I guess I'm confused, how is a technology with a decades-long service life and is basically a capital investment subject to the same sort of label of 'bubble' compared to the explosive growth of something using a commodity model with obsolescence measured in years?

In 2020 I had to occasion to have some OS1 singlemode fiber installed back in 1994 terminated into splice cases and put into use. That fiber sat for basically a quarter century and was then usable when I needed it.

Where I work now I no longer have primary responsibility to deal with cabling infrastructure, but we still light up metro-crossing fiber between locations that could well have sat dormant for decades. Costs to install pathways right now are RIDICULOUS, like upwards of $1000/linear foot for underground conduit work. Stuff installed for a tenth or less that price 20 years ago is paying for itself now even accounting for inflation if I'm not having to spend $50k to go fifty feet between buildings. And unlike point-to-point wireless shots there's no recurring licensing fees to the FCC, there's no service-subscription costs to the wireless equipment manufacturer, and there's basically no lifecycle costs to regularly replace the connection.

If there was any sort of fiber bubble, it was that those looking to profit off of it weren't thinking like a utility, where the ROI takes awhile to see, but the investment in the installation lasts for decades, not months or years. It's only a bubble if you're not a long-term thinker.

So reading, 'riting, and 'rithmatic?

sounds like the sham nonsense isn't limited to the school kids.

I could see that happening.

And to be fair, I don't hate USB-A like I used to, or frequently disconnected and reconnected cables/dongles/ports I like it. It's not as durable as I'd like, but it's physically big enough that if junk makes its way into the port or plug I can clean it out. If the outer housing ends up bent I can bend it back. I can put micro-SD card readers into the socket that are nearly flush with the socket itself.

USB-C is more fragile than USB-A, if something gets into the connector it'll probably damage it. I've seen this happen to phones, and to cables, I could see it happening to computers as well.

True, but when I've been on committees to help make decisions, I tend to play east-german judge, and sometimes this helps steer things away from techbros that don't actually know the tech part of their business and are basically overgrown salesmen who've risen too high in the corporate ladder.

My field is networking rather than software, and while every vendor is usually pushing some new hotness when they do their bring-the-customer-to-the-experience-center, I'm the skeptic asking what they're using for themselves. When Cisco, for example, was leaning hard into their Software-Defined Networking replacement for Trustsec using what they were calling at the time "security group tagging" to basically add tokens that all access-edge and forwarding switches and routers needed to pay heed to in order to enforce end to end security for traffic, I asked them what they were using for themselves. Turns out despite them pushing SDN and SDAccess they were still using Trustsec and the SD- features weren't even functional for an org our size, it couldn't handle the number of endpoints. It couldn't even handle half of the number of endpoints that we saw on a regular basis. And it required basically full vendor lock-in for all access switches, distribution switches, routers, and firewalls to truly work. If you adopted it you were stuck with them basically completely until the end.

I've heard horror stories from other orgs that implemented it too. DNA/Catalyst Center problems, ISE problems, problems when working with endpoints that are not part of the SD- zone, etc.

Or as artists call this, "Mixed Media."

So what exactly are these three Rs that you feel so strongly about?

Isn't a lot of the business use case intended to replace the armies of code-monkeys churning out mediocre code for commercial products? Wouldn't promoting AI in this sense be the org pushing to make itself obsolete? The previous story is literally, "Morgan Stanley Says Its AI Tool Processed 9 Million Lines of Legacy Code This Year And Saved 280,000 Developer Hours."

It also has some rather significant security implications if the company sold that it had sandboxed customer accounts where customer data was only supposed to be exposed to the customer or to an AI system, not to human beings.

The only way I see a Golden Dome scenario working for siloed nukes is if the interceptor missiles are in orbit and basically just moments away from being able to launch to intercept, where the orbit takes them over the adversarial nation.

Since no nation wants their adversary to have missiles in space ready to rain down upon them at a moment's notice, I fully expect that any serious effort to deploy such a system would result in the most diplomatic and economic pressure possible. I don't see Russia or China tolerating armed American satellite anti-ballistic missile platforms orbiting above them any more than the US tolerated the Soviet Union trying to put ICBMs in Cuba.

I used to believe that sort of thing but I don't anymore.

Too many Wallys have patrons who keep them employed even if they actively screw up and cost the company money or create major setbacks. Those managers as patrons have enough clout to both protect themselves and their proteges.

I agree to an extent about your headcount premise, because it's common for an org to implement hiring freezes for a time, and if vacant positions are vacant for long enough then they're deprecated. A useless employee actually on the books during a hiring freeze may prevent that position from being irrevocably lost if the hiring freeze is lifted and the manager is free to try to fill positions again.

I have a Dell XPS 15 that I bought used college surplus that lacks USB-A ports altogether. I can fully believe that a port first used on computers 29 years ago and last physically revised sixteen years ago would be retired from most devices, same as how I watched DE-9 and DB-25 ports for serial and parallel, PS/2 ports for keyboards and mice, and readers and expansion ports like CF, SD, Micro SD, and PCMCIA have gone away.

it's done methodically and slowly because disruptions in service due to rash actions can lead to loss of revenue. Sometimes to significant loss of revenue.

To act rashly without protection from above is a resume-updating experience.

Back in the day a popular expression was, "Nobody was ever fired for buying IBM." This was because IBM was that slow, methodical process and they tended to maintain strong after-sale support so long as you were paying for it. A place I worked for 20 years when through two lifecycles of the AS/400 / iSeries platform because IBM made it very, very easy, safe, and secure to do this. When they finally chose to go away from the minicomputer model to distributed/virtual on bunches of small VMs clustered on chassis like UCS, they had tons and tons of application problems. Whole functions that the old AS/400 software did were not even considered in the new application and either had to be crash-written or the users told to pound hot sand and find another, often harder way to accomplish their objectives. In a couple of cases they ended up migrating through a couple of different subsequent applications because the original post-AS/400 choice was so awful that it was unworkable.

I would tell this Snowflake corporate exec that the burden of proof rests with him, with his company, to show that it's actually better to go with them than with prior solutions. After all, we know how/that those prior solutions worked.

I'm not going to say that you're wrong, but on the other hand I've never seen penalties for having cameras on personal devices covered up or otherwise disabled. I've also never seen an Orwellian level of doublespeak-autocorrect baked into personal devices either.

I wish that the paper wasn't paywalled, I'm curious how the period of the observation works. Telling me it's every ~44min doesn't tell me for what duration we see signal versus for what duration we don't see signal, or how that signal changes over the observation window.

I've found some decent rubberdomes. I used Maxiswitch Gateway 2000 "Anykey" keyboards for awhile, and I also variously used Sun keyboards.

I'd even go so far as to say that the good rubberdomes are far better than crappy mechanical keyboards.