0:00
Welcome to the technical writing success podcast from Kurt Robbins, where we help you get
0:04
smarter than your competition.
0:06
Higher Kurt to coach you or your employees in AI to avoid a pink slip or having your
0:11
competition eat your lunch.
0:13
This is episode 192 and I am Fred Jones.
0:16
And I'm definitely Blake.
0:18
Imagine stepping into a new role and you're just handed this massive, highly technical
0:25
The dreaded giant binder.
0:27
You're trying to figure out a critical system integration that your entire project depends
0:31
You're reading the steps following the logic and slowly this creeping realization washes
0:38
That the person who wrote the manual has clearly never actually used the software.
0:42
I mean, they're using the right nouns and verbs, but the sequence is just totally divorced
0:46
from the reality of the interface.
0:48
It reads like a translation of a translation.
0:50
You can almost feel the writer guessing at how the system behaves under load, you know,
0:54
just hoping the engineering specs they copied actually map onto the real world.
1:00
So this episode reviews a March 16 interview with senior technical writer Robert Delwood
1:05
by show producer Kurt Robbins.
1:07
We are exploring exactly why that happens and how the tech industry is trying to fix it.
1:13
Because Delwood represents a highly specialized and honestly somewhat rare solution to
1:17
this problem of disconnected documentation.
1:20
He operates under a title that wasn't designed in a boardroom but evolved out of pure necessity.
1:28
Our goal today is to really unpack this hybrid role for you listening.
1:32
We'll look at how it functions day to day, synthesize Delwood's five rules for bulletproof
1:36
documentation, and look at how AI is actually being used in the trenches right now.
1:42
Cutting through a lot of the executive hype and to understand the program writer, we really
1:46
have to start with Delwood's origin story because it perfectly shows why this dual skill
1:52
He brings this fascinating collision of backgrounds.
1:54
He's been a hobbyist programmer since the seventh grade.
2:00
Someone who fundamentally enjoys the logic of coding.
2:02
But his formal academic education is actually in journalism.
2:06
He was trained to investigate and report.
2:08
Which is such a rare combo.
2:10
Usually the industry just completely silos those two mindsets.
2:13
You either have the hardcore developer who resents translating their elegant code into plain
2:19
English, or you have the eloquent writer who, you know, breaks out in a cold sweat.
2:24
The moment a command line interface throws an error.
2:28
But right out of school, Delwood lands a job as a technical writer at NASA down in Houston.
2:33
But the environment was apparently just total chaos.
2:36
The building he was assigned to wasn't even finished yet.
2:38
They lacked basic equipment, administrative tools, just any sort of established workflow.
2:43
So faced with that gridlock, he essentially hacks his own environment.
2:47
He sits down at his MS-DOS PC, fires up basic, and codes a custom scheduling tool just
2:53
to manage his own workflow.
2:55
He literally built the infrastructure he needed to survive the job.
2:58
And his manager notices this proactive approach within six months.
3:03
He's promoted from technical writer to developer.
3:06
But he never abandons his journalism instinct.
3:09
Even as a developer, he relentlessly writes the documentation for his own apps.
3:14
And that combination caught the attention of the broader industry.
3:17
A friend at Microsoft eventually recruited him for a brand new rare position they desperately
3:22
needed to fill a writer who could read and write code.
3:26
The programmer writer, which he says is still his favorite title because there's zero ambiguity.
3:33
To make this tangible for you listening, think of a purely non-technical writer documenting
3:38
code as like a restaurant critic who has never cooked a meal.
3:43
Oh, that's a great way to look at it.
3:45
They can tell you the sauce broke, but they have no idea why the sauce broke in the kitchen.
3:49
Delwood is the chef who stepped off the line to write the review.
3:52
He knows the underlying mechanics.
3:54
And the tech industry pays a premium for that chef's perspective because it solves a massive
3:58
logistical bottleneck, which is the empathy gap between writers and engineers.
4:02
They speed the jargon.
4:04
When Delwood works in an office, he intentional requests the desk seated right with the
4:08
developers to dissolve any separation.
4:11
As he reads source code directly, he doesn't have to constantly interrupt a developer's
4:16
flow state to ask questions.
4:17
He just opens the repository and finds the answer himself.
4:20
Boy, I have a pushback here.
4:23
If he's hired to write, isn't spending time coding custom tools just, well, doubling
4:29
his workload and wasting time?
4:31
It sounds counterintuitive, I know.
4:33
But writing those tools is the ultimate efficiency hack.
4:36
Think about the alternative.
4:38
If writers need a custom script to automate formatting, they have to submit a ticket to
4:44
The endless juror backlog.
4:46
It gets prioritized behind critical bugs.
4:48
They might wait six weeks for a simple tool.
4:51
By coding it himself in an afternoon, Delwood completely unblocks his team.
4:56
That's why he insists all tech writers today should learn some programming, even just Python
5:02
Let's take a brief break for a special message from our producer, Kurt Robbins.
5:07
This is Kurt Robbins.
5:08
Thanks for listening.
5:09
I truly appreciate your support.
5:11
I want to let you know that I'm currently accepting new clients.
5:14
My rates are affordable and I have more than 25 years of experience working for enterprise
5:19
companies like Microsoft, Northrop Grumman, Oracle, PNC Bank, FedEx, USA and Wells Fargo,
5:28
If you want to improve your IT documentation and communications, hire me.
5:32
I deliver fast, know how to use AI to improve efficiency and accuracy and love going the
5:37
extra mile to satisfy my clients.
5:40
Thank you for subscribing and listening.
5:42
Back to you, Daphne and Fred.
5:44
Welcome back to the Technical Writing Success Podcast, where we help you get smarter than
5:47
your competition by coaching you in AI.
5:50
So on the flip side of that, if it's so helpful to know how to code, why doesn't the industry
5:55
just have the developers write the manuals?
5:58
They wrote the code, right?
5:59
Well, Delwood is adamant that relying on developers for docs is a total recipe for disaster.
6:06
The biggest issue is the curse of knowledge.
6:09
Developers have such an intricate mental model of the system that they are often fundamentally
6:13
incapable of explaining it to a novice.
6:16
They skip steps because those steps are invisible to them.
6:19
Like trying to explain to a relative how to change a Wi-Fi password over the phone, you
6:23
say, type the IP address.
6:25
Forgetting they don't even know what an IP address is.
6:28
Now, multiply that blind spot by 10,000 lines of enterprise code.
6:30
Plus, there's a brutal economic reality.
6:33
Developers are expensive.
6:35
Making a senior engineer wrestle with a user manual is just a profound misallocation
6:41
It distracts them from their primary mandate, which is writing and optimizing code.
6:46
Delwood even references this classic Three Stooges routine, where Mohan's curly a hammer
6:50
and says, when I nod my head, you hit it.
6:52
And curly hits Mo's head.
6:56
But an API documentation, that kind of ambiguity crashes client servers.
7:02
Which brings us to his operational philosophy.
7:04
Since he started specializing in API docs in 2003, he's formulated five rules for bullet
7:10
proof documentation.
7:11
First, learn a language, broaden how you learned videos, trial, and error to understand
7:16
how developers consume info.
7:18
And rule two is make each call.
7:21
You have to personally test the documentation.
7:23
He actually shared this brilliant anecdote where he found a process ID value was missing
7:27
from an API's return call.
7:29
And QA had totally missed it.
7:31
The in-house devs just knew was supposed to be there so they hallucinated the correct
7:35
Delwood only caught it because he actually executed the call himself as an outsider.
7:39
Which leads to rule three, document it for yourself.
7:43
If he doesn't understand it, he stops and figures it out.
7:46
He acts as an experienced developer who just happens to be ignorant of this specific
7:51
It's a defensive mechanism, really.
7:52
If you document the pain points, you don't have to figure them out again in six months.
7:56
And that transitions right into rule four, which is dog fooding, using your own documentation
8:02
Persistent editing is a virtue for writers, not a weakness.
8:05
It's like an architect being forced to live in the house they just designed for a month.
8:09
It's the only way you'll notice where the drafts are, where if a door opens the wrong
8:14
And his final rule, rule five, don't focus on tools, focus on content.
8:19
Not waste time arguing about Windows versus Mac or Markdown editors, write apps, make
8:24
calls and postman obsessively, and extract maximum detail.
8:29
That deep extraction of truth is the perfect lens for our final topic.
8:33
Because we're in 2026, and we have to talk about artificial intelligence.
8:37
If you read the executive summaries, you'd think AI has completely automated the programmer
8:43
But Delwood provides a massive reality check here.
8:47
He actively uses AI for generating complex algorithms, grammar checking, mapping out
8:54
He said AI was nearly spot on for a recent workflow.
8:57
But there is a massive line he will not cross.
9:00
He never uses AI to write the actual content of the documentation.
9:04
Because the reality of AI coding is that it rarely compiles on the first try.
9:08
It uses the wrong libraries, it misses values, it saves typing, but it absolutely requires
9:13
a human touch to actually get it working.
9:15
If an AI is generating a routine but getting the libraries wrong, aren't we just shifting
9:20
the technical writer's job?
9:21
From being a creator to being a glorified AI babysitter and debugger.
9:26
That tension is a huge issue in the industry right now.
9:28
We're seeing CEOs laying off thousands of team members to save money, explicitly citing
9:35
And in partially reporting Delwood's perspective here, he believes the real threat to documentation
9:40
quality isn't the AI tech itself, it's those executive cost-cutting decisions.
9:44
Yeah, he had a funny quip about that.
9:46
He said true equality will be reached when CEOs are laid off because of AI, noting he's
9:52
seen instances where a magic eight ball could basically do a CEO's job.
9:56
That's hilarious, but it highlights a structural reality.
10:00
AI only knows what has already happened, but technical writers operate at the bleeding
10:06
They're documenting quirks that were invented hours ago.
10:08
The AI literally doesn't have the context yet.
10:11
Maybe a human with empathy and a testing environment can translate that behavior.
10:15
Which brings us full circle.
10:16
We went from a makeshift DOS tool at NASA to the empathy of the programmer writer through
10:21
the five pillars of API docs and finally to a grounded view of AI.
10:26
The big takeaway for you listening, whether you're learning Python or managing a team,
10:29
is the principle of dog fooding.
10:31
Testing your own explanations is universally applicable.
10:34
Empathy requires friction.
10:37
And here's a final thought to mull over.
10:39
As tools get better at generating massive amounts of information, maybe the most valuable
10:43
human skill of the future won't be creating content.
10:46
It might be the ability to deeply understand a complex system and explain it simply to
10:50
another human being.
10:51
Acting is the ultimate empathetic barrier between human intent and machine execution.
10:57
Thank you for listening to the technical writing success podcast from Kurt Robbins,
11:01
where we help you get smarter than your competition.
11:03
Hire us to coach you and your employees in AI to future-proof your career or company.
11:08
Subscribe now to never miss a career-saving episode.