingly have to pay for good journalism.
14 hours ago [-]
vintermann 17 hours ago [-]
That's a qreat discovery by these three qirls. But what IS it with websites that use a font with utterly weird g's? Do they think they need it to stand apart?
retsibsi 16 hours ago [-]
Studies show that your concentration remains sharper if your visual processing system is jarred out of its complacency at frequent, semi-random intervals, preventing you from lapsing into a low-effort 'skim-reading' mindset. It's something to do with the predictive processing model; over time, your brain comes to rely increasingly on top-down processing if its expectations are consistently met, and I guess there is some interplay between visual processing and semantic comprehension in this context.
(Just kidding! I think it is simply an annoying affectation.)
b112 14 hours ago [-]
I had a boss who tried to claim that the "creative process" was helped by him interrupting my coding many, many times per day. With a loud, jovial greeting.
This was year 2000, when you had to write all code yourself, read man pages for info, without Stack Overflow or whatever to guide you. You were more isolated in such tasks, often thinking deeply, especially as a sole dev in a company.
Point is, your joke just brought all that trama back. I'm a peaceful man, but I almost hit him a few times. Major stress. He sort of stopped it when one time, I think I literally has rageful murder in my eyes. He took a step back. and never did it again.
A bendiction... may no other coder experience this horror.
intrasight 15 hours ago [-]
I don't see any of that - but I use reader mode.
Hugsbox 16 hours ago [-]
For real, I thought it was just me but I find the unfinished g's really distracting, even though I understand it's a stylistic choice my brain keeps thinking it's being rendered funky
14 hours ago [-]
akhil_findincal 17 hours ago [-]
[flagged]
18 hours ago [-]
rmason 19 hours ago [-]
For years the holy grail has been for corn to produce its own nitrogen. This could be a small step in the right direction. Interested to see what happens when some of the great American agricultural schools further this experiment.
dofm 18 hours ago [-]
Putting aside the cultural imperialism (this happened in Ireland, which is in Europe, which is also where two of the greatest agricultural schools in the world, if not also the two greatest for this, are based), it also happened over a decade ago, so I guess everyone has tried furthering it by now.
rmason 6 hours ago [-]
I read the article and know the discovery was made in Ireland. This is how science is supposed to work. Is it 'cultural imperialism' if American universities build on the work of these Irish girls? Science would move a lot slower if a countries scientists all had to operate in national silos.
KaiserPro 18 hours ago [-]
I thought all those school/systems of colleges had their budgets cut last year?
blooalien 18 hours ago [-]
> ... what happens when some of the great American agricultural schools further this experiment.
"Brawndo! It's got what plants crave!"
IAmBroom 12 hours ago [-]
> when some of the great American agricultural schools
"Thank God the USA finally sent some scientists to understand things for us!" - the rest of the world. In your head.
https://www.yahoo.com/news/16-old-girls-win-google-science-f...
They posted a video summarizing their efforts in 2014:
https://www.youtube.com/watch?v=vL9N7O0uJoU
Vice did an interview in 2016:
https://www.vice.com/en/article/these-teenagers-science-proj...
Ciara Judge went on to join the research effort to stop Covid in 2021:
https://ciarajudge.github.io/about
And recently earned her Ph.D publishing GPL3 software:
https://github.com/ciarajudge/EpiFusion
> EpiFusion uses particle filtering to simulate epidemic trajectories over time and weight their likelihood according to both case incidence data and a phylogenetic tree using separate observation models, resulting in the inference of trajectories in agreement with both sets of data.
https://doi.org/10.1371/journal.pcbi.1012528
(Submitted URL was https://scienceblog.com/b-three-schoolgirls-in-kinsale-pulle... and I think we'll ban that site.)
I went into the article with the HN submissions’s “(2016)” in my head and was shocked that it kept triggering my AI detector senses. It even checks all of the boxes for the phrases like “One honest footnote”. Had me questioning my ability to detect AI writing.
Then I scrolled up and saw that the article linked to was published in 2026. So yes, I’d agree this is an AI rehash and republish of an old story. An interesting story, but the older articles you linked to are so much better to read and more informative than this AI rehash.
Edit: nevermind, let's just swap the link altogether (more at https://news.ycombinator.com/item?id=49553719).
https://scienceblog.com/about-us/
"Every published piece is editorially directed and reviewed before it goes live. We use AI tools in research, drafting, and editing. Editors decide what claims, framing, and pieces make it to publish."
https://soilscapesolutions.com/products/RhizoGenesis-p668356...
Do a Google search: Methods for Breeding "Super Strains"
Chemical Mutagenesis: Scientists expose the rhizobacteria to alkylating agents like Ethyl methanesulfonate (EMS) to force point mutations.
Physical Radiation: Exposing bacterial cultures to UV light or gamma rays breaks DNA strands, inducing random repair variations.
Experimental Evolution: Google it.
I think the problem with that idea is that the rhizobacteria needs energy to break the nitrogen tripple bond. A lot of energy. It gets that energy by getting carbohydrates from the host plant - which in turn are not available to the plant to grow.
I'm sure there's plants where not fertilizing with nitrogen at all would make the trade-off worth it, but the vast majority of our optimized crops can't spare double-digit percentage points of their total yield to feed bacteria. You're much better of just using natural gas (or green hydrogen) to fix the nitrogen.
My company, Wilbur-Ellis, sells “Nutrio N-Tune” (https://www.cdms.net/ldat/ldJI1000.pdf) which contains spores of Paenibacillus Azotofixans.
Compared to the more well known legume colonizing Rhizobia bacteria, they can host on a wide variety of plant families. Our Paenibacillus Azotofixans can be applied to soil even before a crop is planted and does not need to be directly inoculate seed (though you can do so).
They also don’t seem to have their nitrogen fixing ability suppressed by nitrogen presence, so you can apply both them and fertilizer.
I’m told that under ideal conditions, their presence can behave like adding up to 0.5 pounds of nitrogen an acre per day. While they can’t certainly can’t replace fertilizer applications, that is an impressive amount, especially since it’s a constant source rather than whenever a grower can get around to putting something out.
https://www.worldfoodprize.org/en/laureates/20202026_laureat...
I read that they're trying to cross breed modern corn with one from Oaxaca that has this relationship: https://news.wisc.edu/corn-that-acquires-its-own-nitrogen-id...
My experience with burn areas is that the potash alone will do this. Are you sure that the nitrogen is what's surviving? It could be - sodium nitrate has a fairly high combustion temp of ~300C - but even then it's being boosted by the concentrated potassium left over.
* in a regenerative kiln, volotiles are captured and fed to a burner to provide the bulk of the heat needed to char organic materials, the chared materials can do double duty as soil amendments and long term carbon storage or as fuel for a variety of uses. Also the volotiles can be condenced as liquids that have a wide variety of other uses. What would be interesting is if there might be a way to use an enhanced combustion process to capture extra atmospheric nitrogen in the bio char materials.
The winners are often kids whose parents happen to be experts in some domain.
There are certainly some genuinely bright kids in the mix, but it feels a bit unfair on those who don't have parents "helping" them.
ref past winners:
https://en.wikipedia.org/wiki/Young_Scientist_and_Technology...
It also shows the importance of questioning expert consensus. Its when you find the current consensus is wrong that you find something new.
https://fonts.google.com/specimen/Syne
In science, that's the only way to obtain results: do experiments, gather data.
It's also what sets scientists apart from other people: curiosity. "Hey that's odd, I wonder...". Non-scientists (read: most people) just care about quick answers & how to apply knowledge to their situation.
Note that knowledge in a field, intelligence, 'professional qualification' or whatever, is not essential. These can all be learned or 'sourced' from 3rd parties. But (imho) the curiosity part is essential to make a good scientist.
AI helps to trawl through the mountain of data & knowledge already out there. Re-hash & make it more easily accessible. Discover connections (like proving a math conjecture) that were overlooked before. Which is a good thing. But it can't produce raw data itself. Unless given control of equipment to run experiments on its own. Which might happen btw - if not done already.
The fun thing about research like this is that biology is a 'dirty' science (pun intented): results obtained in experiment A may differ from results in experiment B, due to differences in the local environment: presence/lack of other nutrients, bacteria occuring in A but rare/absent in B, specific varieties of crop used, pollinators, soil pH, etc (yes even local climate).
It's fine if people use AI to lower the bar to entry & make sense of things they don't (yet) understand. But it's sad if people think AI has all answers when it does not, and cannot possibly ever do so. Or even hallucinates at times. Rely on AI as The One Oracle(tm) at your own peril.
"Truth is not decided by majority vote".
A wart-covered pea plant is what they refer to (near Cork, Ireland)
Isn’t the actual underlying science - which I don’t recall seeing on HN before - more interesting?
I would like to know how it turned out - does it work in practice, is it practical to deploy, is it already or soon going to make any impact on real farming?
Why does shit like that work better than quality contextual advertising?
At this point what news site isn't AI written? Except for the few remaining human blogs (which are a delight and a pleasure). Why I detected AI: it said "The honest footnote.." at the end. That's how Claude always ends our long-rambling discussions.
The New York Times, The Wall Street Journal, The Guardian, The Atlantic, NPR, BBC, The Verge, The Economist. That's just off the top of my head. Basically none of the reputable papers will touch AI.
(I guess I have to add an "as far as we know" footnote but I'd be truly shocked if The New York Times was secretly using AI.)
The reality is that everyone is using it, and has been using forms of AI and ML for news writing and copywriting for decades. For example, I remember hearing how Reuters used to use statistical models to auto-generate articles from notes taken by journalists 20 years ago that would then be copy edited.
The reality is that Generative AI is here to stay just like the Internet, but there are ways to use it responsibly (also like the Internet).
[0] - https://www.bbc.co.uk/aboutthebbc/reports/policies/what-were...
[1] - https://www.wsj.com/news/author/wsj-artificialintelligenceto...
[2] - https://www.theguardian.com/help/insideguardian/2023/jun/16/...
[3] - https://www.npr.org/about-npr/1205385162/special-section-gen...
[4] - https://myaccount.economist.com/s/article/How-we-handle-AI-g...
[5] - https://www.niemanlab.org/2026/05/the-economist-launches-a-d...
These were entire divisons that made up the majority of a newsroom. Most members of a newsroom aren't reporters - they're editors, copywriters, translators, graphic designers, narrators, etc.
If the whole point of a good journalist is to provide and synthesize context from sources, validate with experts, and do research then most of those other roles are tangential and can be reduced with human gating.
This is how newsrooms and other similar organizations are approaching and leveraging AI now.
The mainstream ones that pay journalists, for the most part.
These little internet blogs love AI because they’re usually content farms. They want maximum traffic for minimum effort invested.
(Just kidding! I think it is simply an annoying affectation.)
This was year 2000, when you had to write all code yourself, read man pages for info, without Stack Overflow or whatever to guide you. You were more isolated in such tasks, often thinking deeply, especially as a sole dev in a company.
Point is, your joke just brought all that trama back. I'm a peaceful man, but I almost hit him a few times. Major stress. He sort of stopped it when one time, I think I literally has rageful murder in my eyes. He took a step back. and never did it again.
A bendiction... may no other coder experience this horror.
"Brawndo! It's got what plants crave!"
"Thank God the USA finally sent some scientists to understand things for us!" - the rest of the world. In your head.