The only design flaw is the (non-)docking station setup doesn't give it sufficient cooling, and the location of the power button makes it too easy to accidentally press.
Fortunately you can switch so a single press of the power button doesn't shut down the OS. And we're 3D printing desktop supports that fix the cooling situation.
Developers typically get ProBooks with matte non-touch screens. Everyone else gets Envy or whichever they decide.
We usually get 15" laptops specced fairly minimally then immediately pop in a new NVME M.2 along with more RAM. Upgrading them usually only takes a few screws.
There is one laptop here from about 4 years ago that required no screws whatsoever to remove the bottom. That was pretty awesome.
edit: We've had such good luck with HP laptops that we decided to buy a small server. Mistake. Do not attempt. Dell servers, HP laptops.
The only annoyance I have is that when I press capslock some garbage HP software needs to wake up from its deep slumber to show a notification on screen and that locks the system up for like 5 seconds.
Recently acquired an HP Spectre x360 and so far that's working well too, although it is less upgradable than the Elitebook.
Being more serious I’m not sure this is considered in the debate about repairability. If Apple soldering memory to the motherboard is more reliable than using a socket is that better or worse for the life of the product.
But it's most definitely not. The SO DIMM socket is a well proven design, I have never in my life seen or read about one failing. But failing memory is a pretty ordinary thing to happen - rendering the entire computer useless in Apple's case...
Note I’m not defending this practice, just wondering if it’s really the problem it’s made out to be.
Luckily, it's a matter of market expectations around repairability rather than an expectation of increase failure rates.
These are their enterprise machines and there's an expectation in that market for a certain level of on-site repairability, either by purchased warranty plans where the manufacturer sends out a repair tech same or next day, or via on-site tech support being able to purchase spare parts and fix them themselves. That's why you see both repair manuals and spare parts readily available.
Lost productivity from sending a machine to a repair depot for a week or two can vastly outstrip the cost of the machine itself, so laptops which require that for even the most common failure modes will tend to get dinged during the procurement selection process. My company actually gives a choice between Dells or Macs for new hires, and the Macs require business justification and SVP approval, since AppleCare for Enterprise is both incredibly expensive and still very limited in what on-site repairs are even capable of being performed in the first place.
The vast majority of consumers are neither capable of repairing their own devices, nor take that into account when purchasing products. So repairability is prioritized accordingly in the design process for consumer devices.
Protip: The same holds true in most things. Consumer lines tend to optimize form and features[1], whereas business/commercial lines tend to prioritize utility[1] and reliability[2]. Dollar for dollar, you'll get more bang for your buck if you shop at commercial suppliers (such as restaurant supply stores for kitchen stuff).
[1] Features and utility are two different things. Cutting edge and unique features, no matter how gimmicky or little used, can be a selling and marketing point for consumer products. These tend to be pretty worthless in commercial products, at least until they've standardized and matured into mainstream functionality.
[2] Commercial products are designed for long periods of sustained use, and products are designed with the materials, durability, and/or repairability to support that. Consumer products are designed for light, sporadic use and materials/designs are chosen accordingly.