This is highly ridiculous. It's missing entire categories of costs, not to mention selling 1000bu of corn is not easy and you won't get that price for it.
The real question isn't "Can AI do x thing?" but "SHOULD AI do x thing". We know how to grow and sell corn. There is zero that AI can do to make it more "efficient" than it already is.
NAT is not a security measure at all. It just obscures what's behind a firewall, but that is leaky and not reliable from a security perspective. It might make you feel better, but that is not security.
A firewall has nothing to filter, if nothing is routed to it. My IoT devices communicate with a server running in my network. As long as I am behind an IPv4 router, their communications to that server will never make it to the internet, and any communications from the internet have no way of addressing any device on my network. I literally can't add any security to a firewall because there's no communications to handle. Sure, I have personal computers on the same network, which aren't on a separate VLAN because I'm not familiar enough with my router to set that up, so a compromised PC could forward attacks to my IoT devices, but the firewall would be useless at that point.
If I have an IPv6 router, I can miss-configure it in a way where all of my internal communications between IoT devices work as expected, but they also have discoverable addresses on the internet. This would give the firewall something to do, but I'd rather there be no route in the first place.
Also, if I trusted myself to properly configure my router for IPv6, I would put all of my IoT equipment on ULAs, which much like an IPv4 NAT would leave me with nothing to configure in the firewall.
If I were to take your claims at face value, using GUAs with packet filtering is far more reliable and secure than ULAs, and that seems preposterous.
A properly configured firewall for sure adds security, but isolation always wins out.
Yea, people consider NAT a firewall, but at best it stops direct connections from outside. People use this as a rationale to non secure individual devices on the network. Then the moment a single device on your network is compromised (do you really trust that Chinese IOT device?) every host that doesn't have its own firewall is at risk.
With IPv6 you at least say "Holy crap, anyone could connect to this, I better secure it from outside and inside attacks" which is how actual security works.
RDMA is a networking standard, it's supposed to be switched. The reason why it's being done over Thunderbolt is that it's the only cheap/prosumer I/O standard with enough bandwidth to make this work. Like, 100Gbit Ethernet cards are several hundred dollars minimum, for two ports, and you have to deal with SFP+ cabling. Thunderbolt is just way nicer[0].
The way this capability is exposed in the OS is that the computers negotiate an Ethernet bridge on top of the TB link. I suspect they're actually exposing PCIe Ethernet NICs to each other, but I'm not sure. But either way, a "Thunderbolt router" would just be a computer with a shitton of USB-C ports (in the same way that an "Ethernet router" is just a computer with a shitton of Ethernet ports). I suspect the biggest hurdle would actually just be sourcing an SoC with a lot of switching fabric but not a lot of compute. Like, you'd need Threadripper levels of connectivity but with like, one or two actual CPU cores.
[0] Like, last time I had to swap work laptops, I just plugged a TB cable between them and did an `rsync`.
The "R" in RDMA means there are multiple DMA controllers who can "transparently" share address spaces. You can certainly share address spaces across nodes with RoCE or Infiniband, but thats a layer on top
I don't know why that NVIDIA document is wrong, but the established term for doing DMA from eg. an NVMe SSD to a GPU within a single system without the CPU initiating the transfer is peer to peer DMA. RDMA is when your data leaves the local machine's PCIe fabric.
He describes what he sees as a monopsony. That is not misleading. You can have lots of options and still be stuck in this monopsonistic (sp?) world that controls your rights and your financial future.
The real question isn't "Can AI do x thing?" but "SHOULD AI do x thing". We know how to grow and sell corn. There is zero that AI can do to make it more "efficient" than it already is.
Come on.
reply