> most supermarket bread is stabilized, which you might not like aesthetically, but it isn't doing anything to your health
That's a large assertion to make without evidence. Personally, I'm suspicious of all antimicrobial agents' effects on gut microbiota and the knock-on effects on health.
I mean, he's not presenting the evidence here in a hacker news comment, but that doesn't mean the evidence doesn't exist. There are a bunch of different categories of stabilizers, and only some of them have antimicrobial properties -- most are about maintaining moisture and loaf volume over time.
There are three substances that are predominantly used to inhibit mold growth, though: calcium propionate, potassium sorbate, and sodium benzoate.
The first two are extremely well understood and considered safe globally. The third is still considered safe, but might have some interactions with vitamin C that lead to benzene formation to the tune of several parts per billion. The FDA considers this safe, and has for some time.
There have been so many (serious) tv-documentaries about this. Our bodies are the mass-experiments. I remember something from about 20 years ago in Germany, mostly about buns, the stuff that is called 'sandwich', and more 'normal' bread.
It began with (then new) biotechnically produced enzymes, who's into that, and where.
Turned out Denmark is big into it, with corporations most of us have never heard of.
So that stuff goes in all the bread to give it desirable properties, texture, taste and longer shelf-life, not even in big amounts. And that's the crux. Because it can be so little, it isn't even mandatory to declare on the label. Nobody knows what that does to our bodies and their metabolisms longterm.
Aroma came from France mostly.
So they retraced the enzy- and aromatic stuff to Denmark and France, popped up in front of a big Polish bakery, and traced how the half/pre baken stuff moves frozen from there into our supermarkets, to be baken up in store, or being bagged, and brought home.
How will attribution to negative side-effects ever be possible? Instead of pointing to other stuff like black carbon, so2, countless VOCs, PFAs, PFOAs, PTFEs, whatever?
You said dislike. I said suspicious. One's an opinion, the other an invitation for science.
> vinegar, salt, and any fermented foods that suppress microbial agents?
In the gut microbiota broad system sequencing experiments I've consumed reports from, even altering diet from one primary food source to another can result in complete restructuring of the gut ecosystem as quickly as within 24 hours. And switching back to the original food source didn't return to an identical gut ecosystem. Many studies show direct connections between gut microbiota and specific and broad health impacts.
So I think suspicious is right. There is much yet to be learned.
"Ultra-Processed Food" is a phrase. It is not a thing. If you want to use it for its media traction, and "nobody knows what it is," define it before you use it.
Instead, everybody has gone for "it is so mysterious," then goes on to cite a bunch of studies about "ultra-processed foods," treating it as a category that categorical statements can be made about.
Name the chemicals, name the processes. They don't because these chemicals and processes are endorsed by massive conglomerates who can pull advertising and lobbying money. When you're vague as hell about everything, it's a case-by-case basis that only affects people who can't bribe as well as others.
And about the people talking about sugars, salt and fats crammed into everything, I'm convinced they never cook. That's taste, and you can pour it in yourself just as well as a corporation. The reason people love it is because it tastes good. Salt isn't even bad for you, it's in fact very, very good for you because it makes micronutrient-rich foods palatable.
Arguing that standard manufactured foods and foods marketed to children are too rich is a different argument, and most importantly for this conversation has nothing to do with the ultraprocessed foods question. Massive food companies are perfectly happy to remove all the fats and sugars from your food and replace them with some weird alcohol sugar or a new form of teflon they just invented a week ago, and to charge you a premium for the privilege of eating it. They literally did that throughout the 80s and 90s.
The narrative intentionally confusing what the above refers to as "hyperpalatability" (a term which is as good as any) with "ultra-processed" (a propaganda term) is a gift to the food industry. It in fact even favors monopolies who can afford to put in the bizarre chemical constructions they'll use as alternatives, and to pay regulators to look the other way.
"Ultra-processed foods" is a phrase meant for advertising - even the prefix "ultra-" reminds me of a cigarette ad, describing the filter. The filter was a device which ostensibly makes cigarettes healthier, but instead allows you to smoke far more than you would otherwise, and leads to worse health outcomes. Isn't smoking bad? Smoke our ultra-filtered cigarettes and breathe easy!
The featured article/podcast has the person presenting define "ultra-processed foods" as industrial hyper-palatable foods.
However, it focuses more about how the foods were marketed, and in particular, that the same marketing playbook developed by tobacco companies to market Marlboro and Camel cigarettes was then used to sell Kool-Aid and Miller beer.
Basically, the execs at Phillip Morris and R. J. Reynolds realized that beer, Kool-Aid/sugar drinks, Oreos and cigarettes were more or less all the same type of products - affordable, habit-forming small luxury goods people enjoyed putting in their mouths, and applied the expertise they'd developed on one type of this market (cigarettes) to the rest of it.
Harnesses can fairly easily be adapted to other models. Moving capabilities from one model to another is a more involved process requiring distillation, training, etc.
I don't quite understand your point. Why are you moving those capabilities from one model to another, or improving the built-in capabilities of a model, what is the goal? If having the capabilities in the model itself improves the overall capability, then using that better model with the same harness should achieve better results. Or if the capabilities are the same, but they've been moved from the harness into the model, then there should be some other detectable improvement, like to speed or cost.
I don't know about others, I can only speak for myself. But I do appreciate numbers for bare models, numbers for model + harness, numbers comparing different models in the same harness, and numbers comparing several models across several harnesses.
It's a lot of information to ingest, but it gives me some idea of which part of the system is doing which part of the work, how well different harnesses and models interoperate, and more insight into the part of the equation under my direct control as a software developer.
I agree that all else equal, the more comparisons the merrier. But IMO, there has been way too much focus on benchmarks for the bare models, when it seems to me that what actually matters is what you can do with the technology, and that is never limited to what a bare model can do.
> As much as it seems beloved here, people that flash custom Android OSs are the very definition of niche users.
Hmm... Let's try reframing this: "as much as it seems beloved here, people that install their own operating systems on PCs are the very definition of niche users"
I'm absolutely certain that's how IBM felt before the clones. But the ability to install what they wanted on a defacto standard platform is what launched the computing revolution. I think we'd still be living in a sterile monopolistic environment with $10k compilers otherwise.
Folks installing their own ROMs on phones are only niche because they've been pushed out at every opportunity using locked bootloaders, embedded security processors, factory installed secret keys, etc.
Despite all that, there's still thriving communities developing and using custom ROMs on their phones. That demonstrates more than niche demand.
> Hmm... Let's try reframing this: "as much as it seems beloved here, people that install their own operating systems on PCs are the very definition of niche users"
I mean yeah, 99% of people never installed an OS and never will, what's your point here?
> 99% of people never installed an OS and never will, what's your point here?
That the 1% who do build visicalc, Linux, the internet, Google, and every application and innovation that happens outside the corporate wall. The entire ecosystem everyone else ends up using.
And that calling that niche is ridiculous, shortsighted, and shooting oneself as a platform owner in the foot.
What are you talking about? You're making zero sense. I've been in the niche of modifying and installing my own ROMs, then to kernels, and now doing the same things commercially.
It is a niche, it always was. There was never a mass community of users, the most installs you used to see on any given AOSP based project is maybe 20-30K for the most popular devices like Redmis, Pocos or Google Nexus phones.
It's still a niche, and things that users do and used to do to their phones are fundamentally incompatible with Android's security model. Root access, magisk, xposed, overall zero or near zero security validation from the security perspective on all ROMs but Graphene (maybe some others, I haven't followed the sphere for a while).
Again, only because the platforms are locked down and there is no single standardized target for universally bootable ROMs in the way a Linux ISO can boot on any PC.
Kinda sounds like you might be younger and didn't live through the 8bit micro -> PC transition. Standardization of the platform led to a cambrian explosion.
> Again, only because the platforms are locked down and there is no single standardized target for universally bootable ROMs in the way a Linux ISO can boot on any PC.
Couple of things here. Android was never that open to begin with. There were always (and still going) issues with the kernel sources being withheld indefinitely, AOSP side changes are almost never released at all, iirc only Google released anything. Bootloaders were always an issue, not all manufacturers had unlockable (or easily unlockable) bootloaders.
Ironically, Google were always the spearhead of the standartization in the Android world. Project Treble and stuff resulted in the fact that almost any Android device released these days can run a generic Android image, it's not gonna be as functional as the original Android, but it's something.
Yes, I am on the younger side, that's why I know how it was in the Android world. Nobody is interested in reinstalling OS-es on their phones because the amount of effort to create an Android ISO that's as usable as what's preinstalled is truly gigantic. Android flavours are heavily customized for the needs of the OEM, and for the hardware that they're gonna run on.
It's a niche, always was and always will be. A niche powered by a few large players like LOS or GOS, that actually get paid somehow or heavily work for the idea, and an army of teenage unpaid maintainers, who quit the maintainership as soon as they get a properly paying job that takes most of their time (me).
> Nobody is interested in reinstalling OS-es on their phones because the amount of effort to create an Android ISO that's as usable as what's preinstalled is truly gigantic.
It's pretty clear that you have an extremely limited exposure to the custom Android development space, since you never rebutted any of the points I brought up. Custom android was never easy and never will be. And it's fundamentally incompatible with the concept of having a truly secure env on your device (same as in the PC space, ironically enough).
Son, I've been building and installing custom roms since the HTC Dream, was involved in OpenMoko, and have owned nearly every open or Linux-running phone of the last 30 years back to the Sharp Zaurus and iPaqs.
Your custom rom experience sits wholly within a larger category of computer systems experience about which I've been speaking.
The sort of fragmentation you describe is a symptom, not a foregone conclusion.
4 bits per 16 bytes is ~ 1/4 the cost of ECC, which is possible to implement with just an extra cycle or two of latency in the memory controller. MTE seems similarly lightweight. Costs some transistors and a percent of a percent of power budget, but much like ECC it seems a fair bargain.
I was one of those affected. The official CDC numbers are underestimates, because they're only counting cases confirmed by culture. My doc asked enough questions to make sure it wasn't your typical stomach flu or food poisoning and got me the antibiotics ASAP which I appreciated. I assume I'm not counted as a result. I have to assume others aren't as well.
This. As an infectious disease epidemiologist, the number of people who will work their way into laboratory confirmed case reports is a small fraction of the total number of people impacted.
You are almost certainly not reflected in those numbers.
To be clear, ITAR (International Traffic in Arms Regulations) restricts certain roles to "US persons". This is a weird term in US immigration/nationality law. A permanent resident (green-card holder) is classified as a "US person". As a result, SpaceX can employ non-US citizens for ITAR regulated roles.
They're also the folks working on the largest co2 -> methane converters in the world. Because that's how they'll need to make fuel on Mars, and no one else on earth wants to fund an expensive technology to replace an inexpensive gas pumped from the ground. By definition, these reactors will have to be capable of producing enough fuel for a rocket in order to be fit for purpose. That opens the door for carbon neutral operation. It's tough to imagine a greener rocket than a fully reusable one that's carbon neutral running on methane. Sure beats spewing hypergolics or solid rocket fuels.
I can think of academic practitioners with doctorates who had wildly disastrous opinions about their field of expertise too:
Peter Duesberg, a Ph.D. in molecular biology / retrovirology who taught at Berkley and was a member of the National Academy, who maintained that HIV does not cause AIDS and that antiretroviral drugs do more harm than good.
Walter Freeman, M.D. academic neurologist and first chair of neurology at George Washington University who believed that severing frontal-lobe connections could stabilize personality and stop pathological cycles of thought.
Charles B. Davenport, Ph.D. in biology, geneticist and founder of the Eugenics Record Office.
Henry H. Goddard, Ph.D. in psychology, intelligence-testing researcher and later professor of abnormal psychology who believed intellectual disability, poverty, prostitution, and criminality constituted a hereditary family type.
Clarence Cook Little, Harvard Ph.D., mammalian geneticist and prominent cancer researcher who insisted for years that the evidence did not establish a causal relationship between smoking and lung cancer.
Fun game. We could laugh at all of them, and your examples too, if not for the damage they caused.
That's a large assertion to make without evidence. Personally, I'm suspicious of all antimicrobial agents' effects on gut microbiota and the knock-on effects on health.
reply