Vertigo1

We need a PC "un-build" feature where you buy a pre-built with an RTX3080 then rip it apart to get at the GPU, throwing everything else in the bin! :)

Biometric Authentication! It occurred to me that, when Apple launched both TouchID and FaceID, they were subjected to a lot of rigorous testing to see just how secure they were and how difficult it would be to fool them. Much was made of how neither could be fooled very easily. Now virtually every phone and decent laptop seems to come with either a fingerprint scanner and/or face recognition but they're never really tested to see if they're secure or not. Laptop manufacturers in particular seem to expect us to use them for Windows authentication and I often see reviews simply mention their presence without any comment on their security, LTT included. How about a roundup of phones and laptops from major brands to see whether these systems are actually as secure as the Apple versions. I suspect many manufacturers are implementing these systems "on the cheap" as just another box to tick on the spec sheet and, in reality, both could be easily fooled/bypassed.

Because he clearly needed the help. Given this was supposed to simulate a real-world scenario, where someone built their own machine and then brought it in for problem solving, Steve had totally the right approach of totally dismantling everything and putting it all back together again. Incorrectly oriented fans and lack of thermal paste are obvious mistakes that should have been checked for first.

Why is Austin's video mirrored left-right?

If anyone cares, this has now been resolved by a 1.45 firmware update for the NIC. Gigabyte and Asus have already made this available for their boards, I presume other manufacturers have or will do the same.

I noticed Linus hit this very issue during his Gigabyte build live stream this evening. It's a well-known issue which appears to affect all Z490 chipets with the Intel I225-V 2.5Gbe NIC, not just Gigabyte. Basically the latest driver from Intel, and which is available from most of the motherboard manufacturer sites, has a bug whereby the NIC dies on a system reboot. The status will show cable/media disconnected and there'll be no link or activity lights on the socket. This only happens after a soft reboot, powering off and on again will not show the problem. You can fix it by disabling and re-enabling the interface. The solution is to roll back the driver to an earlier version. At present the one on Windows Update (1.0.0.13) is fine. Uninstall the NIC from device manager and opt to remove the driver too. Reboot and allow Windows Update to locate the driver (using Wifi or an alternate NIC for connectivity). This is being discussed in various forums but could really do with being highlighted by a more prominent site/channel like this one to give Intel a kick up the backside to get it sorted.

Not a specific product and too involved for a TechQuickie but what I'd really like to see is a proper, objective look at PCIe 4.0. Personally I think it's little more than a marketing gimmick with extremely limited real-world applications (how many people really need more sequential throughput than a PCIe 3.0 NVMe SSD can provide?!). AMD have pushed it so hard that, in many peoples' eyes, it seems like a must-have feature and a major reason for choosing AMD over Intel at the moment. For all the talk of Intel shilling in the industry, it feels like a great many reviewers are just trotting out the AMD marketing line on this technology. Whether it proves me wrong or validates my opinion, I'd welcome a proper, objective look at PCIe 4.0 and what actual benefits it provides, as well as the downsides such as increased motherboard costs (and fans) and even, as alluded to in your 3990X review, whether some PCIe 4.0 SSDs are being tuned for headline bandwidth figures at the expense of real-world performance.