Honestly if this was possible there are more egregious issues on their part than using AI.
If your backups are stored alongside your production data THEY ARE NOT BACKUPS
Whatever happened to tapes offsite?
The truth is many firms out there don’t have the slightest notion of how to do software engineering properly.
It’s years of wanting IT on a shoestring budget and a “just get it done” dictat.
Not necessarily. I had a student intern at a shop where everybody just directly edited prod and there was no version control system.
At my first job, the software was configured by directly manipulating the SQL database, using UPDATE statements that were created by Excel macros.
The Testing database doubled as the only backup.
They didn’t have Remote Desktop licenses for the server, so only 2 people could work on it simultaneously using admin accounts.
Everyone down to first level support and the secretary had domain admin rights.Oh my god, that’s glorious. I have some pretty sick stories from what my students have seen, but yours is going to be awfully hard to beat.
Crane decided to ask his AI agent why it went through with its dastardly database deletion deed. […] So, the agent ‘knew’ it was in the wrong.
No, you asked the confabulation machine to confabulate a reason/excuse after the fact, and it confabulated something that looks like a reason/excuse. At no point was there knowledge or introspection.
LLMs can’t ’go rogue’, as that would require innate coherence and intent.
They’re explosively imprecise, statistically luke-warm grey goo extrusion sphincters of historical sewage.
Anyone who deploys one without supervision deserves everything it excretes, and anyone impressed by it enough that it resembles intelligence to them is betraying their limited natural capacity.
I don’t know if you are correct or not… But you said it well.
I like how we are posting real news in programmer humor
You have to admit this is pretty funny
yep 100% funny, clown world we are living in, real news could pass as a joke really
It is kind of funny.
It’s extremely funny.
This is just a classic case of bad use of the tools provided. Agents are notorious for making shit up Or getting something that’s just like super close, but not quite accurate.
I bet this dude also probably just uses the same session over and over and over and over again, which clogs up his context window and makes the model less accurate the longer it goes on to.
This probably could have been prevented if it had been forced to show a plan before it tried to do anything. It’s hard to know because the article is so light on details. You also shouldn’t brazenly trust the thing so much. You should run a command and walk away. You should keep an eye on what it is doing.
It’s a bit like giving a junior developer a production key and being like “don’t delete production!” and then walking away.
The way the guy was prompting this agent also leaves a lot to be desired. It’s trained to work on emulating human thoughts, speech patterns. Turns out When giving instructions, it’s really difficult to figure out what to do from a list of things to not do. If the dude just instead told the agent what to do and how he wanted it to work and when it needed to bring things to his attention, instead of telling it to not guess, instead explaining that it needed to use whatever tools to go look up a documentation to understand the context and scope of the project it’s working on It does a better job.
Giving a model the right context to do something is the difference between a model doing something like deleting your production database or your model acting like a magical machine that can get anything done.
Can I say LOL? LMAO, even.
This could have been done by any engineer. You need systems in place that make these things impossible. No easy access to prod environment. Proper backups. Clear APIs.
Generally, companies that have AI integrated to this extent have no engineers remaining who could have made such things impossible.
It starts with automating backups that nobody verifies for years, then continues to off-shoring all development to the cheapest contractors that nobody actively manages, handing over all “keys to the kingdom” to cloud providers, culminating with elimination of 80% of infrastructure and engineering staff in a mad dash to cut costs at any cost. At that point giving AI agents full access is just icing on the cake.
Did they pay Claude a living wage?
Do you treat all your A.I. like that?
Only a living wage can prevent warehouse fires…or data dumps too.
Only a living wage can prevent warehouse fires
Hey, that’s the interns job!
I don’t know much about railway, but it sounds like they had the backup and the database on the same volume. I’m an idiot, but even I don’t do that
Giving the hallucinating lying machine write access seems like a bad idea but what do I know
Honestly I’m as smooth brained as any other vibe coder but even I know not to give it access to my production infrastructure.
where is the humor
It’s standing over here, pointing and laughing at somebody stupid enough to trust Claude.
ok. i guess i’ve seen something like this so many times my only reaction is disappointment
I hope to never lose the simple joy of laughing at others who are suffering the consequences of their stupid, stupid decisions.
it was removed along with the database
