I'm scratching an itch that's been plaguing me since the 1990s... an idea for a computing fabric that deliberately wastes transistors in exchange for performance. I have a hunch it might lead to incredibly cheap Petaflops. Along the way to finding out if it's just a silly idea, or useful, I'm learning VHDL and hope to get a chip made so I can measure the performance in terms of compute/joule and compute/watt. I've written about it here quite a bit, I call it a BitGrid. I give it a 50/50 chance.
Also, I've been watching fringe science as a hobby for a while. I'm hopeful that the Barry-1 satellite[1] which is currently in orbit (and slowly falling towards earth), will soon turn on it's thruster (which is purely electric, no reaction mass/propellant) and shift itself into a 100 Km higher orbit using only solar power and broken laws of physics. I give it a 50/50 chance.
Lastly, I'm just trying to push, ever so persistently, the idea of capability based security into the world. It makes me so damned sad that we've had decades to fix computer security, and know exactly how to do it, but we keep just applying band-aids instead of fixing it. It's like if the power grid never adopted circuit breakers or fuses. - I give that a 5% chance of happening. 8(
-----
[1] https://celestrak.org/NORAD/elements/graph-orbit-data.php?CA...
(Note, this is a public source, not affiliated with the owners of the satellite in question, it's quite handy for other satellites as well)
