If you could metabolize uranium

Miscellaneous and Personal Stuff I Must Share
1

So I saw elsewhere a post about the number of calories in a gram of uranium (20,000,000,000), and somebody joked that it was a good reason not to eat it. Now obviously the body cannot absorb calories from uranium like we do with food but my friend decided to ask ChatGPT what would happen if we could. It is quite good so I thought I would share. This is its reply.

(yes I know uranium can be absorbed by the body and ends up mainly in the bones if not disposed of as waste)

Alright, let’s dive into this hypothetical scenario where the human body could actually process uranium like it does regular food, and ignoring its toxicity and radioactivity.

Calories and Energy Storage :

The average daily caloric intake for an adult is about 2,000-2,500 calories.

Consuming 20 billion calories would be equivalent to the energy intake of approximately 8 million days (or about 21,918 years) for a single person.

Immediate Effects :

If the body could somehow process and store this massive amount of energy instantly, it would lead to an extreme, almost comical, energy overload.

The body would need to store this energy, primarily as fat, since it’s the most energy-dense storage method.

Weight Gain :

1 pound of body fat is roughly equivalent to 3,500 calories.

Consuming 20 billion calories would result in an approximate weight gain of over 5.7 million pounds (20,000,000,000 / 3,500 ≈ 5,714,286 pounds).

Physiological Impact :

Such an enormous, instantaneous weight gain would be physically impossible for the body to handle. The sheer volume and mass would crush bones, organs, and lead to immediate fatality due to the body’s inability to cope with such extreme changes.

In a more controlled, theoretical sense, if the body could gradually process and store these calories, it would result in extreme obesity and severe health complications like diabetes, cardiovascular diseases, and mobility issues.

Energy Expenditure :

Even with an extremely active lifestyle, burning off 20 billion calories would be impossible within a human lifespan. To put it in perspective, running a marathon burns about 2,600 calories. To burn off 20 billion calories, you’d need to run over 7.7 million marathons.

In conclusion, even if the human body could process uranium and utilize its caloric content, the consequences of ingesting such a massive amount of energy would be catastrophic. The body’s natural limits would be overwhelmed, leading to severe and likely fatal health issues.

2

Is it fruitful to point out what an obvious misunderstanding of basic physics this is?

Stranger

3

Not that this distinction is going to matter given the overall nonsense of the premise, but you’ve conflated thermodynamics calories with dietary calories. 1 dietary calorie is actually 1000 thermodynamics calories, so you need to shave off 3 zeros on your big number.

Dietary calories are formally called kilocalories in physics.

4

I was a bit surprised that ChatGPT didn’t pick up on that. The first several responses were “Please don’t do this.” LOL

5

Who needs physics when you have AI! :wink:

6

LLMs are easily fooled when specific terms mean different things in different contexts. It sees “calories” in physics and dietary contexts but pretty much never at the same time, so it can’t know the difference.

7

The calories in uranium are not released at once when ingested, are they? Uranium decays slowly, so if you managed to life forever it would be an appropiate snack every now and then. Good for interplanetary travel without FTL technology.

8
9

LLMs and other generative AI don’t really ‘know’ anything in the context of the physical environment. Despite the fact that many want to claim that they have extensive models of the real world, what they actually have are extensive models of the text or images that they have been trained on in the form of relations between their data, and which they use to produce similarly structured text or images that conform to the prompt. Hence why they are so easily ‘fool’ or make trivial mistakes in basic comprehension, and why they frequently produce nonsense and fabrications when asked for factual information and citations.

There are no ‘calories’ in uranium in the metabolic sense, which is actually a measure of the chemical energy released in a bomb calorimeter when a sample of food is combusted (and has only a rough correlation to actual energy released during digestion). The energy released by radioactive decay and nuclear fission is, as @gnoitall said, a completely different concept, and there is no way to convert that energy into fat or other elemental mass.

I know that the o.p. (and others, I assume) are aware of this, and that the chatbot response is just posted as a ‘funny’, but people are actually relying on LLMs to provide factual responses and then inserting them into real documents under the assumption that the source is authoritative because is reads well. I’ve recently had to deal with a technical analysis where someone used an LLM to produce cromulent-reading bullshit that was inserted into an analysis that was well-written enough to pass through multiple reviewers who apparently didn’t critique it with sufficient thoroughness to peer through the nonsense technobabble, and if not caught would have very much served as the basis for a critical decision.

Stranger

10

Pay more attention to the text right at the bottom of the ChatGPT page :grin:

11

Keep in mind, I was not the one conversing with ChatGPT in this instance; however, I have had some interesting conversations with ChatGPT just for fun, and it often surprises me when it picks up on little errors and oddities. Not always of course, but it would not have surprised me at all if I had said that for metabolic calories you would need to divide by 1000. I think perhaps the reason it didn’t is because it just doesn’t make any sense at all as uranium cannot be eaten as food (Matter Eater Lad not withstanding). Part of it as well, is that ChatGPT really tries to be helpful barring certain topics. While LLMs are not area of expertise, I suspect this is an intentional part of the design. If you’re making an assistant tool, then you don’t really want it respond “I don’t know about that one” all the time (looking at you Alexa).

But yes, ultimately I posted this just because I thought it was funny. No other conclusions should be taken. :slight_smile:

12

You know what could metabolize Uranium? Andromeda:

Dr. Jeremy Stone : We just found out Andromeda functions like a little reactor. Converts matter to energy, energy to matter, directly

13

Fiddle.
Go-Go Godzilla.
Who feeds on radiation.

14

I mean, everyone is easily fooled by this one. What do you expect, when you give two different units the same name aside from capitalization?