They mention standardisations and cost savings in their paper, as well as solving the heat load per cell problem by decreasing cell size. They also mention that there’s been a lot of micro-CPV module designs but that they haven’t been scaled up. Some quotes below:

Various researchers and developers have been exploring different micro-CPV module designs [5], [6], [7], [8], [9], [10], [11], [12], [13]. Most approaches have been tested on small prototypes or minimodules, while fewer have been realized with aperture areas (Aap) above 200 and 800 cm2,[…]

By decreasing the sizes of the primary optics and the solar cells, the heat load per cell is minimized. This reduction allows for sufficient heat spreading via the circuit board, enabling the direct assembly of solar cells onto the circuit board on glass.

At Fraunhofer ISE, we have developed a micro-CPV module concept [17], [18], [19], [20], [21], which is based on parallelized manufacturing processes and commercially available components.

The final module features a panel size of 24” × 18”, which is a standard in the microelectronics industry, facilitating machine adaption without necessitating special adjustments.

• https://ieeexplore.ieee.org/document/10938951

What are concentrating photovoltaics? One of the ways to increase the output from the photovoltaic systems is to supply concentrated light onto the PV cells. This can be done by using optical light collectors, such as lenses or mirrors. The PV systems that use concentrated light are called concentrating photovoltaics (CPV). The CPV collect light from a larger area and concentrate it to a smaller area solar cell. This is illustrated in Figure 5.1.

Also, from the article - 33.6% efficiency in real-world conditions:

A 60 cell-lens prototype was studied for a year. In “real-world” conditions, CPVs achieved up to 33.6% efficiency. The 36% mark was posted at 167 degrees Fahrenheit. The prototype showed no signs of degradation, according to IE.

I’ve had smooth sailing both with Fedora GNOME and Fedora KDE. And it’s another “common choice” with good enterprise support thanks to Red Hat, just like Ubuntu and all their baby distros.

Yeah, if the user can’t verify the truthfulness of the LLM generated text then it has no use as it can’t be trusted.
I’d say the thing LLM is best at for me is giving me a starting point, in the same way that it is easier to fix someone else’s writing than it is to write something from a blank slate. Oh and copilot is generally up to date on where in Microsoft 365 Admin you find an option right now, compared to all the outdated official and unofficial guides out there.

Many comms moved to piefed, are they counted in that graph?