lee/hpr-knowledge-base
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial commit: HPR Knowledge Base MCP Server

Browse Source 
- MCP server with stdio transport for local use
- Search episodes, transcripts, hosts, and series
- 4,511 episodes with metadata and transcripts
- Data loader with in-memory JSON storage

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Lee Hanken
2025-10-26 10:54:13 +00:00
commit 7c8efd2228
4494 changed files with 1705541 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

335
hpr_transcripts/hpr4097.txt Normal file
View File

@@ -0,0 +1,335 @@
Episode: 4097
Title: HPR4097: Will they take our jobs? Of course they will.
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr4097/hpr4097.mp3
Transcribed: 2025-10-25 19:35:59
---
This is Hacker Public Radio episode 4,097 for Tuesday the 16th of April 2024.
Today's show is entitled, Will They Take Our Jobs?
Of course they will.
It is hosted by D.O.D.D. dummy and is about 34 minutes long.
It carries an explicit flag.
The summary is I blathered on about my thoughts on robots taking our jobs.
Hello, this is D.O.D.D. dummy with today's episode of Hacker Public Radio.
This is one of those episodes that I recorded several times.
I did an uploaded and now it kept seeming like I was already kind of late, but anyway I'm
going to record this again and then upload it today itself.
We'll see if you're listening to this then I got it done.
My plan is to only do bare minimum processing so maybe take out the automatic tools that
take out empty spaces and that's really going to be it.
Now the topic for today is Will The Robots Take Our Jobs?
I know that's been a topic for quite a while now, but I think it's changed here in the
last year or so and it seems to keep changing at least in my mind.
Anyway, that's going to be the topic and so okay, so in I specifically save robots because
I know there's some debate or people are calling it AI and it's not really AI.
But for me, the question isn't really our large language models, actually AI or not.
I mean that's, it's an interesting question and I guess they're not, but it's not really,
in my mind, it's not the main question to me.
And so the main question is at least in my mind is Will The Robots Take Our Jobs?
And I specifically save robot because I'm not, I'm really not speaking specifically
about AI because I think once we get to the fact that it's really AI, then for sure will.
I'm just talking about the kinds of things that current gen or soon to be current gen automation
can do.
And so to me that incorporates everything from doing current jobs faster and hopefully
with more precision or more accuracy, better results, quicker timing, et cetera, et cetera.
And I know there's a lot of people who are pretty good in the tech world who keep saying
no, it's not possible.
If you're creative or if you're human, you bring something to the table that the robots
can't bring to the table and all of that to me, I agree with.
But the thing is that's not really the concern is that it's going to create, at least I don't
think it should be the concern.
The robots replace the creative among us.
And let's just say, and you guys probably are at least a good judge of this as I am.
But in the real world, the kinds of things I want to consume, whether it be actual products
or digital things, to me, I think probably on the high mark what I would consider creative
or genius is probably less than 30%, I think almost for sure less than 30% of what's out
there.
So let's assume that those 30% are safe.
AI will never be able to robots, AI will never be able to take their jobs.
Let's say that's for their safe.
What I've been saying for quite a while is they don't have to replace the top 20% or
the top 30, or even the top 40%, 50% to cause an existential crisis.
And if I think about all of the professional people, all the people that have jobs that
I know of, I would say most of them, probably around 50%, they're just mediocre.
And a lot of places robots are already good enough to take their jobs.
The only reason they happen is because the people who employ them haven't found the means
or the wherewithal to replace them.
And so maybe right now it's deemed too cheap or not feasible, it's a little early.
But like I said, to me that's 50% of the workforce I think are doing.
Our mediocre are worse, it's my opinion.
I don't have any stats, but that's kind of how it plays out in my experience.
So those are the people that, and let's say I'm wrong and it's only half of that.
Okay, and let's say it's even half of that.
So that gets us down from 25 to 12, 12 and a half people.
So I think it's relatively certain that we can already replace 12 and a half percent
of the workforce.
And if we do that, I think we're already at existential crisis level.
Now I'm not talking about that on top of whoever's already not looking for jobs.
I mean the people right now who are gainfully employed, we could probably replace 12 and
a half percent is kind of low-hanging fruit.
And then what's that going to do to the economy, the governments?
That's already getting pretty close to the Great Depression era here in the U.S.
And it's probably going to have a snowball effect.
So, and like I said, I know I'm on social media that I kind of participate in.
It's mostly IRC and Mastodon.
Like I said, a lot of people that I think are pretty sharp, technically speaking, you
know, they keep saying like they're poo-pooing the idea and that it's total bunk.
But so, but there's also more and more people who have been saying, yeah, you know, a year
ago, maybe two years ago, I didn't think there was any way that, you know, these large
language models could do my job.
And so, but now more and more of those people are saying, hey, you know, this current
iteration, it's, it's competent.
It might be able to greatly assist me at any rate or just kind of do some of my job.
And so, to me, that is kind of indication, at least anecdotally, that we are getting
closer and closer to where the robots could take our jobs.
And on the case of some of these, let's say, coders as one, the, you know, what I consider
at least, decent coders are saying that now, yeah, the large language models are making
their jobs a whole lot easier and they can do things faster.
I know Martin, what do I call him, Martin, who put two Martin, or he's probably would
not appreciate that moniker these days.
He said that he coded something in rust, a language he doesn't even know, and it's, and he's
got it working.
So what's my point here?
Let's assume that Martin doesn't have to worry about his job, because he does seem to
have, at least from where I sit.
He seems to have done things and be sharp enough that he could probably justify his existence
longer than some of us.
But what that very well could result in is some junior programmers not getting hired
or being let go.
If he can do the other stuff and then do rust, and other languages he doesn't even work
in, then that, to me, is a little bit of proof that we are getting to the point where robots
will take our jobs, at least a good chunk of them.
A large enough percentage to, like I said, cause an existential crisis, causes to have
to be in a world where governments take care of us just because we're human, not because
of the job we have or the work we do.
And so it'll be interesting to see how it goes in the next couple of years.
Cause I know, I don't know about where you work, but where I work, they're always telling
us that human resources are, you know, are us, where they're most valuable commodity
or most valuable.
They try to say it in a way that we don't feel like we're cogs in a machine.
But anyway, that's what they mean.
But still, they lay you off, and they have no issues telling you, even if the manager
who's telling you this does the higher ups, they don't have any qualms about telling
you.
Thanks for coming in today, but by the way, we don't need you anymore.
So we're going to, you know, let you go back to your desk and pick up your stuff and
get you out of here.
They don't have any issues doing that.
So it's going to have to be governments somehow do, you know, taking care of us in that
regard.
And that's been true for a while, you know, there's been a talk universal basic income
as comes to mind.
But we've never really moved very, very well on that.
And I've always thought that it was because it didn't affect some kind of a critical mass
of either the type of people that were impacted or the percentage of people.
But I think we're getting pretty close to that.
Now, in my own world, it's been a, I don't know, it's been a while now, I don't exactly
know.
I did a, I asked Chad GTP three to three dot five to a three file compare and cobalt,
which is, which is what I'm, I'm for the most of my career, I've been a cobalt programmer.
And it does serviceable job.
I did have to clean it up a little bit, but after a couple iterations of, of me asking
for changes, it, um, it did a serviceable job.
And now it probably would have taken me who I'm pretty familiar with file compares in
cobalt.
And it probably would have taken me about, probably about a day and a half to, to run,
to code that the way they did it.
And some of the, um, let's say some of the more junior people I work with would have
probably taken them close to a week.
So I find that kind of interesting.
Now, um, what are some other examples, um, oh, let's go creative.
So let's creative is in the arts and, um, so, you know, there's, uh, well, they'll never
be able to replace Picasso or actual real artist because that you need to, that takes humanity.
Well, I agree with that, but the thing is, you can
sell a lot of artwork, um, and we do right now, a lot of mediocre artwork gets sold, um,
but it's low price and it's, you know, quote unquote, the starving artist.
Um, well, I would pay, you know, I would definitely buy and be satisfied with AI generated
artwork, you know, as long as it's cheaper than the starving artist.
Um, and part of that is because I'm not an art connoisseur.
It just kind of has to look nice.
Sometimes it doesn't even have to look nice.
It needs to just kind of blend in with the rest of the room.
And I don't think I'm alone in that regard.
And so the thing is, I think AI is, and I'm going to say call it AI, even though I realize
it's not AI because that seems to be the, the name that is sticking.
Kind of similar to like, even though I don't believe in how the, you know, the quote unquote
mainstream folks use the term hacker, sometimes I just call, I slip into that myself.
Um, but so I think AI could already replace a lot of the mediocre artist.
And so it's true, we're going to still need the genius artists, but you know, we don't
need a lot of the lower artist and they still need to do something.
Now you could say, well, they're not really an issue because they'll just keep being
waiters like they are now.
Um, well, yeah, maybe, maybe that's the case.
So like I said, in my mind, it's already good enough to replace, you know, large percentages
of the working artist.
Um, and I don't really know a lot about working artists, but I think it's probably already
good enough to, to replace a lot of the, um, the, um, what do they call a commercial artist,
you know, make our ads and things like that for, for companies.
Um, and if it's not enough to totally replace them, like I said, I think it's probably a good
enough assistance for the, for the, for the good folks and they can make it to where you
don't need nearly as many people.
And, um, let's take another one.
And I just looked at just like a few minutes ago and let's see, it was posted on YouTube.
It was posted by, um, I think it was posted later earlier today.
And today is, um, April 7, 2024.
And, um, it's Teresa's chaotic corner on YouTube.
She's a writer and apparently she's making a living that she's a writer.
She, you know, she didn't say that's her hobby as far, unless I'm just missed it, but it
doesn't look like it.
And so she's asking the question, will AI end creative jobs?
And, um, it's a writer, a writer contemplates unemployment.
And she does this with a free one, so not even a paid one.
And she had it create two stories, one, um, and like I said, I don't know who this is.
Um, she sells her work, but I don't think she's like, you know, um, a AAA, our, um, writer.
I don't think so.
Um, so I think she would fit, I'm not going to slide her by saying she's mediocre, but
like I said, I don't think she's a AAA writer.
And she asked it to create two stories, two short stories, um, one in the style of, um,
Stephen King with the subject of, uh, well, um, with the subject of AI taking over human
jobs.
And the other one, I forgot what she said, the other one was, but the other, oh, the
other one was to create a short story that's, uh, romance, um, story in the style of some
author that, I don't remember who she, who it was, but some author, um, in with, um, dark
sith, dark mall and sith.
So two star, star trek, star wars characters.
And one of them, uh, apparently was castrated and so, um, well, sketched, castrated.
And what she said was those stories were competent.
Now she said she doesn't think even one of those authors needs to worry about their jobs,
but again, to me, those are not the authors that were really concerned about here, um, because
by the time Stephen King has to worry about his jobs, you know, probably most of the other
writers have already lost their jobs.
And then that meets society has to deal with all of these writers not having to work.
Now she called them the stories both competent.
And she said, um, what did she say?
She said that if you presented those in high school, or maybe even college, as short
stories in your class, it would, it would kind of be okay.
And like you, you would pass.
Now to me, that already is a pretty good advancement.
We have someone here who's a working writer who has, you know, evaluated the Chaget,
I don't know if it's Chaget TV, but evaluated the large language model stories and said,
yeah, this is probably high school or college level.
Now to me, what that says is that's pretty decent support for the thought that, okay, we're,
you know, we're not too many iterations away from, um, from actually being, you know, low
level professional.
And it hasn't been that many iterations since people said, oh no, that don't ever happen.
The stories sound stupid.
They don't make any sense.
Nobody would think a human, no one would think a three-year-old or a five-year-old or
a seven-year-old did this, let alone a professional.
So, um, and so to me, those are the kind of obvious things.
Now I want to bring one more point from my personal experience.
So in my career since so, let's see, I started around 93 and around 96, I guess, yeah,
that's probably about right.
I created my first thing that was improved automation at work.
And since then, kind of my main things that I've done, my main things that I've done
have worked at work, what I consider a value, or improving automation where we had, um,
we had competent people, but for whatever reason they just didn't automate.
And I assume it's a combination of the people who didn't want to automate.
There were some things where I'm like, I remember I took over, this guy was out sick for like,
I don't know, he's on vacation for a couple of weeks, maybe a month, but at least a couple
of weeks.
And I was to take over what he was doing.
And so he gives me the training on what he was doing.
And it was updating spreadsheets about, I don't know, 400 rolls with probably twice as
many fields as that was what he did every day and he did all this manually.
And it was on a, it was on a unique system and he ran some, I don't know, he ran some
commands or maybe strips, scripts and he got, um, actually I don't know, his combination
of looking things up, but he was copying them manually from, from there to the spreadsheet.
And he did this every day and he got done around newly, maybe a little bit later.
So and he started, I don't know, he started at least at eight, so four hours a day, this
is what he was doing.
And so I'm like, holy smokes.
So I did it that way a couple of days and then I was going insane.
Oh, I did it that way a couple of days and then I found out that I made some mistakes.
Well, I'm human, you know, we sometimes make mistakes and if you know me, as the geochos
my grandma used to say, so anyway, when it comes, you know, to air is human to forgive
his divine and where as far as you're concerned, you're super human.
So I thought that was pretty funny.
The point being I make probably more mistakes than most people.
So what I found out was I sent out the, and this, he sent it out in the form of like
a report spreadsheet that he sent out to everybody because apparently people have worked
with being driven by this.
And so I sent it out the first day and like within five minutes it came back, hey, this
is all wrong and I'm like, holy smokes, sure enough, I copied something wrong and screwed
up a formula and I was able to fix it because I had backups to see, you know, what I messed
up.
So that happened a couple of days in a row and I'm like, okay, this has got to change.
So then I took some time and I scripted it so where I created the spreadsheet instead
of, you know, I created it and some of the things that he was doing totally manually and not
just cutting and pasting, but where he's actually deriving values before cutting and pasting.
So I scripted the whole thing and then on the, I don't know, sometime that week, I don't
remember exactly what day, but sometime that first week I had it or it was all, the spreadsheet
itself was totally generated.
And then imagine this, I didn't have any, when I sent it out, I didn't have any, you
know, any reports that it was wrong and so after that it took me like five minutes or
less a day to send it out and I sent it out and then, you know, my, the boss said, hey,
this usually doesn't come out until around noon.
How did you get it out so fast and it looks right?
Well, yeah, because I automated it and so I probably cost him his job actually because
I didn't see him too much later after that.
But the point is that, so that's just one thing and I've automated several other things,
one of the things I did on one of the companies I worked for, they would bring on like 10
or 15 seasonal folks and then I automated a lot of what they were doing and then they
didn't need those seasonal folks anymore and several other things.
And then there was a few things that I automated to where the automation didn't cut jobs
but it improved quality and saved mistakes.
It could have eliminated jobs that I'm not really aware of, like maybe one or two jobs
weren't needed anymore because we didn't have as many errors to fix.
And so the point of that is I'm just an idiot and I definitely wasn't using a, I was just
looking at what we're currently doing and can we automate any of it easier and better to
make it easier and better than what we were doing.
And the answer is we could, we just didn't spend the time to do that previously for whatever
reason.
Now one of the things to note here is that in my mind some of this LLMs, since their
tasks, LLMs are trained only up to a certain point and the criticism about or a reason
that people would say that AI won't be taking our jobs is because AI can only do what's
already been done, it doesn't have that creative spark.
So okay sure, let's assume that's correct, maybe it really is correct, I don't know.
Like I said to me that's not the point here.
But in the case of, in that how this applies is the work that we're already doing that
just isn't automated very well.
That has been done so, and so it shouldn't know about that.
So that can probably already help us automate things and if we automate things again, maybe
it's a significant enough percentage of jobs will be not needed that we again have, you
know, pile on to our existential crisis.
And so like I said, I saw that just, I mean I've probably been, my project had probably
amounted to the elimination of about a hundred jobs over my career.
And like I said, I'm just one idiot who is doing low-hanging fruit and probably most
of you listening to this could say, hey, oh that's what you're doing, well why aren't
you using, I don't know, Python or Pearl or Bash or, for heaven's sakes, you can even
do that with cobalt, okay, maybe not too many of you are saying that last one.
But like I said, you would be able to look at a lot of the stuff that I've done and say,
oh yeah, surely we could do that, we can automate that.
So in now we get these LLMs that make that process faster and better and even if it,
like even if it doesn't amount to actual AI, I think it's going to help us be able to
automate things and do things faster with fewer people.
And so those amongst us who are doing, you know, we're not really, we're just kind of
going to work, we're not really doing the, you know, the mental heavy lifting.
Those people I think are really, their jobs are really at risk and so yeah, to me it's
not very far, it's not really hard to see, that's something that's kind of related to it.
If you look back some iterations on where quote unquote AI was, it's really been making
leaps and bounds and so I don't understand why people, so many otherwise smart people
and leaders in the industry can imagine a couple, you know, iterations down the road to
where, yeah, we're still going to need people, but what percentage of the people are we going
to need? And I would say we don't need, you know, maybe like 5% of the current people
who are making, you know, kind of the boring wall art that people buy just so there's a
piece of wall art there, not because it's some great, you know, work of art.
I think we're probably going to need about 3% of those people that we currently have.
In this fiction writers, I don't know that to me, it's probably, like I said, this one
artist or writer, now of course she might not, you know, I only knew about her today,
so who knows, maybe she's not, maybe she's not indicative and maybe she doesn't understand
enough to judge, but seems to me like she does. If we can write short stories that are at
high school grade level now, surely we can be too far away from being able to write stories
that I would read or listen to, especially if they're almost free or maybe they are going
to be free, even if it's free with ads. I'm not against free with ads. Now, so we'll see,
I kind of hope I'm wrong because my job is supposed to be one of the easiest or one of
the ones that are on the chopping block. So I hope I'm wrong, but it sure seems to me
to be pretty obvious and you don't have to go too far into the future or look too far in the past
to see the progress is being made. See, is there any other related points that I thought worth making?
No, those are the main ones. Okay, I'm going to end this.
You have been listening to Hacker Public Radio at Hacker Public Radio. Today's show was
contributed by a HBR listener like yourself. If you ever thought of recording a podcast,
you click on our contribute link to find out how easy it leads. Hosting for HBR has been
kindly provided by an onsthost.com, the internet archive and our syncs.net. On the Sadois
status, today's show is released on our Creative Commons Attribution 4.0 International License.