From an engineering point of view their control loop must be much faster. Things that fly need much tighter loops than we do that lumber about on the ground.
Watch a bird come in fast and then land on a narrow twig.
Bird neurons are roughly half the length of mammal neurons. This means you can pack in around 8x more neurons in a given volume.
> CUDA is a huge reason NVidia is so big right now. Their GPUs are broadly similar to AMD. Their software libraries are better.
This was pre-2026 story. CUDA isn't a moat anymore.
Now it's mostly market dynamics: People always ship ad optimize things for Nvidia because it's the most widely used platform, and people buy it because it has the most software.
This is much less of a barrier than it used to be though. Buggy AMD GPU drivers remain a problem though.
> it is the compiler to use it that is the hard hard part
We're in a wildly different world from the one Itanium got to live in and compilers are something with a super tight automated verification loop that can make LLMs thrive. Add that plus cheap formal verification and I think this concern goes away almost completely.
I own two of their devices and some of the decomp fairies flittering through the gaming industry might just take a visit to their stack for some internal prodding…
Not only that, but if you were to point Astra, Argon, or Opus at their patents and tell it, "Find some prior art that will invalidate these claims", I suspect the odds of success would be better than even.
Maybe it is time to run a friendly LLM and claim to be worried citizen who is seeking relief by reversing the firmware and excluding GPL lisence violations from their nightmares.
Yes, a lot of the ideas are not new.
Nevertheless, it is currently an active field of research. One of them is "Integrated Sensing and Communication" (ISAC [or ICAS/JC&S/...]).
One idea is to exploit the signal's structure - despite simply trying to correlate the received line-of-sight signal with its delayed copies.
Especially OFDM has a nice structure. You can estimate delays (-> distances) as frequencies along the subcarriers and Doppler shifts (-> velocities) as frequencies along consecutive OFDM symbols. Estimators like maximum likelihood, ESPRIT, MUSIC, and compressed sensing etc. are your friends here for high-resolution results.
Unfortunately OFDM has a problem with large Doppler shifts (velocities). Therefore, people are looking at other multiplexing schemes like OTFS or AFDM.
there's still / again a number of Czech companies in the sector ... the "passive Radar" came from a Czech cold war project (Tamara). Yes it had a spotlight moment when some F-117 were shot down in the Yugoslav war.
Both are true, it is a single point of failure and a single point of convenience.
What always puzzles me about this stuff and critical infrastructure like nuclear powerplants, is why is an internal network connected to the public Internet at all?
If you have to connect it to the Internet, then why is it talking to unknown devices?
In 2014 Microsoft cancelled the dedicated SDET (Software Development Engineer in Test) role, triggering significant layoffs.
reply