Re: kids these days

On Fri, 4 Oct 2024 20:28:53 -0400, "Edward Rawde"
<invalid@invalid.invalid> wrote:

"Cursitor Doom" <cd@notformail.com> wrote in message news:lts0gjd4gsl84be5je9hcmen2dolt931fr@4ax.com...

On Fri, 4 Oct 2024 14:56:37 -0400, "Edward Rawde"
<invalid@invalid.invalid> wrote:
>

"Cursitor Doom" <cd@notformail.com> wrote in message news:2c80gjt42h2f04f40i1i3n05j2pe4c3jqa@4ax.com...

On Thu, 3 Oct 2024 20:20:21 -0400, "Edward Rawde"
<invalid@invalid.invalid> wrote:
>

>
"john larkin" <jl@glen--canyon.com> wrote in message news:rb5ufj1pc4uk139u9n0rljvrliqacpllq3@4ax.com...

On Thu, 03 Oct 2024 23:03:24 +0100, Cursitor Doom <cd@notformail.com>
wrote:
>

On Wed, 02 Oct 2024 19:53:49 -0700, john larkin <JL@gct.com> wrote:
>

On Wed, 2 Oct 2024 16:45:37 +0100, Clive Arthur
<clive@nowaytoday.co.uk> wrote:
>

On 30/09/2024 19:11, john larkin wrote:
<snip>

>
If they get the DC part about right, I ask them for any other
comments. All sorts of things could be mentioned.
>
With the base looking at 5K, it's unlikley to oscillate. It would be a
miracle if any kid even mentioned emitter follower oscillation. Or
noise, or tempcos, or anything else.
>

>
Along with a colleague, I interviewed someone for a repair technician's
job a few years back.  Among the questions was a simple common emitter
single transistor stage which we asked him to explain.
>
He blew us away.  He knew *far* more detail than either of us.  Turned
out he was a shit-hot analog designer looking for a less stressful job
as he wound down to retirement.  He turned out to be brilliant at his
new job, and mentored a lot of younger people.  He left when the company
was bought by a large US corporation with the concomitant mind-numbing
treacle-wading bullshit. [Me too!]

>
I see the trend, good circuit designers retiring and not being
replaced.

>
Maybe not yet, but pretty soon AI will do it better than humans.

>
Don't see how a simple quesion has enough information to generate a
complex design.

>
Take a modular approach until such time as the algos improve.
>

https://www.flux.ai/
>
Why do we have garbage like Windows and Outlook if AI is available?

>
I haven't used either for very many years. Linux is *way* better in so
many ways.

>
Outlook was always garbage. I currently set up emclient for anyone who wants an installable client which can handle many different
email addresses.
A Linux version of emclient would be nice but not likely to happen.
>
I use Windows 10 for daily work but Hyper-V has Windows xp and two debian servers.
I don't use a Linux desktop, just putty for command line and winscp for file access.
Just log in as root over SCP and use notepad++ to edit any file on the Linux system.
I also have linux boxes running proxmox.
>

>

Because garbage made money and closed source meant no-one else could laugh at the code.

>
Ha! ha! Well said, Edward; spot on!
>

If Windows is ever rewritten by AI then it's likely to be in a way which does whatever is necessary to make more money.

>
We could get AI to come up with something a little better than just
another version of Windows, I'd imagine. ;-)

>
Depends on who trains it, and what they train it to do, and what they train it to be.
>

>
Well, so long as Bill Gates doesn't train it, it'll do just fine.

>
The BASIC interpreters were fine, but they were written in assembler and refined and refined for small memory footprint.
This had the side effect of making them super efficient and about as bug free as you can get.
They were also small enough that one person could understand all or nearly all of the code.
>
Problems started when multiple people started writing DOS in c.
Programmers were now shielded from what the processor was actually doing so issues such as unchecked buffer in just about everything
arose.
I've no idea whether Gates himself ever wrote anything in c but I suspect he did not.
>
I wonder whether AI will eventually get smart enough to look at current code at assembler level and refine it in a similar way.
>

Should be a given. Compilers have been optimising higher level
languages for *years* and without the use of AI.