Re: Just take me out back and shoot me...

Ok.  I'll jump in.

The very first interaction with chatGPT I had (back in december 2022) was
the following.

Write me a (Python) program that traverses the whole file system and
inserts a random number of tabs or spaces in any text file with a given
extension, say '.py'.

After a few iterations (it took a while), it did a decent enough job.

Then I asked: now run it on your code base.

A lengthier exchange ensued, where chatGPT refused, politely, but
resolutely.

I concluded that chatGPT somewhat respects Asimov's third.

A more recent example with DeepSeek was about ELisp.  Here is where I talk
about it:
https://within-parens.blogspot.com/2025/04/emacs-lisp-programming-with-deepseek.html

Cheers

MA


On Fri, Aug 8, 2025 at 9:52 AM Yuri Davidovsky (as work at disclosure dot
ie) <[email protected]> wrote:

>
>
> > On 7 Aug 2025, at 22:17, Eben Bruyns (as eben at sdk dot co dot nz) <
> [email protected]> wrote:
> >
> > I think there's certainly value in these tools, but I don't think the
> value is as much as the industry wants you to believe.
>
> I find LLMs are very valuable for me as a learning tool (smart search
> engine, basically). I also use it occasionally for boilerplate generation
> (give me a class describing a wav file structure) and in these two aspects
> these tools are well worth the money for my use case. However as dbm noted,
> you have to know what you are asking, otherwise you won’t be able to filter
> bullshit out.
>
> These systems are not good coders (and probably won’t be in the near
> future) because programming is not about being able to churn lines of code
> out but rather being able to model a specific problem in a mental abstract
> space (your head, basically), and then being able to express this model in
> computer code. Programming is not about writing code, but about modeling
> problems.
>
> Equating programming with putting out lines of text is the same as
> equating bicycle riding with turning pedals — this is the easiest and most
> boring part of the entire process, but not the process itself.
>
> _______________________________________________
> Lisp Hug - the mailing list for LispWorks users
> [email protected]
> http://www.lispworks.com/support/lisp-hug.html
>


-- 
Marco Antoniotti, Professor, Director         tel. +39 - 02 64 48 79 01
DISCo, University of Milan-Bicocca U14 2043   http://dcb.disco.unimib.it
Viale Sarca 336
I-20126 Milan (MI) ITALY

REGAINS: https://regains.disco.unimib.it/