Casey Bralla via plug on 9 Jul 2025 04:51:39 -0700
|
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
- From: Casey Bralla via plug <plug@lists.phillylinux.org>
- To: PLUG Philadelphia Linux Users Group <PLUG@Lists.PhillyLinux.org>
- Subject: [PLUG] More AI Comments
- Date: Wed, 9 Jul 2025 07:51:32 -0400
- Dkim-signature: a=rsa-sha256; b=MrPHJZUjCZXXPXLklf/W6TXy1tiFvolK3OZ5mv9tFo3k028sJX5jfN2n5ao0f7p/AvN6kvi00YaZbYokB0zGIHnpXbLMUBF98j9EUUsN7skELZRUGOJ/NLQlg2mJcSNnihmTEY0fxI9Im39kBSpjpBylhBYB22mdouMPJ+TU8OrCVFDuNVpRpsMWw1odAqP4SvAhc2dHDQKDFjkJ4AaqARbM/K8hldm7BSb71mNUOvdRsOineULOrhyNXy7K9p8OT/+OB4LWrtxKGpNnvwD4LhFLq3K+fYvWFyvtpRlfZFshcqj6jE1uVwhvMj/gn2v1f1Wt6DdK08s5lGhis9v9Ug==; s=purelymail3; d=nerdworld.org; v=1; bh=AhoR1ZQYrp5uUIN2XE6nRaDt4pQgGy/Kbx555+uodcc=; h=Received:Date:From:Subject:To;
- Dkim-signature: a=rsa-sha256; b=BaaJb/yi/rtRTqfCy6BmpCxNBv0rgejzkSbiCjQLgUncg11oEx/S8Pnh6+CxYM55OVZsAB5PNXgzmynmlMikgotGlrnzPdVvuCS0HbsIZIN8VgoTpBm/CpGBSdQiRFpaY9z8YJxmQRhnaMammoQZVKQRcb5Wz6Ia2W4MEQW02QmopvaXbxCU0fB6ROtp3Xhr20JJY5BLwYI2+sAAtT467sfAIwBjg4Ea85C6iqReQ248ANSnk7EhZF50yP5B4gd4shvLsj/chBjuikj5S9SFQcxR2DpyVV1cf8h3fQiq3HCbAeWBbRrks4O2j5sjeSNV8vDIRbvKqyUelp4ByF9RGw==; s=purelymail3; d=purelymail.com; v=1; bh=AhoR1ZQYrp5uUIN2XE6nRaDt4pQgGy/Kbx555+uodcc=; h=Feedback-ID:Received:Date:From:Subject:To;
- Feedback-id: 44196:7652:null:purelymail
- Reply-to: Casey Bralla <MailList@NerdWorld.org>
- Sender: "plug" <plug-bounces@lists.phillylinux.org>
- User-agent: Mozilla Thunderbird
I confess to being pretty cynical about AI ("There's no 'I' in
'AI' "). To me, it was kind of like the talking dog: It wasn't
so remarkable for what it said, but that it could speak at all.
But lately I've been putting it to good use and have (somewhat)
changed my mind. I now feel it is a very useful tool. And just
like some tools, it can go horribly wrong and/or misused to bad
effect. But also like tools, it can make someone MUCH more
productive.
Lately, I've been working on a distributed temperature/humidity
sensing system for my home. It's a combination of wireless
Arduino Nanos, SHT-31 sensors, and an FTP server that collates the
data and presents it on a web page. This means I've had to use
Arduino code, python, Apache configuration, HTML, and CSS. I"m
somewhat familiar with them all, but far from proficient in any of
them.
Here's how I used various AIs:
- Write basic code snippets that I can assemble into larger
programs. This is especially helpful since I don't remember all
the syntax and structural details of these various languages.
This has tripled (or more) my code-writing speed.
- Debug my programs and tell me why various parts of them don't
work like I expected. I can often copy and paste the entire
code into the AI and it will (usually) identify my mistakes.
This is better than having someone else review it because the AI
review is more thorough, and it's less humiliating for a
computer to point out the blindingly obvious mistakes I've made.
Dave Plummer just published a good video (https://www.youtube.com/watch?v=-AJoByRGkgU&t=424s)
on which AIs are best at various tasks. Here's what I've been
using:
- ChatGPT. Good for overall coding and presenting suggestions
for improving the code
- Grok. Disappointing. Lots of output, but it never seems to
be very helpful. Often talks about what "should" work.
Probably better for current events since it seems to be tied to
X.
- Gemini. Not bad. Can present links to data sites.
- Perplexity. I don't use this much. Good for web search
- Claude. Dave loves this one the best, but I haven't used it
much yet.
I would love to hear other's opinions and conclusions about using
AI in Linux work.
Casey
___________________________________________________________________________
Philadelphia Linux Users Group -- http://www.phillylinux.org
Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce
General Discussion -- http://lists.phillylinux.org/mailman/listinfo/plug