Ask HN: Is synthetic data generation practical outside academia?
4 points by cpard 9h ago 2 comments
Ask HN: Has anybody built search on top of Anna's Archive?
284 points by neonate 3d ago 146 comments
Millions in west don't know they have aggressive fatty liver disease, study says
85 robaato 99 6/5/2025, 5:47:52 PM theguardian.com ↗
“The researchers found that just under 3% of people in the UK, France and Germany, and 4% of those in the US have MASH, but diagnosis rates were below 18%. That means about 20 million people in the US, UK, Germany and France are living with MASH but only 2.5 million people have a diagnosis, leaving more than three-quarters – about 16.7 million people – unaware they have the condition.”
2.5/20 is 12.5%, which is under 18% but a very weird and specifc way to put it; 16.7 is more than 3/4 of 20 indeed (by a lot), but adding 2.5 to 16.7 is about 19 not about 20. This just all seems randomly off in various ways that make little sense to me. Anyone has any good theory how such sentences escape editorial edits, or can find a simple typo or two that make this paragraph coherent again?
I can't find much published research on it to be fair, but I think the science in this field is lagging behind people's personal experiences.
If there's evidence to the contrary let me know, I'm not trying to spread misinformation. It's just one of the things I consistently recall reading over the years.
Edit since I'm being downvoted:
https://pmc.ncbi.nlm.nih.gov/articles/PMC6893587/ (prolonged fasting, ~8 days)
> The improvement of FLI correlated with the number of fasting days (r = −0.20, p < 0.0001)
https://eglj.springeropen.com/articles/10.1186/s43066-021-00... (ADF rat model)
> MSRDF rats showed cure of grade-1 NAFLD and significantly decreased LW than other groups and normalized HOMA-IR, HbA1C TC, LDL-C, ALT, and CRP.
https://www.cell.com/cell-metabolism/fulltext/S1550-4131(22)... (exercise + ADF, humans)
this is the main thing I could find.
https://prolonlife.com/ sells a prepackaged fasting-mimicking diet. plenty of reviews online about the subjective effects on energy levels and soforth during the fast.
I didn’t like it. day 2.5-3 will put me back into the headspace of food scarcity and even knowing that the next meal was sitting in the box and that this is temporary … it was a mental challenge for me. if you’ve never experienced food scarcity, it can be all-consuming and seriously warp your cognition and emotional baseline.
But this is a sample size of 1 and results definetly vary wildly between folks.
The real danger is if you dont swap back and just created yourself an eating disorder.
My understanding is that if you are healthy and you fast, it's great. If you are actually ill and fast it's still great but it only hides your illness and you are on a very bad path (eating disorder)
Yhe issue with anorexia is that it works as cycle - if ypu have genetic predisposition, starwing affects metabolism, your discomfort about food gets worst and you are in it.
Hunger is truly a powerful driver.
I cleaned up my diet about a month ago, and have accidentally done some 24 hour fasts when I was busy and it's been fine. By the time I do eat, I'm really not even hungry, though my stomach may be growling a bit. The first time I ever did this, I had horrible headaches and felt miserable for a while, but subsequent times have been easier.
Good sleep maters too. Bad sleep will throw your hormones out of whack. I'm extremely hungry when this happens, and crave all the wrong things. Knowing what's going on helps a little.... just a little.
I find all this much easier than just trying to cut back by 200 calories with what I normally eat. It's all about hormones.
Long story short: meat and vegs + fruits. It takes a while.
To put some numbers to it:
https://pmc.ncbi.nlm.nih.gov/articles/PMC10564080/
> Only five out of the 1304 studies on NAFLD involved IF.
Here is one that mentions there may be some efficacy to the idea and no harm.
https://pmc.ncbi.nlm.nih.gov/articles/PMC8958240/
> In conclusion, current evidence suggests that intermittent fasting in patients with NAFLD is a feasible, safe, and effective means for weight loss, with significant trends towards improvements in dyslipidemia and NAFLD as illustrated through non‐invasive testing (NIT).
If someone has NAFLD, they can either sit around and eat cake for 20 years waiting for the science, or they can try doing some fasting, which is very low risk (assuming they don't have other issues going on), and find out very quickly if it works for them. Sure, it's an n of 1 in that case, but who cares, if they are the test subject it only matters if it works on them.
I'd add to this that the carbs should be kept low and the diet having quality foods outside of the fasts. Eating aforementioned cake during a feeding window every day is going to leave a person miserable, burning muscle, and still leave the hormones all screwed up. Insulin needs to be controlled and lowered. Fasting does that quickly, but don't abuse it during your meals on a regular basis.
From what I've read elsewhere, fasting can help in the early stages to reverse it, but once real damage occurs that sticks around.
In terms of muscle loss, from what I’ve read, muscle loss is more of an issue for low calorie carbohydrate based programs. When fasting there might be some muscle loss when at the very start, but then it tends to preserve muscle, and like you said, adequate protein intake and resistance training can mitigate that.
Those, like Peter Attia, who saw muscle loss from long term keto and fasting over several years, weren’t using fasting as a medical intervention. He was already metabolically healthy, but maintained a pretty extreme fasting protocol in an effort to gain longevity benefits, effectively experimenting on himself based on some results of early studies around the benefits of fasting for longevity.
The general theory now seems to be making the fasts much less frequent once healthy. Maybe only a couple times per year.
I’m sure this area of study will continue to evolve slowly.
Long calories deficit can lead to permanent brain damage too. And heart damage.
> Available evidence suggests that any form of caloric restriction may be beneficial and specific forms of IF should be tailored to the individual.
Also important to notice that once the liver is damaged it's not recommended to fast:
> Additionally, it is important we investigate the possible risks of fasting in patients with cirrhosis, which is currently not recommended.
I weigh 72kg at 178cm height. At peak, when I got tested I weighed around 86kg. I was barely overweight and definitely not obese and yet...
E.g. I'm 178cm as well but when I was 71kg I was visibly "fat", or at least skinny fat to be charitable. I'm currently 67kg but very lean and somewhat muscular.
But it isnt pushed hard because it is difficult to steer the ship of humanity. Like how doctors will say "lose some weight" but they arent really expecting miracles on it as they know the battle that is.
If someone wants to try limiting meat consumption as an "n=1" experiment to see how it affects their body composition and other biomarkers then go ahead. Just don't expect a major impact from that one factor.
I eat around a pound of beef a day, a gallon of yogurt a week, and almost everything (eg oatmeal) is made with copious amounts of butter.
I think unless you know something specific about your genetics, just eat plain natural foods people have been eating for millennia and you will be perfectly fine. Basically, buy plain meat, vegetables, grains, and dairy and prepare them yourself. Don’t eat preservatives, corn syrups, or novel vegetable oils like canola. Maybe they’re fine but there’s no reason to risk it. Also do your best to make sure what you eat followed the same rule; eg my beef was grass fed and finished and was not fed skittles in a feedlot. Diet affects animal meat as well, just as it does us.
I feel the hfcs is the bigger issue here because it is put in everything in the US, but moderating both of these will help.
Look at labels and put back anything with hfcs.
https://www.nih.gov/news-events/nih-research-matters/how-hig...
https://pmc.ncbi.nlm.nih.gov/articles/PMC6549781/
And actually one of humanity's major evolutionary advances.
It makes sense to try to eliminate it even though it's "only" a small difference. Might as well remove the difference at all and look out for things with no HFCS shoved in it for no reason.
>HFCS 42 is mainly used for processed foods and breakfast cereals, whereas HFCS 55 is used mostly for production of soft drinks.
In other words, the type of HFCS that's "shoved into countless foods" has less fructose than table sugar, not more. If fructose is the villain here, that actually constitutes an improvement over table sugar.
Similarly, it's not suprising when pasta sauce has some sweetness added -- grandma also likely added a bit of sugar if she found tomatoes too acidic, which many do.
The only thing that matters is that it's sugar. HFCS isn't somehow worse. If you're trying to eliminate sugar overall then sure, of course avoid HFCS. But if you're fine with a certain moderate amount of sugar per day, then the relatively small amounts of HFCS in things like pasta sauce and peanut butter are fine. The same way the sugar or honey in teriyaki sauces is. They count towards your daily allotment of sugar. For people trying to eat relatively healthily, avoid the soda but there's no reason to worry about the HFCS in ketchup or normal amounts of tomato sauce, for goodness' sake. The only reason to avoid HFCS entirely is if you're truly cutting sugar out of your diet entirely. Otherwise they're just substitutes for practical nutritional purposes.
These choices add up, which is the point I was trying to make originally (though I agree I did not do a good job of it); I understand I was being pedantic so I understand the nature of the responses to me. The point is that small differences, isolated, don't matter, but in aggregate they absolutely do. We make arguments like this all the time in software when trying to write correct, performative code -- the milliseconds add up, and so do the grams of sugar.
The anti-HFCS movement, despite having its targets aimed for wrong reasons, is still aiming at the right thing: being more mindful of what's in the things we put in our bodies.
As for honey (and fruit, for that matter) - they are full of beneficial compounds that help your body regulate blood sugar.
To illustrate this, someone I know of with type I diabetes who ate natural honey didn't need to inject as much insulin as when eating commercially processed/heated honey - those beneficial compounds are destroyed during the process. Same for fruit vs fruit juice...
Corn syrup might or not be different, Im no expert in that field, but simply having highly refined sugars without the filler is monocropping your diet in weird ways. In the same way soil is being killed by mono cropped agricuture, we are doing the same with our gut biome. The flow on impacts are still being discovered.
The glucose and fructose than act the same as glucose and fructose from HFCS.
> JVL, HEM, and CJK acknowledge institutional support to ISGlobal from grant CEX2023-0001290-S, funded by MCIN/AEI/10.13039/501100011033, and the Generalitat de Catalunya, through the CERCA Programme. CDB is supported in part by the Southampton National Institute for Health and Care Research Biomedical Research Centre (NIHR 203319). Funding statement: This work was supported by Novo Nordisk and Echosens via a grant to ISGlobal. The funding sources had no role in the study design, writing of this manuscript, or decision to submit the paper for publication, but did carry out the data modelling.
https://www.thelancet.com/journals/lanepe/article/PIIS2666-7...
No comments yet
> JVL, HEM, and CJK acknowledge institutional support to ISGlobal from grant CEX2023-0001290-S, funded by MCIN/AEI/10.13039/ 501100011033, and the Generalitat de Catalunya, through the CERCA Programme. CDB is supported in part by the Southampton National Institute for Health and Care Research Biomedical Research Centre (NIHR 203319).
> Funding statement: This work was supported by Novo Nordisk and Echosens via a grant to ISGlobal. The funding sources had no role in the study design, writing of this manuscript, or decision to submit the paper for publication, but did carry out the data modelling.
Fun to see my local government funding something that appeared on HN :)