Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.misc > #93752 > unrolled thread

Personal agents on Linux: what should stay under the user's control?

Back to article view | Back to comp.os.linux.misc

Contents

  Personal agents on Linux: what should stay under the user's control? Martin Davies-Roundhill <enduphere.xyz@gmail.com> - 2026-10-15 14:52 +0000
    Re: Personal agents on Linux: what should stay under the user's control? John Ames <commodorejohn@gmail.com> - 2026-10-15 07:57 -0700
      Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-15 19:33 +0100
        Re: Personal agents on Linux: what should stay under the user's control? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2026-10-15 19:40 +0000
    Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-15 17:01 +0200
    Re: Personal agents on Linux: what should stay under the user's control? Beej Jorgensen <beej@beej.us> - 2026-10-15 17:29 +0000
    Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-15 19:27 +0100
      Re: Personal agents on Linux: what should stay under the user's control? rbowman <bowman@montana.com> - 2026-10-15 18:47 +0000
        Re: Personal agents on Linux: what should stay under the user's control? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2026-10-15 19:40 +0000
          Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-15 23:35 +0200
            Re: Personal agents on Linux: what should stay under the user's control? rbowman <bowman@montana.com> - 2026-10-16 01:41 +0000
              Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-16 10:47 +0100
                Re: Personal agents on Linux: what should stay under the user's control? rbowman <bowman@montana.com> - 2026-10-16 18:40 +0000
                  Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-16 23:17 +0100
                    Re: Personal agents on Linux: what should stay under the user's control? rbowman <bowman@montana.com> - 2026-10-17 02:53 +0000
                      Re: Personal agents on Linux: what should stay under the user's control? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2026-10-17 18:33 +0000
            Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-16 10:46 +0100
            Re: Personal agents on Linux: what should stay under the user's control? John Ames <commodorejohn@gmail.com> - 2026-10-16 07:58 -0700
              Re: Personal agents on Linux: what should stay under the user's control? Leroy H <lh@somewhere.net> - 2026-10-16 15:59 +0000
              Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-16 18:46 +0200
                Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-16 18:32 +0100
                  Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-16 19:41 +0200
                    Re: Personal agents on Linux: what should stay under the user's control? rbowman <bowman@montana.com> - 2026-10-16 18:26 +0000
                      Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-16 22:40 +0200
                    Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-16 23:11 +0100
                      Re: Personal agents on Linux: what should stay under the user's control? rbowman <bowman@montana.com> - 2026-10-17 03:06 +0000
                      Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-17 12:24 +0200
                        Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-17 11:31 +0100
              Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-16 18:46 +0200
                Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-16 18:35 +0100
                  Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-16 19:43 +0200
                    Re: Personal agents on Linux: what should stay under the user's control? rbowman <bowman@montana.com> - 2026-10-16 18:34 +0000
                      Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-16 22:47 +0200
                        Re: Personal agents on Linux: what should stay under the user's control? rbowman <bowman@montana.com> - 2026-10-17 02:40 +0000
                          Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-17 12:33 +0200
                            Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-17 12:04 +0100
                              Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-17 13:40 +0200
                            Re: Personal agents on Linux: what should stay under the user's control? Stéphane CARPENTIER <sc@fiat-linux.fr> - 2026-10-17 11:56 +0000
                              Re: Personal agents on Linux: what should stay under the user's control? John Bokma <contact@johnbokma.com> - 2026-10-17 13:59 +0200
                                Programming Books (was Re: Personal agents on Linux: what should stay under the user's control?) John Bokma <contact@johnbokma.com> - 2026-10-17 15:01 +0200
                                  Re: Programming Books (was Re: Personal agents on Linux: what should stay under the user's control?) rbowman <bowman@montana.com> - 2026-10-17 18:26 +0000
          Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-16 10:45 +0100
            Re: Personal agents on Linux: what should stay under the user's control? "Carlos E.R." <robin_listas@es.invalid> - 2026-10-16 13:56 +0200
              Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-16 15:38 +0100
            Re: Personal agents on Linux: what should stay under the user's control? Rich <rich@example.invalid> - 2026-10-16 15:19 +0000
              Re: Personal agents on Linux: what should stay under the user's control? The Natural Philosopher <tnp@invalid.invalid> - 2026-10-16 17:09 +0100
                Re: Personal agents on Linux: what should stay under the user's control? rbowman <bowman@montana.com> - 2026-10-16 18:21 +0000

Page 1 of 3  [1] 2 3  Next page →

#93752 — Personal agents on Linux: what should stay under the user's control?

Hi all,

I have been thinking about the jump from chatbot to personal agent. Once software can take actions, the question is not only which model is most capable. It is also where it runs, what it remembers, whether I can inspect and stop what it is doing, and whether I can move to another tool without losing my context.

For Linux users, which matters most: local execution, open components, portable memory, an auditable action log, or simply being able to replace the service? Is there a sensible balance between a more capable hosted model and a less capable system that stays under the user's control?

I made a short, non-commercial video essay about Peter Steinberger's OpenClaw and the possibility that personal-agent tools could renew interest in open-source and practical computing. It is one perspective, not evidence that OpenClaw solves those control questions:
https://youtu.be/TuDJWsdj38Y

I'm Martin Davies-Roundhill. This is a non-commercial project: no product, fundraising, paid offer, or request for money. I'm sharing this episode because it connects with the Linux/open-source question. I'm interested in disagreement and concrete experience. Would you run an agent locally if the capability were weaker than a hosted one, or is that trade-off not worthwhile yet?

[toc] | [next] | [standalone]

#93753

*Everything* stays under my control if, y'know, I just don't get it
into my head to hand the keys to my system to a stochastic parrot.

[toc] | [prev] | [next] | [standalone]

#93768

On 15/10/2026 15:57, John Ames wrote:
> *Everything* stays under my control if, y'know, I just don't get it
> into my head to hand the keys to my system to a stochastic parrot.
> 
Stochastic parrot? Douglas Adams would have loved that. Better than a 
babel fish.

Somewhere I have a carton strip featuring Polly the Cosmic 
Halluco-parrot. An identity that appears to those who have take far too 
much (or just enough) LSD.

https://www.redbubble.com/i/sticker/Polly-the-cosmic-Halluco-parrot-by-Bryan-Talbot/171496343/7sgk

I really liked that cartoon. Picked it up at a rock festival sometime in 
the 70s

Brainstorm comix.

-- 
"The great thing about Glasgow is that if there's a nuclear attack it'll 
look exactly the same afterwards."

Billy Connolly

[toc] | [prev] | [next] | [standalone]

#93779

On 2026-10-15, The Natural Philosopher <tnp@invalid.invalid> wrote:

> On 15/10/2026 15:57, John Ames wrote:
>
>> *Everything* stays under my control if, y'know, I just don't get it
>> into my head to hand the keys to my system to a stochastic parrot.
> 
> Stochastic parrot? Douglas Adams would have loved that. Better than a 
> babel fish.
>
> Somewhere I have a carton strip featuring Polly the Cosmic 
> Halluco-parrot. An identity that appears to those who have take far too 
> much (or just enough) LSD.
>
> https://www.redbubble.com/i/sticker/Polly-the-cosmic-Halluco-parrot-by-Bryan-Talbot/171496343/7sgk
>
> I really liked that cartoon. Picked it up at a rock festival sometime in 
> the 70s
>
> Brainstorm comix.

Small independent comics are sure a lot of fun.  I still
have my copy of "Berford Seaman's Flabby Thighs and Butter"
("At last!  A Canadian comic you can read and not understand!").

And then there's David Boswell's series
"Reid Fleming: World's Toughest Milkman"...

-- 
/~\  Charlie Gibbs                  |  In this world there are
\ /  <cgibbs@kltpzyxm.invalid>      |  two kinds of people:
 X   I'm really at ac.dekanfrus     |  1. Those who can extrapolate
/ \  if you read it the right way.  |  from incomplete data.

[toc] | [prev] | [next] | [standalone]

#93754

On 15/10/2026 16:52, Martin Davies-Roundhill wrote:
> Hi all,
> 
> I have been thinking about the jump from chatbot to personal agent. Once software can take actions, the question is not only which model is most capable. It is also where it runs, what it remembers, whether I can inspect and stop what it is doing, and whether I can move to another tool without losing my context.
> 
> For Linux users, which matters most: local execution, open components, portable memory, an auditable action log, or simply being able to replace the service? Is there a sensible balance between a more capable hosted model and a less capable system that stays under the user's control?
> 
> I made a short, non-commercial video essay about Peter Steinberger's OpenClaw and the possibility that personal-agent tools could renew interest in open-source and practical computing. It is one perspective, not evidence that OpenClaw solves those control questions:
> https://youtu.be/TuDJWsdj38Y
> 
> I'm Martin Davies-Roundhill. This is a non-commercial project: no product, fundraising, paid offer, or request for money. I'm sharing this episode because it connects with the Linux/open-source question. I'm interested in disagreement and concrete experience. Would you run an agent locally if the capability were weaker than a hosted one, or is that trade-off not worthwhile yet?

My experience with copilot CLI on Linux is that it works quite well.

-- 
Static tumblelog generator: https://github.com/john-bokma/tumblelog/
Available as Python or Perl. Example tumblelog: https://plurrrr.com/

[toc] | [prev] | [next] | [standalone]

#93761

In article <11aqpbm$2ga9d$1@dont-email.me>,
Martin Davies-Roundhill  <enduphere.xyz@gmail.com> wrote:
>I'm Martin Davies-Roundhill. This is a non-commercial project: no
>product, fundraising, paid offer, or request for money. I'm sharing
>this episode because it connects with the Linux/open-source question.
>I'm interested in disagreement and concrete experience. Would you run
>an agent locally if the capability were weaker than a hosted one, or is
>that trade-off not worthwhile yet?

Are you classified as human?

(Just checking, since you posted a very, very similarly boilerplatey
message and video to rec.bicycles.tech. And I just finished reading _Do
Androids Dream of Electric Sheep_.)

-- 
Brian "Beej Jorgensen" Hall | beej@beej.us

[toc] | [prev] | [next] | [standalone]

#93767

On 15/10/2026 15:52, Martin Davies-Roundhill wrote:
> For Linux users, which matters most: local execution, open components,
> portable memory, an auditable action log, or simply being able to
> replace the service?

What matters most is not having it at all.

-- 
"The great thing about Glasgow is that if there's a nuclear attack it'll 
look exactly the same afterwards."

Billy Connolly

[toc] | [prev] | [next] | [standalone]

#93772

On Thu, 15 Oct 2026 19:27:55 +0100, The Natural Philosopher wrote:

> On 15/10/2026 15:52, Martin Davies-Roundhill wrote:
>> For Linux users, which matters most: local execution, open components,
>> portable memory, an auditable action log, or simply being able to
>> replace the service?
> 
> What matters most is not having it at all.

I've lived this long without a butler, valet, or personal agent. I think I 
can make it for a few more years. I did play with a Python interface but 
didn't get further than the point where the chatbot wanted money, or at 
least a registration. 

[toc] | [prev] | [next] | [standalone]

#93778

On 2026-10-15, rbowman <bowman@montana.com> wrote:

> On Thu, 15 Oct 2026 19:27:55 +0100, The Natural Philosopher wrote:
>
>> On 15/10/2026 15:52, Martin Davies-Roundhill wrote:
>>
>>> For Linux users, which matters most: local execution, open components,
>>> portable memory, an auditable action log, or simply being able to
>>> replace the service?
>> 
>> What matters most is not having it at all.
>
> I've lived this long without a butler, valet, or personal agent. I think I 
> can make it for a few more years. I did play with a Python interface but 
> didn't get further than the point where the chatbot wanted money, or at 
> least a registration. 

Hear, hear.  I find it amazing that people can find things essential
to life itself that didn't even exist 30 years ago.  How ever did
human civilization last this long?

Actually, the question should probably not be how did we survive
without these things, but how long we're going to survive with them.

-- 
/~\  Charlie Gibbs                  |  In this world there are
\ /  <cgibbs@kltpzyxm.invalid>      |  two kinds of people:
 X   I'm really at ac.dekanfrus     |  1. Those who can extrapolate
/ \  if you read it the right way.  |  from incomplete data.

[toc] | [prev] | [next] | [standalone]

#93786

On 15/10/2026 21:40, Charlie Gibbs wrote:
> On 2026-10-15, rbowman <bowman@montana.com> wrote:
> 
>> On Thu, 15 Oct 2026 19:27:55 +0100, The Natural Philosopher wrote:
>>
>>> On 15/10/2026 15:52, Martin Davies-Roundhill wrote:
>>>
>>>> For Linux users, which matters most: local execution, open components,
>>>> portable memory, an auditable action log, or simply being able to
>>>> replace the service?
>>>
>>> What matters most is not having it at all.
>>
>> I've lived this long without a butler, valet, or personal agent. I think I
>> can make it for a few more years. I did play with a Python interface but
>> didn't get further than the point where the chatbot wanted money, or at
>> least a registration.
> 
> Hear, hear.  I find it amazing that people can find things essential
> to life itself that didn't even exist 30 years ago.

Like syntax highlighting? I like it when computers do the boring parts 
for me ;-)

-- 
Static tumblelog generator: https://github.com/john-bokma/tumblelog/
Available as Python or Perl. Example tumblelog: https://plurrrr.com/

[toc] | [prev] | [next] | [standalone]

#93790

On Thu, 15 Oct 2026 23:35:34 +0200, John Bokma wrote:

> On 15/10/2026 21:40, Charlie Gibbs wrote:
>> On 2026-10-15, rbowman <bowman@montana.com> wrote:
>> 
>>> On Thu, 15 Oct 2026 19:27:55 +0100, The Natural Philosopher wrote:
>>>
>>>> On 15/10/2026 15:52, Martin Davies-Roundhill wrote:
>>>>
>>>>> For Linux users, which matters most: local execution, open
>>>>> components,
>>>>> portable memory, an auditable action log, or simply being able to
>>>>> replace the service?
>>>>
>>>> What matters most is not having it at all.
>>>
>>> I've lived this long without a butler, valet, or personal agent. I
>>> think I can make it for a few more years. I did play with a Python
>>> interface but didn't get further than the point where the chatbot
>>> wanted money, or at least a registration.
>> 
>> Hear, hear.  I find it amazing that people can find things essential to
>> life itself that didn't even exist 30 years ago.
> 
> Like syntax highlighting? I like it when computers do the boring parts
> for me ;-)

At least with my setup Vim highlights keywords, data types, and strings in 
C. I don't know how much that adds.  Using % to bounce between curlies is 
more useful.

[toc] | [prev] | [next] | [standalone]

#93837

On 16/10/2026 02:41, rbowman wrote:
> On Thu, 15 Oct 2026 23:35:34 +0200, John Bokma wrote:
> 
>> On 15/10/2026 21:40, Charlie Gibbs wrote:
>>> On 2026-10-15, rbowman <bowman@montana.com> wrote:
>>>
>>>> On Thu, 15 Oct 2026 19:27:55 +0100, The Natural Philosopher wrote:
>>>>
>>>>> On 15/10/2026 15:52, Martin Davies-Roundhill wrote:
>>>>>
>>>>>> For Linux users, which matters most: local execution, open
>>>>>> components,
>>>>>> portable memory, an auditable action log, or simply being able to
>>>>>> replace the service?
>>>>>
>>>>> What matters most is not having it at all.
>>>>
>>>> I've lived this long without a butler, valet, or personal agent. I
>>>> think I can make it for a few more years. I did play with a Python
>>>> interface but didn't get further than the point where the chatbot
>>>> wanted money, or at least a registration.
>>>
>>> Hear, hear.  I find it amazing that people can find things essential to
>>> life itself that didn't even exist 30 years ago.
>>
>> Like syntax highlighting? I like it when computers do the boring parts
>> for me ;-)
> 
> At least with my setup Vim highlights keywords, data types, and strings in
> C. I don't know how much that adds.  Using % to bounce between curlies is
> more useful.
> 
Oh THAT sort of highglighting

Geany does all of that. Some of it is actually useful.


-- 
Outside of a dog, a book is a man's best friend. Inside of a dog it's 
too dark to read.

Groucho Marx

[toc] | [prev] | [next] | [standalone]

#93906

On Fri, 16 Oct 2026 10:47:33 +0100, The Natural Philosopher wrote:

> On 16/10/2026 02:41, rbowman wrote:
>> On Thu, 15 Oct 2026 23:35:34 +0200, John Bokma wrote:
>> 
>>> On 15/10/2026 21:40, Charlie Gibbs wrote:
>>>> On 2026-10-15, rbowman <bowman@montana.com> wrote:
>>>>
>>>>> On Thu, 15 Oct 2026 19:27:55 +0100, The Natural Philosopher wrote:
>>>>>
>>>>>> On 15/10/2026 15:52, Martin Davies-Roundhill wrote:
>>>>>>
>>>>>>> For Linux users, which matters most: local execution, open
>>>>>>> components,
>>>>>>> portable memory, an auditable action log, or simply being able to
>>>>>>> replace the service?
>>>>>>
>>>>>> What matters most is not having it at all.
>>>>>
>>>>> I've lived this long without a butler, valet, or personal agent. I
>>>>> think I can make it for a few more years. I did play with a Python
>>>>> interface but didn't get further than the point where the chatbot
>>>>> wanted money, or at least a registration.
>>>>
>>>> Hear, hear.  I find it amazing that people can find things essential
>>>> to life itself that didn't even exist 30 years ago.
>>>
>>> Like syntax highlighting? I like it when computers do the boring parts
>>> for me ;-)
>> 
>> At least with my setup Vim highlights keywords, data types, and strings
>> in C. I don't know how much that adds.  Using % to bounce between
>> curlies is more useful.
>> 
> Oh THAT sort of highglighting
> 
> Geany does all of that. Some of it is actually useful.

What other kind is there? I use Geany on the Pi since it comes with 
Raspberry Pi OS. It isn't bad. My preference is gVim. I seldom use the GUI 
features but it doesn't eat a terminal like Vim. However Pi's DE is Gtk2 
based and vim-gtk doesn't work.

[toc] | [prev] | [next] | [standalone]

#93941

On 16/10/2026 19:40, rbowman wrote:
>> Geany does all of that. Some of it is actually useful.
> What other kind is there?

For some reason I was thinking of predictive text where it always wants 
to type exactly the wrong thing.


> I use Geany on the Pi since it comes with
> Raspberry Pi OS. It isn't bad. My preference is gVim. I seldom use the GUI
> features but it doesn't eat a terminal like Vim. However Pi's DE is Gtk2
> based and vim-gtk doesn't work.
> 
I cant remember why I ended up with geany. I used to use wordstar years 
ago, Vi is all right but not for the whole 900 pages. Emacs had a heady 
whiff of Compsci about it and the text editor in mint was sorta OK, but 
a bit 'meh'.
Someone said 'try geany' and I did. It  was good enough. I'll not say more

-- 
“Those who can make you believe absurdities, can make you commit 
atrocities.”

― Voltaire, Questions sur les Miracles à M. Claparede, Professeur de 
Théologie à Genève, par un Proposant: Ou Extrait de Diverses Lettres de 
M. de Voltaire

[toc] | [prev] | [next] | [standalone]

#93960

On Fri, 16 Oct 2026 23:17:53 +0100, The Natural Philosopher wrote:

> On 16/10/2026 19:40, rbowman wrote:
>>> Geany does all of that. Some of it is actually useful.
>> What other kind is there?
> 
> For some reason I was thinking of predictive text where it always wants
> to type exactly the wrong thing.
> 
> 
>> I use Geany on the Pi since it comes with Raspberry Pi OS. It isn't
>> bad. My preference is gVim. I seldom use the GUI features but it
>> doesn't eat a terminal like Vim. However Pi's DE is Gtk2 based and
>> vim-gtk doesn't work.
>> 
> I cant remember why I ended up with geany. I used to use wordstar years
> ago, Vi is all right but not for the whole 900 pages. Emacs had a heady
> whiff of Compsci about it and the text editor in mint was sorta OK, but
> a bit 'meh'.
> Someone said 'try geany' and I did. It  was good enough. I'll not say
> more

WordStar was bundled on a CP/M 'portable' I bought. In text mode it was 
indeed a decent programming editor compared to vi.  Note: when I say 'vi' 
I mean the original Bill Joy version. Most Linux distros symlink 'vi' to 
'vim' so many who say 'I use vi' never had the joy of using vi.

Back when storage was a concern iirc vim was about 2 MB, emacs was 23 MB. 
I don't need an editor to tell my fortune and I am allergic to lisp.

When I moved to x86 DOS I bought Brief. It was designed as a programming 
editor and quite good. Sadly Borland bought and buried it.

[toc] | [prev] | [next] | [standalone]

#94055

On 2026-10-17, rbowman <bowman@montana.com> wrote:

> WordStar was bundled on a CP/M 'portable' I bought. In text mode it was 
> indeed a decent programming editor compared to vi.  Note: when I say 'vi' 
> I mean the original Bill Joy version. Most Linux distros symlink 'vi' to 
> 'vim' so many who say 'I use vi' never had the joy of using vi.
                                                 ^
Shouldn't that be capitalized?  :-)

-- 
/~\  Charlie Gibbs                  |  In this world there are
\ /  <cgibbs@kltpzyxm.invalid>      |  two kinds of people:
 X   I'm really at ac.dekanfrus     |  1. Those who can extrapolate
/ \  if you read it the right way.  |  from incomplete data.

[toc] | [prev] | [next] | [standalone]

#93836

On 15/10/2026 22:35, John Bokma wrote:
> On 15/10/2026 21:40, Charlie Gibbs wrote:
>> On 2026-10-15, rbowman <bowman@montana.com> wrote:
>>
>>> On Thu, 15 Oct 2026 19:27:55 +0100, The Natural Philosopher wrote:
>>>
>>>> On 15/10/2026 15:52, Martin Davies-Roundhill wrote:
>>>>
>>>>> For Linux users, which matters most: local execution, open components,
>>>>> portable memory, an auditable action log, or simply being able to
>>>>> replace the service?
>>>>
>>>> What matters most is not having it at all.
>>>
>>> I've lived this long without a butler, valet, or personal agent. I 
>>> think I
>>> can make it for a few more years. I did play with a Python interface but
>>> didn't get further than the point where the chatbot wanted money, or at
>>> least a registration.
>>
>> Hear, hear.  I find it amazing that people can find things essential
>> to life itself that didn't even exist 30 years ago.
> 
> Like syntax highlighting? I like it when computers do the boring parts 
> for me ;-)
> 
Not when they simply get it wrong

-- 
Outside of a dog, a book is a man's best friend. Inside of a dog it's 
too dark to read.

Groucho Marx

[toc] | [prev] | [next] | [standalone]

#93877

On Thu, 15 Oct 2026 23:35:34 +0200
John Bokma <contact@johnbokma.com> wrote:

> Like syntax highlighting? I like it when computers do the boring
> parts for me ;-)

You will note, though, that syntax highlighting is completely possible
to do *without* bringing nondeterministic, compute-intensive chatbots
into the equation.

[toc] | [prev] | [next] | [standalone]

#93885

On Fri, 16 Oct 2026 07:58:00 -0700, John Ames wrote:

> 
> You will note, though, that syntax highlighting is completely possible
> to do *without* bringing nondeterministic, compute-intensive chatbots
> into the equation.
>

Syntax highlighting depends in large measure on human perceptual
factors which cannot readily be quantified and hence are beyond
the ken of any AI.

Operators are more important than variables.  Reserved words have
the highest priority.  How can a color scheme reflect and rank such
importance?

To humans, some colors look awful when juxtaposed with some
other color.  How can a highlighting scheme be made pleasing
to a human?

Trying to implement an optimal color scheme is to fall into
a massive black hole.  Escape is impossible.

I just use what is given even though all are somewhat ugly
and irritating to the senses.

[toc] | [prev] | [next] | [standalone]

#93889

On 16/10/2026 16:58, John Ames wrote:
> On Thu, 15 Oct 2026 23:35:34 +0200
> John Bokma <contact@johnbokma.com> wrote:
> 
>> Like syntax highlighting? I like it when computers do the boring
>> parts for me ;-)
> 
> You will note, though, that syntax highlighting is completely possible
> to do *without* bringing nondeterministic, compute-intensive chatbots
> into the equation.

Sure. It's just an example of a feature that once was thought useless 
but of the n00bs.

-- 
Static tumblelog generator: https://github.com/john-bokma/tumblelog/
Available as Python or Perl. Example tumblelog: https://plurrrr.com/

[toc] | [prev] | [next] | [standalone]

Page 1 of 3  [1] 2 3  Next page →

Back to top | Article view | comp.os.linux.misc

csiph-web