Sorry, but I heavily suspect that you didn't read the original article and I suggest you to do so.
Security has nothing to do with 1394's market adoption, in the 2010s, some desktops still have 1394, and some laptops still have ExpressCard, and you don't see anyone complains about security, only some security researchers.
The reason was partially cost, but also a bad business decision by Steve Jobs, which made Intel and Microsoft stopped their efforts for pushing 1394, largely damaged it before it even got a chance for wider adoption.
> FireWire's principal creator, Apple, nearly killed it before it could appear in a single device. And eventually the Cupertino company effectively did kill FireWire, just as it seemed poised to dominate the industry.
> Intel sent its CTO to talk to Jobs about the change, but the meeting went badly. Intel decided to withdraw its support for FireWire—to pull the plug on efforts to build FireWire into its chipsets—and instead throw its weight behind USB 2.0.
> Sirkin believes that Microsoft could have reversed the new licensing policy by citing the prior signed agreement. "Microsoft must have thrown it away," he speculated, because it would have "stopped Apple in its tracks."*
--
> Good to know that USB4 implementations are making the same mistake as FireWire implementors did in choosing performance over security.
DMA is optional and can be disabled.
Speaking of usability, DMA is limited to the Thunderbolt part (which is ultimately the PCI-E part) of USB 4 standard. You don't have to use USB 4's PCI-E. Unless the USB device really needs PCI-E, like RAID storage box or a GPU, disabling DMA is not a problem. Requiring explicit user consent for PCI-E is also a solution, operating systems already does it for Thunderbolt devices. Same treatment can be applied to USB 4.
And again, calling out DMA support is only shooting the messenger while ignoring the underlying problem. DMA is essential for any high-performance system bus, and it can a problem regardless of who supported it, or whether it's connected to an external port. A Wi-Fi controller handles untrusted data, but despite being an internal device on the motherboard, it's still potentially an exploit vector for DMA attackers.
> (or an OS with fuzzed IOMMU and also controller firmwares)
Only IOMMU is needed (and possibly a good driver, but not strictly required, see QubesOS). If your IOMMU is good, you don't have to trust firmware or hardware, since arbitrary DMAs are blocked by IOMMU, just like how memory protection (MMU) works but for I/O. Memory spaces are isolated.
The IOMMU vulnerabilities I previous mentioned are already patched, and a complete bypass of IOMMU is unlikely to occur in the future. Driver bugs are still an issue, more RAM regions can be exposed by drivers than what's really needed, and potentially you can pretend to be any hardware and feed bad data to trick the most vulnerable driver to expose more RAM regions, and I expect to see more bugs.
Turn on IOMMU, and apply OS patches often, and you'll probably be fine.