JATth

Deploy the hypnobots! Marketing +10000%!

The timeline with the current AI sounds like ripoff from “the paper clips” game, where the only task is to make the number go up. If you play that game, you’ll spot that dark patterns used have spread everywhere since.

You have seen the equations? 99% you don’t get anything useful or somehow split the fucking universe, so that time-reversed flow of information won’t break the causality. (And it fucking does break the universe, the local bubble we are in. Please, don’t google false-vacuum.) We don’t even know the speed of light in one-direction, because information travels at the fucking speed of light, whatever that is. Yet, you measure a spin of electron, and the other fucker in the opposite side of the universe, is known to be opposite in the instant.

The action is minimized by nature, in one way or other, even if this means punching a hole in the space-time continuum, to start a tree-of-universes descending from the host.

/s (I had enough beer for today.)

I don’t know, if you gather one proton’s and two neutrons worth of energy, localized entirely within the supposed space of an atom, you could generate mass from energy. However, what that energy converts into, is entirely random, so I would guess most of the time you just get realllly high energy photons out of it.

Initially, x86 CPUs didn’t have a FPU. It cost extra, and was delivered as a separate chip.

Later, GPU is just a overgrown SIMD FPU.

NPU is a specialized GPU that operates on low-precision floating-point numbers, and mostly does matrix-multiply-and-add operations.

There is zero neural processing going on here, which would mean the chip operates using bursts of encoded analog signals, within power consumption of about 20W, and would be able to adjust itself on the fly online, without having a few datacenters spending exceeding amount of energy to update the weights of the model.

arxiv.org: a pre-print source, this is not peer-reviewed yet. So, until other papers refer to this paper, it has low significance in the literature. The paper has references to other papers and to previous corpus of knowledge on the subject, which is good. However, this paper is based on simulations only.

Crucially, the scheme identified here does not negatively impact electricity production, and is also compatible with the challenging tritium breeding requirements of fusion power plant design because (n, 2n) reactions of 198Hg drive both transmutation and neutron multiplication

Monte Carlo transport calculations show that neutrons produced in a tokamak power plant can convert the abundant isotope 198Hg to stable 197Au via the (n,2n) channel, yielding several tonnes of gold per plant-year without compromising the tritium breeding ratio.

The decommissioned blanket material would increase in value, and the Mercury-198 in the blanket doesn’t majorly impact its effectiveness in transmuting lithium to tritium.

The “worst” gold isotope half-life is less than a year, so only a modest cool-down is needed for the output:

197Au to be Class 7 when activity concentration is > 2700 pCi/g, which is reached after 13.7 years for the initial concentration listed.

An even more stringent constraint can be applied for any gold that will be regularly handled by the general population. As a highly conservative requirement, we can stipulate that this gold must be less radioactive than a banana. Due to 40K content, bananas have an activity of ∼ 3520 pCi/kg, or about 420 pCi for a single banana. To meet this requirement, a troy ounce of gold with the initial isotope mix shown in Table 2 must sit for about 17.7 years to be below a banana equivalent level of activity.

What strikes me here with this paper is the suggested liquid blanket, so this would be a kin of D-T Fusion MSR. We now have two proposed technologies behind being able handle hot radioactive liquids.

After using dnf a bit:

• All the default answers are backwards to me, so dnf quite literally ignores my input.
• dnf search did not show, by default, if a matching package is already installed.
• Perfect perhaps for newbies, since dnf asks you trice.

yeah… arch is not leaving me anytime soon. The option to makepkg from source a few custom packages is very neat.