Developers Joke About 'Coding Like Cavemen' As AI Service Suffers Major Outage (arstechnica.com) 28
An anonymous reader quotes a report from Ars Technica: On Wednesday afternoon, Anthropic experienced a brief but complete service outage that took down its AI infrastructure, leaving developers unable to access Claude.ai, the API, Claude Code, or the management console for around half an hour. The outage affected all three of Anthropic's main services simultaneously, with the company posting at 12:28 pm Eastern that "APIs, Console, and Claude.ai are down. Services will be restored as soon as possible." As of press time, the services appear to be restored. The disruption, though lasting only about 30 minutes, quickly took the top spot on tech link-sharing site Hacker News for a short time and inspired immediate reactions from developers who have become increasingly reliant on AI coding tools for their daily work. "Everyone will just have to learn how to do it like we did in the old days, and blindly copy and paste from Stack Overflow," joked one Hacker News commenter. Another user recalled a joke from a previous AI outage: "Nooooo I'm going to have to use my brain again and write 100% of my code like a caveman from December 2024."
The most recent outage came at an inopportune time, affecting developers across the US who have integrated Claude into their workflows. One Hacker News user observed: "It's like every other day, the moment US working hours start, AI (in my case I mostly use Anthropic, others may be better) starts dying or at least getting intermittent errors. In EU working hours there's rarely any outages." Another user also noted this pattern, saying that "early morning here in the UK everything is fine, as soon as most of the US is up and at it, then it slowly turns to treacle." While some users criticized Anthropic for reliability issues in recent months, the company's status page acknowledged the issue within 39 minutes of the initial reports, and by 12:55 pm Eastern announced that a fix had been implemented and that the company's teams were monitoring the results.
The most recent outage came at an inopportune time, affecting developers across the US who have integrated Claude into their workflows. One Hacker News user observed: "It's like every other day, the moment US working hours start, AI (in my case I mostly use Anthropic, others may be better) starts dying or at least getting intermittent errors. In EU working hours there's rarely any outages." Another user also noted this pattern, saying that "early morning here in the UK everything is fine, as soon as most of the US is up and at it, then it slowly turns to treacle." While some users criticized Anthropic for reliability issues in recent months, the company's status page acknowledged the issue within 39 minutes of the initial reports, and by 12:55 pm Eastern announced that a fix had been implemented and that the company's teams were monitoring the results.
Oddly enough.. (Score:5, Insightful)
productivity actually went up.
Re: (Score:1)
Lines of code went down
Number of bugs went down
Understanding of the code base went way up
So... (Score:5, Funny)
> Developers Joke About 'Coding Like Cavemen' As AI Service Suffers Major Outage
So "coding" then?
So am I a cave man ? (Score:5, Insightful)
I have never used an AI to write code. The code that I write I understand and am able to document.
Re:So am I a cave man ? (Score:5, Insightful)
Slashdot is owned by people who boost the most ridiculous shit. If it isn't AI, it's cryptocurrencies. And they persist in doing so to the actual people who know it's a con. It's so pathetic.
Hey, Slashdot, ACTUAL DEVELOPERS were not making jokes about coding like cavemen. We were doing our work and didn't notice any outages. "Vibe coders", maybe, but vibe coders aren't actual developers, they're con-artists.
Re: (Score:2)
I almost never bother reading the articles, most of which are paywalled, anyway. I'm just here for the comments.
Speaking of bad jokes and ridiculous shit, my mod points are suspiciously absent again. I'll never understand the logic of randomly taking them away to "encourage" me to use them.
Re: (Score:3)
The code that I write I understand and am able to document.
We recently had a new round of replacement hiring; and among the candidates, we had two people who used AI to implement the small required proof-of-knowledge project. They both used ChatGPT to help with the project. One of them could explain what the code was doing, and indicate what was needed to modify the code to do something else, and one of them has no idea. We immediately shit-canned the latter, and the former moved forward. Personally, I think they should have both been shit-canned immediately; but I
Re:So am I a cave man ? (Score:5, Interesting)
Anyone who can't write their own code without ChatGPT or its ilk needs to be either forced to take programming classes or to find another job.
I don't know. I'm torn over this. This could well be the future - not fully coding by AI, but commonly using AI to assist.
I've built an arithmetic logic unit using NAND gates. I've programmed in assembler. I've programmed in-line assembler in C programs. Sure, I have a good appreciation on what happens at a low level, but is that really relevant today? What benefit does someone starting today have if they know this sort of stuff?
Everyone today uses frameworks, huge libraries of pre-built functions, IDE's that refactor and lots of other magic that back in the day, editing C code in vi, I couldn't even dream of. I remember using Borland C for the first time when it was new, and being amazed at what the IDE could do, and how it improved my productivity.
Progress happens, we have to move with the times. If someone can complete the task, and uses some wiz-bang IDE with AI integration - then as long as what they create is fit for purpose, does it really matter that they have never soldered a circuit board?
Re: (Score:2)
Anyone who can't write their own code without ChatGPT or its ilk needs to be either forced to take programming classes or to find another job.
I don't know. I'm torn over this. This could well be the future - not fully coding by AI, but commonly using AI to assist.
I've built an arithmetic logic unit using NAND gates. I've programmed in assembler. I've programmed in-line assembler in C programs. Sure, I have a good appreciation on what happens at a low level, but is that really relevant today? What benefit does someone starting today have if they know this sort of stuff?
Everyone today uses frameworks, huge libraries of pre-built functions, IDE's that refactor and lots of other magic that back in the day, editing C code in vi, I couldn't even dream of. I remember using Borland C for the first time when it was new, and being amazed at what the IDE could do, and how it improved my productivity.
Progress happens, we have to move with the times. If someone can complete the task, and uses some wiz-bang IDE with AI integration - then as long as what they create is fit for purpose, does it really matter that they have never soldered a circuit board?
AI like Anthropic is trained on open-source code at Github, GitLab, etc. If you're an open-source developer using the same framework the AI was trained on vociferously, AI can write good code. Trust me. True, one needs to be discerning as a professional what to ultimately accept, but it can have a programmer's multiplier effect, and the code can still be approved by the team before final repo commitment.
Using GIT and branches, one can be bold! Using AI one can also afford the time to experiment and then ul
Re: (Score:2)
In short, yes.
And it's going to be an entire age before people realize that was actually a better way to be in many regards, and learn something from it.
There are undeniable and dramatic benefits to technological improvement, but trying to throw every technological advance at every problem is the kind of stupid that can only be fostered by capitalism, where anything that doesn't make money is a waste of time that leaves you falling behind in income, and then you get walked on by someone with more capital. H
The sad truth behind the jokes... (Score:5, Insightful)
Re: (Score:3)
Re: (Score:2)
Indeed.
In the 1980s I had to write some code in assembly language. I've mostly forgotten how to do that, but I have no desire to go back to those days.
Re:The sad truth behind the jokes... (Score:4, Interesting)
Probably because the CPU was shit? Intel?
By contrast, I'd love to have the excuse to immerge myself back into 68k assembly. Attached to an ATARI ST or AMIGA would be cherry on top. Or learn ATARI Transputer. Or learn original ARM Archimedes.
Re: (Score:2)
By contrast, I'd love to have the excuse to immerge myself back into 68k assembly.
I never got the opportunity to learn 68k assembly, but my first assembly was on the 6809. Motorola assembly language was a joy. When I had to move to Intel assembly language on the 8088, I immediately discovered how badly it was designed and how absolutely painful it was to use.
Like you, I would welcome the opportunity to re-immerse myself in Motorola assembly language, while I would dread the necessity, should it ever arise, to have to re-immerse myself in Intel assembly language.
Re: (Score:3)
Most of my assembly experience was on a Commodore 64 on their MOS 6510 chip. Very primitive, but for the time, very advanced. I'm glad I learned out to do it, but I don't like making my own nails from scratch, in order to build a building. These days, I love C# / .NET because it has utility classes for literally everything. I can spend my time focusing on what I want the software to do, not on making components to build the software.
Re: (Score:2)
No, my assembly was mostly on Commodore 64. Its 6510 processor was made by MOS Technology, a Commodore subsidiary.
I also did some assembly on a Wang minicomputer in the 1980s, and later, a bit of x86 assembly in the early 90's.
I don't care to go back to any of them.
Re: (Score:2)
True but not sad. Forget AI, if you switch from coding in C++ to Python, within months you will start to lose fluency with C++. Whatever you do stops you from doing almost infinitely many other things.
It's more than that. I myself had multiple prolonged phases when I was doing almost exclusively C++ or Python programming, and sure one forgets some details of the syntax, which is relatively easy to remind oneself of again when switching back.
But these LLM addicted coders are different, it is really like they loose their entire programming abilities, including the ability to think of suitable algorithms or data structures for a given requirement.
Switching programming languages may be like switching from
Re: (Score:3)
Apocolypse Meh? (Score:3)
Too much cloud dependence (Score:2)
Having an up-to-date local machine running a code LLM should be the goal for any org that's that dependent on AI to get work done.
Its so damn useful (Score:3)
Anthropic's Claude Sonnet 4 is the premium coding engine in my experience. They charge extra for it and for good reason. I use other cheaper models if I am running low on credits but Claude 4 definitely rules if you value your time.
Jeah. (Score:2)
You mean coding like people who actually need to know what they're doing - as in be actual professionals?
Oh well, I guess 99% of Indians took the day off.
Cry me a river (Score:2)
Oh, the horror. People actually had to do their job instead of copying and pasting mashed up versions of other people's code using some third party service.
Problem: relying on one resource (Score:2)
I program IBM RPG professionally, and while I do use several AI tools for reference, relying on one is simply absurd. I always keep multiple tools available.
As to using AI to write code, for my use case, AI is a great tool to explain code and to suggest rewrites to code, but for actually writing new code, it's not reliable. It's great for pointing me in the right direction, but in the end, my programming experience is what helps the most.
"blindly copy and paste from Stack Overflow" (Score:2)
Where do these idiots think the "AI" gets its code from? From fucking Stack Overflow and every other site on the internet. You're just copying and pasting something a tin can cobbled together from random crap it found during its "training". This is not an improvement. These "jokes" prove their brains have already atrophied past the point of being worth hiring.