• 0 Posts
  • 116 Comments
Joined 3 years ago
cake
Cake day: June 10th, 2023

help-circle

  • When it comes to PC OEMs I’ve observed that right now Dell has really good driver support. They’ve got increasingly good utilities for keeping drivers up to date and they’ve been doing a good job of loading drivers and their utilities into Microsoft’s relevant repositories where it makes sense, and that driver support tends to actually last multiple years. I can often pull down a new UEFI update for a 5 year old Dell PC, which is not something I can say of most hardware manufacturers.

    So at the threat model of an enterprise org, I’d prefer Dell for that reason alone. Lenovo and HP have tried to implement some of that, HP seems to have given up after building the bare minimum and Lenovo has their typical wonky software that will become good after a few years if they keep investing developer time into it, but knowing Lenovo there’s about a 60% chance some new executive will come in and change direction, and the software will be made increasingly unusable then later discontinued due to lack of use

    However for my personal computers, there’s a high chance it won’t even be running Windows so I just buy based on hardware & price alone


  • My very first desk job was an outsourced support role where 99% of calls we simply found the answer in the user manual and provided that to them. The other 1% was usually something isoteric we’d forward on to someone within the company. The amount of callers who’d say “I’ve read the user manual cover to cover and I just can’t figure out how to…” And I’d just try to page 12 on the PDF and read them the instructions word for word

    At the scale of HP, I can see the logic. You know that, say 60% of calls are directly covered by the knowledge base because you have those metrics. That’s means 60% of their support overhead could be eliminated if they somehow got people to read those documents. Hardware sales usually have very thin margins and a customer contacting support can easily cost more in support than the entire profit margin of the product (and often it’s a self-inflicted problem) and of course an RMA for most products basically negates all profit from that sale. It’s a real business challenge and the asshole solution is to simply tie people up for 15 minutes in the phone system before connecting to a human to see how many people hang up and how much that reduces support load






  • The price is irrelevant, because they aren’t for you or regular consumers. They’re already reserved and being shipped to AI data centers.

    I mean this is the standard operating procedure for all top end data center products, they aren’t sold on consumer marketplaces but can be purchased by suppliers with existing contracts and relationships

    As they ramp up yields larger capacity drives will slowly trickle into more consumer channels until eventually the 40+TB drives are like the 8-12tb drives are today






  • The thing with adding lanes is induced demand. By nature of there being more space for cars on that road more drivers will choose take that road over other roads. Cars don’t magically come into existence, people drive them, and people drive them for a reason, most commonly to go to/from somewhere

    Trains (and bikes and buses) take cars off the road. Every person riding on a transit solution that isn’t a car is a individual vehicle trip saved. When every vehicle contains an average of 1.2 people in it, you’ve got very close to 1:1 vehicle reduction for every trip that’s not taken by car

    So to your point, are some number of non-drivers choosing not to drive because of traffic? Probably a small number of them. But a complete transit system that has the real world effect of fewer cars on the road will mean few people owning cars. Why would a family own 2 cars when one is parked most of the time? Why spend $20k on a new (to you) car if you’re barely using the one you have/had? Fewer cars means less cars on the road which means less traffic. This is the dream.


  • Goddamn we need so much high speed rail, and yesterday

    And so much more standard speed rail too! There’s tons of railroad lines all over the country, and even many old stations still standing. Let’s start building RDCs again (or better, a modern equivalent) and start running passenger services on all of those lines

    The challenges are several fold. For one, there’s basically no manufacturers of passenger railcars left in the country. Occasionally a network upgrade will lead to one being spun up for a few years then it’ll shut down once the order is fulfilled because there’s no consistent market for passenger railcars in North America.

    I’d propose using a mix of historic British Rail procurement practices and current military contract practices where you put up a pot of money for up for say 3 companies to develop a prototype railcar meeting a specific spec. Make the spec for a fairly basic car and be ready to update station platforms for ADA compliance rather than forcing the cars to be compatible with 20 different platform heights and designs. Then test those 3 prototypes and the winner receives a bonus as the design is purchased by the federal government, and next you license that design out for all manufacturers in the country to produce, followed by an ongoing order of say 48 railcars per year from 5 different manufacturers and you have 248 railcars per year (enough to replace Amtrak’s entire current fleet within 10 years) from 5 different companies (reducing risk of one company mucking it all up) all manufactured with local labor and you have a standard design that is already in active production for other operators to order as well. Repeat this process for more equipment and designs as needed, and suddenly you have a bunch of known standard designs that your network can be built to and you have health competition between manufacturers which will be big enough (because each will have around 50-100 million dollars a year in revenue from that one ongoing federal contract alone) to start performing their own independent R&D to make their own unique stock to try to attract more orders from rail operators

    This is what the federal government exists for, making ambitious infrastructure projects like this possible. Policians just aren’t interested in thinking big enough