Talk about an extra slap in the fuckin face… getting blamed for something your replacement did. Cool.
Amazon blames human employees for an AI coding agent’s mistake | Two minor AWS outages have reportedly occurred as a result of actions by Amazon’s AI tools
Submitted 3 weeks ago by remington@beehaw.org to technology@beehaw.org
Comments
Soulphite@reddthat.com 3 weeks ago
Tharkys@lemmy.wtf 3 weeks ago
That’s in the SOP for management.
Soulphite@reddthat.com 3 weeks ago
True. In this case, these poor saps being tricked into “training” these AI to eventually render their jobs obsolete.
Megaman_EXE@beehaw.org 2 weeks ago
It blows my mind that people are going forward with this AI nonsense and that it has infected key infrastructure. I feel like im taking crazy pills here. I could kind of understand if it actually worked. Like if it genuinely worked as well as they said? I could totally understand it. I would still question it, but it would make more sense.
Penguincoder@beehaw.org 2 weeks ago
A computer (AI) can never be held accountable. Therefore, a computer (AI) must never make any decision.
kibiz0r@midwest.social 2 weeks ago
This is a terrible idea for Amazon, the cloud services company.
But for Amazon, the AI company? This is them illustrating the new grift that almost any company can do: use AI to keep a plausible mirage of your company going while reducing opex, and sacrifice humans when necessary to dodge accountability.
But his job wasn’t even to supervise the chatbot adequately (single-handedly fact-checking 10 lists of 15 items is a long, labor-intensive process). Rather, it was to take the blame for the factual inaccuracies in those lists. He was, in the phrasing of Dan Davies, “an accountability sink” (or as Madeleine Clare Elish puts it, a “moral crumple zone”).
pinball_wizard@lemmy.zip 3 weeks ago
described the outages as “small but entirely foreseeable.”
LMAO
MNByChoice@midwest.social 3 weeks ago
Yay! Extra mental load of having to ask the AI “correctly” and then keep up one’s skills to be able to review the AI’s work! Extra bonus for being blamed for letting anything slip past.
At least the junior that fucked up will learn something from the experience and can buy a round of beers (if the junior is paid well enough, otherwise the seniors have to buy the junior a beer while talking it out).
Powderhorn@beehaw.org 3 weeks ago
I’m reminded of a time I was in a bar in Georgia at a conference. It was in the hotel, and a high-ranking editor for the then-reputable Washington Post bought me a beer. He let me take a sip before launching into how much “immature shit [I] need to get out of [my] system” before being ready to be “Post material.”
Where is any industry going to be in a decade, when no one’s been mentored?
Petter1@discuss.tchncs.de 3 weeks ago
Well, AI code should be reviewed prior merge into master, same as any code merged into master.
We have git for a reason.
So I would definitely say this was a human fault, either reviewer’s or the human’s who decided that no (or AI driven) review process is needed.
If I would manage devOps, I would demand that AI code has to be signed off by a human on commit taking responsibility with the intention that they review changes made by AI prior pushing
pinball_wizard@lemmy.zip 3 weeks ago
If I would manage devOps, I would demand that AI code has to be signed off by a human on commit taking responsibility with the intention that they review changes made by AI prior pushing
And you would get burned. Today’s AI does one thing really really well - create output that looks correct to humans.
Limerance@piefed.social 3 weeks ago
Realistically what happens is the code review is done under time pressure and not very thoroughly.
TehPers@beehaw.org 2 weeks ago
This is what happens to us. People put out a high volume of AI-generated PRs, nobody has time to review them, and the code becomes an amalgamation of mixed paradigms, dependency spaghetti, and partially tested (and horribly tested) code.
Also, the people putting out the AI-generated PRs are the same people rubber stamping the other PRs, which means PRs merge quickly, but nobody actually does a review.
The code is a mess.
heluecht@pirati.ca 2 weeks ago
@Petter1 @remington at our company every PR needs to be reviewed by at least one lead developer. And the PRs of the lead developers have to be reviewed by architects. And we encourage the other developers to perform reviews as well. Our company encourages the usage of Copilot. But none of our reviewers would pass code that they don't understand.
JakenVeina@midwest.social 2 weeks ago
That’s actually the correct approach, they just got the wrong humans. It’s the executives that are FORCING developers to use AI as a metric that are responsible here, not the developers themselves.
AllNewTypeFace@leminal.space 3 weeks ago
AI can never fail, it can only be failed
XLE@piefed.social 3 weeks ago
If a person is going to be blamed, it should be the one that mandated use of the AI systems…. Because that’s exactly what Amazon was doing.