Free Republic
Browse · Search
General/Chat
Topics · Post Article

Skip to comments.

AI coding tools are like that helpful but untrustworthy friend, devs say
The Register ^ | Thu 12 Jun 2025 | Thomas Claburn

Posted on 06/14/2025 7:56:18 PM PDT by nickcarraway

A survey from AI biz Qodo finds robo-coding productivity gains are unevenly distributed

Exclusive Software developers largely appreciate the productivity improvements they get from AI coding tools, but they don't entirely trust their output, according to a survey conducted by AI coding biz Qodo.

As a result, some potential productivity gains get lost to manual reviews deemed necessary to check the AI's work.

Qodo offers "an agentic code quality platform for reviewing, testing, and writing code," so it has an opinion on such matters.

For its report titled "The State of AI Code Quality 2025" – provided in advance to The Register – Qodo earlier this year conducted a survey of 609 developers using unspecified AI coding tools at a variety of organizations in different industries, ranging from startups to enterprises. A whopping 82 percent of the respondents said they use the tools at least weekly, and 78 percent reported productivity gains from them.


TOPICS: Business/Economy; Computers/Internet; Education
KEYWORDS: ai; aintthatpeculiar; coding; computerscience; doge; europeanunion; meta; patch

1 posted on 06/14/2025 7:56:18 PM PDT by nickcarraway
[ Post Reply | Private Reply | View Replies]

To: nickcarraway

lol. AI can’t code.


2 posted on 06/14/2025 8:06:24 PM PDT by dadfly
[ Post Reply | Private Reply | To 1 | View Replies]

To: dadfly
"lol. AI can’t code."

I've used AI to develop several nice programs that, over the years, I just didn't have time to do. It's not so much the coding that I have trouble with AI performing, but rather the testing.

3 posted on 06/14/2025 8:57:30 PM PDT by The Duke (Not without incident.)
[ Post Reply | Private Reply | To 2 | View Replies]

To: The Duke

humm. well, i guess it might be time saving by putting out boiler plate print statements that would take me a few hours to type out by hand.

but, when i was in the computer science program at UC (ug and grad) we had two grades for programs: A for it works perfectly and basically F for anything else. admittedly that was back in the 70’s. different time, perhaps different standards. so for me, if i have to fix someone’s or something’s program for them, it’s always an F. ergo my statement, ‘AI can’t code.’


4 posted on 06/14/2025 9:29:28 PM PDT by dadfly
[ Post Reply | Private Reply | To 3 | View Replies]

To: The Duke

Yes, iterative testing with AI can be a serious headache. I had one just start removing blocks of working code when trying to work out a bug in a different part. With no explanation as to why.


5 posted on 06/14/2025 9:36:04 PM PDT by FrankRizzo890
[ Post Reply | Private Reply | To 3 | View Replies]

To: The Duke

I’ve used it to throw together a quick framework for scripting languages like Python and Powershell, but it VERY often hallucinates commandlets, parameters, and syntax that literally doesn’t exist.

If entities like Microsoft are vowing to have up to 50% of their code written by AI, they’ll need to dial in reality, because unless they’re going to have it invent a completely unique and AI-specific coding language, it can’t write stable code into existing frameworks.


6 posted on 06/15/2025 12:38:51 AM PDT by rarestia (“A nation which can prefer disgrace to danger is prepared for a master, and deserves one.” -Hamilton)
[ Post Reply | Private Reply | To 3 | View Replies]

To: dadfly
"but, when i was in the computer science program at UC (ug and grad) we had two grades for programs: A for it works perfectly and basically F for anything else. admittedly that was back in the 70’s. different time, perhaps different standards. so for me, if i have to fix someone’s or something’s program for them, it’s always an F. ergo my statement, ‘AI can’t code.’"

And, when I was writing code for the Space Shuttle Mission Simulator (SMS) and NORAD we were similarly picky. But, when storing business data that can be re-created ... not so much.

The code to control AZ/EL for SETI telescopes was also a bit critical. Folks got unreasonably upset when they had to replace gear boxes due to excessive acceleration and jerk.

Go figure.

7 posted on 06/15/2025 6:27:55 AM PDT by The Duke (Not without incident.)
[ Post Reply | Private Reply | To 4 | View Replies]

To: dadfly

I was first exposed in the early 70s like you, but still write code today.

It’s a different world now, to say the least.


8 posted on 06/15/2025 8:03:49 AM PDT by MV=PY (The Magic Question: Who's paying for it?)
[ Post Reply | Private Reply | To 4 | View Replies]

To: The Duke

Yes, code is more important if it can bend metal (developed early Z80 motion control for printing presses).


9 posted on 06/15/2025 8:07:13 AM PDT by MV=PY (The Magic Question: Who's paying for it?)
[ Post Reply | Private Reply | To 7 | View Replies]

To: nickcarraway

https://patch.com/california/across-ca/doge-gets-access-social-security-data-what-ca-residents-should-know

Doge ‘Patch’ across California:

Patch News:

Politics & Government

DOGE Gets Access To Social Security Data: What CA Residents Should Know

The plaintiffs called the high court’s order “a sad day for our democracy and a scary day for millions of people.”

Miranda Ceja’s profile picture:

Verified Patch Staff Badge
Posted Sun, Jun 15, 2025 at 3:13 am PT

https://patch.com/california/across-ca/doge-gets-access-social-security-data-what-ca-residents-should-know


10 posted on 06/15/2025 10:15:53 AM PDT by Grampa Dave (Drones have forever changed fighting wars/with Navy/Army top water or below and anything that flies)
[ Post Reply | Private Reply | To 1 | View Replies]

To: nickcarraway

AI select best answer from data stored?.


11 posted on 06/15/2025 10:35:04 AM PDT by Vaduz
[ Post Reply | Private Reply | To 1 | View Replies]

Codex / o1 pro rarely make mistakes. Codex will perform its own unit testing, and if you know what you're doing, you can suggest A/B testing, negative testing, etc.

Naysayers can bleat, but Codex-generated code is already in most everything everyone uses right this moment.

12 posted on 06/15/2025 12:08:10 PM PDT by StAnDeliver (TrumpII)
[ Post Reply | Private Reply | To 1 | View Replies]

To: The Duke

yes, i guess that would be a bit upsetting for the MEs :).

based on your previous comment, i was wondering if a test AI could be trained to write and run simple unit tests, given general set of requirements, and then learn from how and when a system/emulation was failing to write and run some integration testing. that could be useful and take some of the boredom out. it could certainly keep trying to exercise as much of the code as possible, maybe compile stats on intermittent failures, and maybe even start detecting failure patterns.


13 posted on 06/15/2025 1:06:56 PM PDT by dadfly
[ Post Reply | Private Reply | To 7 | View Replies]

To: rarestia
Hey... *I* very often hallucination commandlets! (There goes one now!)
14 posted on 06/15/2025 2:02:06 PM PDT by The Duke (Not without incident.)
[ Post Reply | Private Reply | To 6 | View Replies]

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.

Free Republic
Browse · Search
General/Chat
Topics · Post Article

FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson