A Few Thoughts on the Future of Tech

Josh:
00:00
We've come to the end of another season, but instead of talking about another computer component,

00:05
I wanted to talk about where I think computing is going in the next 5 to 10 years.

00:10
This is just me throwing ideas out there, so take it all with a grain of salt.

00:15
And as we know, the world can change overnight, so maybe these are accurate, or maybe they will never come to fruition.

00:24
The wrong thing to do is just go out and buy a computer and then learn about it.

00:29
You'll learn, but you'll learn a lot of things that maybe you didn't want to learn.

00:33
A computer that you buy today will likely be obsolete 6 months from now, and there's not a dang thing that you can do about it.

00:42
My name is Josh, and I'm able to keep this podcast independent and advertisement-free because of support from listeners like you.

00:51
If you are finding value in what I'm doing here, consider becoming a paid supporter at members.sideofburritos.com.

00:59
And as a thank you...

01:00
members get early access to new videos, ad-free versions of everything, bonus content, and access

01:06
to a live monthly Q&A. Thanks for considering. Now let's get back to the show. Unless you've

01:13
been off-grid for the last few years, you might have noticed AI has seeped into every facet of

01:18
our lives. From support chat pop-ups on websites, to smart appliances, which were already a dumb

01:24
concept, shoving AI into them to make them smarter, but also at the same time dumber. And we also have

01:31
circular investment in AI, which has created a bubble just waiting to burst. I have four predictions,

01:38
theories, or just general gut feelings about where I think things are going.

01:43
AI will take jobs. We're already seeing companies firing developers, shoving AI in its place, and

01:49
hitting go. AI has the benefit that it can work 24-7 and scale very quickly. The downside of this

01:56
is that it's not the magic it has been marketed to be.

02:00
I use it quite a bit, and I pushed back on it initially, but as someone in tech who wants to see and learn what's going on, I would be doing myself a disservice if I ignored it on principle.

02:11
It does help, I can't deny that. I can dump a bunch of logs from my server into it, ask what happened at x time, and it does a great job piecing things together.

02:21
I'd also be lying if I said some of the models weren't pretty decent at helping code, but it has a level of confidence that makes me uncomfortable.

02:29
I know behind these answers are a bunch of GPUs and some AI models trying to provide an answer it thinks I want, but that confidence gives it a level of validation that I don't think a lot of people question.

02:41
They take the answers at face value and run with it, and that truly scares me.

02:48
We already have issues of headline reading, trusting random information from strangers online, but at least that was somewhat easier to analyze.

02:56
With AI, people can get the answers they want.

02:59
With AI, people can get the answers they want.

03:00
instead of the right answer. And that's a problem. The second is that you will own nothing and be

03:06
happy. I think that saying can sum up my next point. Companies want you to pay for everything

03:12
on a monthly subscription basis. And I think once the AI bubble bursts, all this hardware and data

03:19
centers just sitting there will need a purpose. And I think they're going to come for our personal

03:24
computers next. Years ago, Google had a product called Stadia, which they killed, like most things

03:30
they do. It was a cloud gaming service you connected into, and the games streamed from their server.

03:38
This was back in 2019, and one of the guys I worked with had a subscription to it.

03:43
So we tested it out at work, playing some games during the workday, because why not? It was pretty

03:49
cool and worked well. But years later, thinking about it, I can see the repurposing of AI hardware

03:56
going to this use. You'll buy a shell of a computer that can...

04:00
next to a data center where your supposed personal computer is running. You'll do all your work on

04:05
the shell of a computer, which is basically just a screen, keyboard, and mouse, and all your data

04:10
and processing happens remotely in the data center. A company I used to work at had something like

04:17
this set up, basically a terminal server that allows centralized management of workstations

04:22
and users. Fin clients connect in, usually over RDP, and give the employee access. It works great

04:29
in a corporate environment, but it seems like a next likely step if they want to monetize what we

04:34
used to purchase once and turn that into a reoccurring subscription. Just like Uber and Lyft

04:40
and every other SaaS startup in the last decade, it will start out dirt cheap, they will then kill

04:45
the competition, and once everyone's locked in, raise rates. The third thing is that I think we've

04:52
been spoiled with hardware. I think it made a lot of code turn to lazy code. Why spend time optimizing,

04:59
like they did better?

05:00
in the day when you only had one megabyte of ram when you can just throw cheap hardware at it and

05:04
make it work but since corporations are purchasing all the available hardware which is causing the

05:10
prices of gpus and now ram to skyrocket i think we might see a shift to making what is already out

05:17
there or maybe older hardware work longer which i am a big proponent of so long term what do i think

05:25
this means i think it'll lead to leaner simpler more efficient code and i don't think that's ever

05:32
a bad thing but we'll see and that brings us to the last item which is that i think we will see a

05:38
shift away from the cloud for corporations i don't know how many of you listening have ever interacted

05:44
with aws azure or gcp but the original promise was put your apps data and services with us and you can

05:53
essentially fire all your sys admins and save money well over the last decade or however long it's been

05:59
that

06:00
has not happened. AWS or any of those other massive cloud providers still require a massive

06:06
amount of administration and dedicated employees. It's not set and forget like they would make you

06:12
believe. Not to mention the extraordinary markup on hardware. From storage to bandwidth, the rates

06:18
are insane. The truth is that hardware has become much cheaper, if you ignore the RAM prices lately,

06:24
and it's a one-time expense. I don't think this will happen in the next few years, but I think in the

06:30
next 5 to 10, we'll see more and more companies bringing things back in-house. I think there's

06:36
been a lot of fear-mongering over the years about hosting things yourself and how difficult or

06:41
unreliable it is. The fact is, is that it's not. Enterprise hardware has many redundancies built

06:47
into it and it's extremely reliable. And at the end of the day, we still see AWS, GCP,

06:54
and Azure outages that take down massive parts of the web. So yes,

07:00
you can shift the blame to someone else when your customers ask what happened, but to what end.

07:06
So there's some of my thoughts. Again, these are just my opinions. Don't base any of your

07:11
decisions on them. But maybe they can give you something to think about. That's the end of

07:17
season three, and I'll see you next season with something completely different.

07:24
In the Shell is written, researched, and recorded by me.

07:29
And if you ever wondered how a computer gets drunk

07:32
and takes screenshots, that's it. Take care, and I'll see you next season.