> Finally, we use our dataset and LRE-estimating method to build a visualization tool we call an attribute lens. Instead of showing the next token distribution like Logit Lens (nostalgebraist, 2020) the attribute lens shows the object-token distribution at each layer for a given relation. This lets us visualize where and when the LM finishes retrieving knowledge about a specific relation, and can reveal the presence of knowledge about attributes even when that knowledge does not reach the output.
They're just looking at what lights up in the embedding when they feed something in, and whatever lights up is "knowing" about that topic. The function is an approximation they added on top of the model. It's important to not conflate this with the actual weights of the model.
You can't separate the hallucinations from the model -- they exist precisely because of the lossy compression.
Did the author even read the abstract let alone the paper?
> We also found that while risk for the six outcomes in our analysis combined was minimized at 0 g unprocessed red meat intake per day, the 95% uncertainty interval that incorporated between-study heterogeneity was very wide: from 0–200 g d−1. While there is some evidence that eating unprocessed red meat is associated with increased risk of disease incidence and mortality, it is weak and insufficient to make stronger or more conclusive recommendations.
Doesn't sound like it particularly slams years of shoddy research.
> Did the author even read the abstract let alone the paper?
The author has a different purpose in publishing as evidenced in the well crafted headline, as correlated with ratio of commenters here engaging on that headline w/o reading the abstract.
> While there is some evidence that eating unprocessed red meat is associated with increased risk of disease incidence and mortality, it is weak and insufficient to make stronger or more conclusive recommendations.
I did a good bit of research and experimentation on diet. Tons of conflicting studies. I just eat things that don't make me feel bad after eating them and hope for the best.
I wonder if lab-grown meat is also carcinogenic.[1]
Funds directed at educating the public on the amount of protein they actually need instead of high-protein diets would go a long way towards reducing energy consumption and emissions to help the climate. Lab-grown meat seems like the wrong direction when there are numerous good protein sources already.
> Getting people to stop beef in countries where it's not a problem is a waste
It's a problem. Most americans consume way more meat than what they could possibly need and they do it because they've heard they need lots of protein which could come from many other sources besides dead animals.
Again, America is not the only country in the world and is not the country that I was talking about when I said "beef in countries where it's not a problem".
> Most americans consume way more meat than what they could possibly need and they do it because they've heard they need lots of protein
I doubt that's true. Meat has always been rich people's food, especially beef, and people like consuming like rich people.
> which could come from many other sources besides dead animals
Is the problem ecological or moral here? If the dead animals were more efficient than the other sources (which they sometimes are), would it still be a problem?
Most americans are consuming 100+ grams of protein a day (its very likely all of that coming from meat - I don't know anyone who consumes a lot of beans).[1] Average person needs about .36 grams of protein per pound of body weight.[2]
When they are grazing on non-arable land and the alternative is to produce no food at all here.
About proteins, from the article you linked (the Harvard one):
> For a relatively active adult, a daily protein intake to meet the RDA would supply as little as 10% of his or her total daily calories. In comparison, the average American consumes around 16% of his or her daily calories in the form of protein, from both plant and animal sources.
> Based on the totality of the research presented at the summit, Rodriguez estimates that taking in up to twice the RDA of protein "is a safe and good range to aim for." This equates roughly to 15% to 25% of total daily calories, although it could be above or below this range depending on your age, sex, and activity level.
I'm not sure how you come to the conclusion that Americans are eating too much proteins, considering the article is not saying it, or even saying the opposite.
Most americans would not meet suggested exercise levels for "relatively active" even if that was at the 30 min. per day level.
16% is 60% more than 10%. Most americans are consuming meat for at least 2 meals a day, possibly 3 for those who are eating sausage or bacon for breakfast.
The general point is that americans consume way more meat than what they actually need. This has been promoted heavily in the high protein diet fads recently. Educating people on what they actually need for a healthy diet would go a long way and cost a lot less than alternatives like lab-grown meat.
Your article suggested that you could aim for 15-25% and not 10%. 16% is at the bottom of that range. I don't understand what point you're trying to make here.
People eat more meat than they need. Most people would be healthier if they consumed less of it or none and substituted vegetables/beans/nuts for those calories while still getting proper amounts of protein from non-meat sources.
Reducing meat consumption reduces a bunch of other consumption of resources letting people get their calories more efficiently.
> Meat is considered one of the prime factors contributing to the current biodiversity loss crisis.
> the livestock sector is a major stressor on many ecosystems and on the planet as a whole. Globally it is one of the largest sources of greenhouse gases (GHG) and one of the leading causal factors in the loss of biodiversity, and in developed and emerging countries it is perhaps the leading source of water pollution.
> farmers would reduce their land use of feed crops; currently representing about 75% of US land use, and would reduce the use of fertilizer due to the lower land areas and crop yields needed. A transition to a more plant based diet is also projected to improve health, which can lead to reductions in healthcare GHG emissions, currently standing at 8% of US emissions [1]
What's not to understand? Stop eating meat. We don't need lab-grown meat. We don't need meat substitutes.
I don't understand why you're mixing a health claim that you failed to prove with an environmental claim that I agree with. If your point is that using arable land to grow cattle or cattle feed is a waste, then I agree with you. If your point is that we should stop eating meat for health reasons, then I wasn't convinced by what you showed me, and disagree with you. If your point is that we should stop eating meat for moral reasons (as in animal rights), then I disagree with you. If your point is that we should stop eating all meat for environmental reasons, then I disagree with you, as not all meat has the same impact, and I think we should avoid blanket statements like that, especially when they are mostly based on what happens in the USA and not around the world. If your point is that we should stop eating meat that has the most ecological impact, like American or Brazilian beef, then I agree. If your point is not any of those things, then I failed to understand it.
> If cultured protein is going to be even 10 percent of the world’s meat supply by 2030
Why do we need protein from meat? The money to engineer cultivated protein could be directed to educating simple americans on the fact that they're consuming way too much protein already. Most americans consume twice the daily protein they need.[1][2]
Cutting out something you don't actually need rather than inventing lab grown meat is an elegant answer to a complex problem.
It isn't like asking people to be vegan or something extreme like that. Just be aware of what your body actually needs and don't go overboard or be wasteful.
Why would that be some kind of controversial message or idea?
Inventing lab grown meat is a good way to make the problems of scaling the raw inputs for custom grown human organs to be sufficiently cheap that everyone can have them.
Building a factory to make aluminium foil is extremely expensive if that's all you're trying to do: it's a lot cheaper if there's already a global mining industry producing aluminium in many near-finished states.
I loathe the idea that everyone getting replacement organs is some kind of good thing to be shooting for.
I have a condition where organ replacement is fairly common. I very much wish the world invested more effort in keeping people like me actually healthy rather than celebrating the macabre prospect of giving more of us replacement organs.
So this is a problem space I've thought about a fair amount and I have zero sympathy for an argument for engineering organ replacement for everyone.
You seem unaware of the many and myriad reasons organ transplants are performed. Like you get that by the time doctors are considering it, it's because the alternative is they think you're either (1) going to die soon when it becomes necessary or (2) is necessary right now.
Did you know there are people who survive COVID and wind up in kidney failure from the stress on their body? What's your answer to them? Oh right: hope you can get a kidney and then enjoy life on immunosuppressant drugs.
But you know, go tell those dialysis patients on the waiting list that actually they're not that important.
You seem unaware of the many and myriad reasons organ transplants are performed.
I'm not. I'm just skeptical that putting more time, energy and money into headline grabbing "heroics" actually makes people healthier and I am very concerned that it only turns people like me into guinea pigs for people who want some limelight more than they want (people like) me to experience some kind of reasonable quality of life.
> it only turns people like me into guinea pigs for people
Did you also know that you are legally allowed to decline medical procedures? You can even sign yourself out of a hospital AMA ("Against Medical Advice") if you don't like what's happening.
You are also completely free, and generally advised, to seek alternate medical opinions.
The existence of a medical procedure or option has not, and never does, obligate you to take it.
The absence of alternatives is my concern. If we optimize for better organ replacement instead of optimizing for how to help people keep their existing organs functional, it's really an asinine thing to pretend they have some kind of meaningful choice.
"Oh, well, now that we've let your organs decline this far, you can get a transplant and maybe live. Or you are free to decline it and almost certainly die." is not a meaningful choice.
I've gotten ridiculous amounts of flak for making real choices about my health. Much of the world would like me to know I'm evil incarnate for doing prosaic things like eating better as a first line of defense.
I guess you could say this, but given that the planet is essentially burning it seems like we're getting way ahead of ourselves with this type of "solution" when we could be working on more important problems.
Plenty of brown rice, soy, nuts, beans etc for people to stop eating meat altogether if they actually needed more protein (they probably don't). ¯\_(ツ)_/¯
Maybe we should direct funds towards advertising meat-free and the benefits of not consuming carcinogenic meat instead of all of this meat replacement. [1]
You can eat all the inferior substitutes, I'll stick to grass fed beef every day. There's no replacement for animal protein, it's the healthiest thing you can eat. Meat is well known to increase testosterone levels, promoting all sorts of beneficial effects on your metabolism and mental well being.
I wouldn't say eating meat is the healthiest thing you can eat.
It seems like most people need fiber for a diverse and resilient gut biome.
But the dangers of cholesterol are largely overblown.
Sugar, oxidation, insulin levels, and stress play a far more pervasive roles.
Statins are largely ineffective for preventing heart disease except for a small percentage of older white males.
There's also a low causal relationship between dietary cholesterol and the cholesterol found in your blood.
And most cholesterol in the human body is produced in the liver.
The biomechanics of the body are complex and difficult, but one underrated aspect of being able to produce meat in a lab is you would be able to repurpose that technology to produce human organs for people who need new kidneys, livers, hearts, lungs, etc.
It's a very valuable technology, even if you're not interested in eating it.
The problem with all non-animal sources of protein is their ratio of protein to carb+fat. I eat ~.8gr of protein per lb of body weight per day to maintain my muscle mass. If I had to get that protein from non-animal sources, I'd be way over my total daily calorie intake.
Tablespoon of peanut butter really isn't any more work. It's probably less.
> How many beans equal a serving of meat?
> In general, 1 ounce of meat, poultry or fish, ¼ cup cooked beans, 1 egg, 1 tablespoon of peanut butter, or ½ ounce of nuts or seeds can be considered as 1 ounce-equivalent from the Protein Foods Group.[1]
As a vegan Hindu it makes me sick thinking about making a food have the taste and texture of a dead animal. But I do understand the desire to keep eating what is familiar
The other day I decided to try a beyond burger for a change. I had to throw it away because I could not eat it, because eating something even close to meat was unappetizing.
lol thanks. I read that totally backwards, and after having that burger tonight i concur with the commenter. They're pretty damn good, I'd happily eat another
This was said as a current meat eater, but if I keep down this recent ethical veganism youtube rabbit hole i might not be that much longer
I think it's hard for me to say exactly without a side by side comparison, partially because this one had different toppings than my normal burgers do, but it seems fairly comparable in flavor and I would buy these again. I've also had an impossible burger a few years back at a nicer restaurant and that was good too.
I think both are to the point that if I didn't know what I was eating and didn't deeply try to analyze and compare, I could eat one without any idea that it's not meat.
I'm now somewhat interested the next time I see my family to see if I can pull off a ruse of making burgers for dinner from these and see if anyone even notices, particularly my brother who would say there's no way you could trick him XD
> Finally, we use our dataset and LRE-estimating method to build a visualization tool we call an attribute lens. Instead of showing the next token distribution like Logit Lens (nostalgebraist, 2020) the attribute lens shows the object-token distribution at each layer for a given relation. This lets us visualize where and when the LM finishes retrieving knowledge about a specific relation, and can reveal the presence of knowledge about attributes even when that knowledge does not reach the output.
They're just looking at what lights up in the embedding when they feed something in, and whatever lights up is "knowing" about that topic. The function is an approximation they added on top of the model. It's important to not conflate this with the actual weights of the model.
You can't separate the hallucinations from the model -- they exist precisely because of the lossy compression.