[277392] in TECHWR-L
[TECHWR-L] - Re: Differences in perception of the consequences large
daemon@ATHENA.MIT.EDU (sereshk 146)
Wed Oct 29 11:28:07 2025
Date: Wed, 29 Oct 2025 08:15:03 -0700 (PDT)
From: sereshk 146 <sereshk146@gmail.com>
To: TECHWR-L Email Discussion group <techwr-l@techwr-l.com>
Cc: Doug Cuff <dgcuff@gmail.com>
Message-Id: <d4e2f42c-47a1-4202-9b44-a5b0c8ff0844n@techwr-l.com>
In-Reply-To: <5298a08e-e896-4fd8-adf1-16f6ce549dddn@techwr-l.com>
MIME-Version: 1.0
Content-Type: multipart/mixed;
boundary="----=_Part_91804_242916126.1761750903444"
Reply-To: techwr-l@techwr-l.com
------=_Part_91804_242916126.1761750903444
Content-Type: multipart/alternative;
boundary="----=_Part_91805_1940202346.1761750903444"
------=_Part_91805_1940202346.1761750903444
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Thanks for sharing this!
I couldn't agree more - almost: Except, I am a full-fledged tech writer,=20
happy about it and, I know a thing or two about AI - and one thing is very=
=20
true about it in programming, writing or anywhere else you want to use it:
* It should fit the purpose as per programming of its specific algorithms;=
=20
for complex tasks that can become pretty costly, i.e. expensive...
* It should always be clear, where the data it was programmed and trained=
=20
on comes from.
* Results of any output should be checked carefully.
Therefore, no, tech writers won't be a 'yesterday's model', not by a long=
=20
chalk. ;) Same way developers won't:=20
You need human and actual brains behind the work.
AI hasn't got any... it's got algorithms that look for patterns, basically,=
=20
and either re-assemble their findings from new data in a likely manner - or=
=20
make predictions.
What *can* make them appear human is the fact that similar language models=
=20
are used to make them 'talk', printing it in an output box, or using voice=
=20
generators to make it appear to be a human voice...
And I am also testing and using AI for special projects I wouldn't finish=
=20
on my own because complex CSS/HTML-coding is involved. It wouldn't probably=
=20
be done in the first place without AI, because as so often is the case in=
=20
any business, resources are limited.=20
When you know how to 'prompt' it (in more senses then one :) ) it can be=20
helpful.
But one should look out for cost and effect (in terms of effort as well as=
=20
power consumption...)
No, humans will not 'go out of fashion'...=20
AI can become pretty dangerous when in the wrong hands; but it is a tool=20
and machine, first and last and therefore cannot replace humans.=20
At the beginning and end it is humans, any day.=20
Happy 'prompting'!
Kind Regards
Nina
On Wednesday, 29 October 2025 at 15:56:37 UTC+1 Doug Cuff wrote:
> I meant to close with the sentence "But I'm quite a bit happier that I'm=
=20
> NOT starting out as a software developer."
>
> Mea culpa. I left out the not. Again. At least I didn't type "now" when I=
=20
> meant "not."
>
> Doug
>
> On Wednesday, 29 October 2025 at 10:27:30 UTC-4 Doug Cuff wrote:
>
>> Long before ChatGPT was a gleam in Sam Altman's eye, the stance of most=
=20
>> people including managers was "Anyone can write!" Now that AI is at our=
=20
>> throats, the rats are deserting a ship that they perceive to be sinking.=
=20
>> Who needs technical writers?
>>
>> I'm going to try to remain amused as long as I can, especially after a=
=20
>> week in which I learned that software developers are more easily replace=
d=20
>> by AI than tech writers are.
>>
>> One big idea behind BASIC was "Anyone can program!" And that's where I=
=20
>> started. I dabbled with BASIC, with perl, with Inform, with PHP, with SQ=
L=20
>> queries, and with Python. Not enough to do it for a living, but enough t=
o=20
>> grasp the portable concepts.
>>
>> So this week and last, I asked a LLM platform (I won't specify) to code=
=20
>> me up:
>>
>> - a fix for PowerPoint slides counting up (5) instead of down (2 slides=
=20
>> remaining)
>> - a Pandoc filter to convert Word tables to Sphinx's list-tables (becaus=
e=20
>> my Pandoc doesn't support the --list-tables argument)
>>
>> It went fast. It went faster because I knew enough about programming to=
=20
>> spot some incorrect assumptions and fix bugs, but it still went fast. It=
=20
>> went slower because I kept adding robustness checks and because I caved =
to=20
>> endless suggestions to improve, but it still went fast.
>>
>> Sure, I'm happy I'm not twenty-somethin' years old and starting out as a=
=20
>> technical writer. But I'm quite a bit happier that I'm starting out as a=
=20
>> software developer.
>>
>
--=20
TechWR-L is the =E2=80=9CTechnical Writers List=E2=80=9D =E2=80=94 an email=
discussion list focused on all issues in technical communicatio
---=20
You received this message because you are subscribed to the Google Groups "=
TECHWR-L Email Discussion group" group.
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to techwr-l+unsubscribe@techwr-l.com.
To view this discussion visit https://groups.google.com/a/techwr-l.com/d/ms=
gid/techwr-l/d4e2f42c-47a1-4202-9b44-a5b0c8ff0844n%40techwr-l.com.
------=_Part_91805_1940202346.1761750903444
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
<div>Thanks for sharing this!</div><div>I couldn't agree more - almost: Exc=
ept, I am a full-fledged tech writer, happy about it and, I know a thing or=
two about AI - and one thing is very true about it in programming, writing=
or anywhere else you want to use it:<br />* It should fit the purpose as p=
er programming of its specific algorithms; for complex tasks that can becom=
e pretty costly, i.e. expensive...</div><div>* It should always be clear, w=
here the data it was programmed and trained on comes from.</div><div>* Resu=
lts of any output should be checked carefully.<br /><br /></div><div>Theref=
ore, no, tech writers won't be a 'yesterday's model', not by a long chalk. =
;) Same way developers won't:=C2=A0<br />You need human and actual brains b=
ehind the work.</div><div>AI hasn't got any... it's got algorithms that loo=
k for patterns, basically, and either re-assemble their findings from new d=
ata in a likely manner - or make predictions.</div><div>What *can* make the=
m appear human is the fact that similar language models are used to make th=
em 'talk', printing it in an output box, or using voice generators to make =
it appear to be a human voice...</div><div><br /></div><div>And I am also t=
esting and using AI for special projects I wouldn't finish on my own becaus=
e complex CSS/HTML-coding is involved. It wouldn't probably be done in the =
first place without AI, because as so often is the case in any business, re=
sources are limited.=C2=A0</div><div>When you know how to 'prompt' it (in m=
ore senses then one :) ) it can be helpful.</div><div>But one should look o=
ut for cost and effect (in terms of effort as well as power consumption...)=
</div><div><br /></div><div>No, humans will not 'go out of fashion'...=C2=
=A0</div><div>AI can become pretty dangerous when in the wrong hands; but i=
t is a tool and machine, first and last and therefore cannot replace humans=
.=C2=A0</div><div>At the beginning and end it is humans, any day.=C2=A0</di=
v><div><br /></div><div>Happy 'prompting'!</div><div><br /></div><div>Kind =
Regards</div><div>Nina</div><div class=3D"gmail_quote"><div dir=3D"auto" cl=
ass=3D"gmail_attr">On Wednesday, 29 October 2025 at 15:56:37 UTC+1 Doug Cuf=
f wrote:<br/></div><blockquote class=3D"gmail_quote" style=3D"margin: 0 0 0=
0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><div=
>I meant to close with the sentence "But I'm quite a bit happier t=
hat I'm NOT starting out as a software developer."</div><div><br><=
/div><div>Mea culpa. I left out the not. Again. At least I didn't type =
"now" when I meant "not."</div><div><br></div><div>Doug=
</div><div><br></div><div class=3D"gmail_quote"><div dir=3D"auto" class=3D"=
gmail_attr">On Wednesday, 29 October 2025 at 10:27:30 UTC-4 Doug Cuff wrote=
:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 0.8ex;bo=
rder-left:1px solid rgb(204,204,204);padding-left:1ex"><div>Long before Cha=
tGPT was a gleam in Sam Altman's eye, the stance of most people includi=
ng managers was "Anyone can write!" Now that AI is at our throats=
, the rats are deserting a ship that they perceive to be sinking. Who needs=
technical writers?</div><div><br></div><div>I'm going to try to remain=
amused as long as I can, especially after a week in which I learned that s=
oftware developers are more easily replaced by AI than tech writers are.</d=
iv><div><br></div><div>One big idea behind BASIC was "Anyone can progr=
am!" And that's where I started. I dabbled with BASIC, with perl, =
with Inform, with PHP, with SQL queries, and with Python. Not enough to do =
it for a living, but enough to grasp the portable concepts.</div><div><br><=
/div><div>So this week and last, I asked a LLM platform (I won't specif=
y) to code me up:</div><div><br></div><div>- a fix for PowerPoint slides co=
unting up (5) instead of down (2 slides remaining)</div><div>- a Pandoc fil=
ter to convert Word tables to Sphinx's list-tables (because my Pandoc d=
oesn't support the --list-tables argument)</div><div><br></div><div>It =
went fast. It went faster because I knew enough about programming to spot s=
ome incorrect assumptions and fix bugs, but it still went fast. It went slo=
wer because I kept adding robustness checks and because I caved to endless =
suggestions to improve, but it still went fast.</div><div><br></div><div>Su=
re, I'm happy I'm not twenty-somethin' years old and starting o=
ut as a technical writer. But I'm quite a bit happier that I'm star=
ting out as a software developer.</div></blockquote></div></blockquote></di=
v>
<p></p>
-- <br />
TechWR-L is the =E2=80=9CTechnical Writers List=E2=80=9D =E2=80=94 an email=
discussion list focused on all issues in technical communicatio<br />
--- <br />
You received this message because you are subscribed to the Google Groups &=
quot;TECHWR-L Email Discussion group" group.<br />
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to <a href=3D"mailto:techwr-l+unsubscribe@techwr-l.com">techwr-l+unsub=
scribe@techwr-l.com</a>.<br />
To view this discussion visit <a href=3D"https://groups.google.com/a/techwr=
-l.com/d/msgid/techwr-l/d4e2f42c-47a1-4202-9b44-a5b0c8ff0844n%40techwr-l.co=
m?utm_medium=3Demail&utm_source=3Dfooter">https://groups.google.com/a/techw=
r-l.com/d/msgid/techwr-l/d4e2f42c-47a1-4202-9b44-a5b0c8ff0844n%40techwr-l.c=
om</a>.<br />
------=_Part_91805_1940202346.1761750903444--
------=_Part_91804_242916126.1761750903444--
