Hacker News Reader

Comment by acidburnNSA | Hacker News Reader

Big dream of mine would be to align nuclei of nuclear fuel atoms just so and then induce fission in such a way as to get one delayed neutron precursor and one other quick-to-stability fission product. This would allow fission power without any long-lived waste products or afterglow heat cooling challenges that dominate accident risk. Physicist friends have told me it's impossible. I've only accepted impractical for now.

View on HN

You're thinking about the atom very classically, At the scale of the nucleus things just don't "exist" in "places". Processes are truly random and things literally don't have position/momentum/rotation/alignment until you do the thing that requires them to decide where they were and what they were doing at the time.

Simpler than nuclear physics is just the electron. There is no meaningful answer to where it is around an atom at any particular time. You can either get a location or a momentum or half the information about each if you poke it, but that's just its response to being poked, it wasn't "actually" there until you poked it.

From what I recall, quantum things have well defined states, even if those states may not correspond to position / momentum / rotation / alignment.

By correctly molding the energy landscape it may be possible to set the states and state transitions up in a beneficial way for what he proposed.

Eh, not really. You can futz with the probability distribution, like a fast neutron will cause a different distribution of fission products than a slow one... but it is still a very random process. You can't control it like an expert at a billiards table. Especially the strong force mediated interactions between particles in the nucleus. Some people just won't believe you though.

> Processes are truly random

You can get a Nobel prize or two by proving this.

We don't know about random yet, just that there's no hidden variable.

Some say that it has not been fully proved there are no hidden variables.

https://youtu.be/ytyjgIyegDI?t=86 (Sabine Hossenfelder on Superdeterminism)

The universe... just one big wave function perhaps

https://en.m.wikipedia.org/wiki/Wheeler%E2%80%93DeWitt_equat...

I'm a big believer in Energy Wave Theory.

You want an msr/LFTR and breed away the bad waste.

Now that’s what I’d call nuclear engineering!

More seriously, spin polarized D-T fusion is known to have an enhanced reaction cross section, so there are labs out there researching how to implement it more reliably.

In software terms, this would be as difficult as switching out specific bits from a running program to fix bugs.

Certainly not impossible, but impractical as far as we can see.

that’s pretty much how windows update works (or used to work) to attempt to hot patch certain things without a reboot. Compiled functions in windows DLLs have a 5 byte prologue of nop instructions which is just long enough to overwrite with a jmp instruction to hook the function call and redirect it. https://devblogs.microsoft.com/oldnewthing/20110921-00/?p=95... In WinXP they started compiling this nop prologue in on purpose due to how useful it was. Before that, reverse engineers would need to get a bit lucky and find “room” for a jmp in some code path that was guaranteed to hit, in order to patch an executable without crashing it (iirc, i’m fuzzy on the details). Anyway, certainly not impossible, and we’d al be surprised by what can be made practical if the need is great enough.

> 5 byte prologue of nop instructions

Has this been around long enough that CPUs optimise it out?

I presume there is a long list of CPU optimisations that are specific to the quirks of Windows object code . . .

Reply on news.ycombinator.com