Hacker Neue

Preferences
wjnc parent
This article could be so much better: How large are the estimated stores of ore that underwent natural fission? How much energy did it release and over how much time? When? Would this be noticable (and to whom)? So many questions, so little information.

I only know (or knew) high school physics, and when entering this in Claude I get an answer but am unable to verify the answer. Claude says 680 kWh gained per 0.03 grams of U-235 lost due to fission. I am left wondering into what the U-235 fizzed into (sorry, pun) and if I should take that into account.

Edit: There we go with modernity. I went to Claude instead of Wikipedia. Wikipedia at least has the answers. Thanks u/b800h. 100kW of heat on average. I can start filling in the blanks now.

adev_
The 'natural reactor' in Oklo has been discovered by some french researcher from the CEA in the 70s.

There is an entire scientific publication on the topic if it interests you:

https://www.sciencedirect.com/science/article/abs/pii/S00167...

the_arun
I wonder why Claude’s answers aren’t equal or better than Wikipedia - assuming Wikipedia is one of the training datasets. Is the temperature causing it to be probabilistic & other sources are carrying more weight?
lazide
You can think of a LLM as a type of lossy compression of knowledge.

With that in mind, is it really surprising that you don’t get the ‘right’ answer out? Any more than if you compress an image with JPEG, a given pixel isn’t the ‘right’ color anymore either?

They’re both close (kinda) at least, which is the point. If you wanted the exact right answer, don’t use lossy compression - it’ll be expensive in other ways though.

ksenzee
What a great metaphor. I’m adopting that immediately, thank you.
danielbln
Wikipedia is the best first point of entry, but if you do use Claude, just tell it to do web search for you: https://claude.ai/share/73e67582-3e03-454b-aa12-e8906bd7b3fd
croes
Why not just do a web search?
geocrasher
I can't speak for users of Claude, but as a user of Perplexity, having an LLM do a web search has uncovered sources I'd never have considered. The only time I use Google anymore is when I know exactly what I'm looking for.

When I'm in research/discovery mode, I use Perplexity. Its search/analysis is a lot slower than a Google search, but saves me time overall and generally gives me solutions that I'd have to spend time sorting through a Google search to find, in less time than it takes to do so.

wjnc OP
Claude gave a great answer at the link, at least for me. There might be a plus in learning as well since the answer is well structured with a recognizable style. Say, the scientific article above, has a distinct style and really was not high school physics level.
pfdietz
Uranium was very enriched back at the formation of the Earth, so for a given geometry it would have been much more reactive.

However, uranium ores are often formed due to redox processes, since U(VI) is much more soluble than U(IV). So maybe concentrations wouldn't have been as common back before the Great Oxygenation Event about 2.4 Gya. Still, that leaves ~600 Mya between that point and this reactor, which would be not quite one half life of U235.

This item has no comments currently.