Elevated error rate across multiple models • Hacker News Board

Re: Elevated error rate across multiple models

#81

Post by throwaw12 » Tue, Jun 23, 2026, 2:53 PM UTC

root-parent wrote: post #41

throwaw12 wrote: post #31

Imagine a future where Anthropic holds your company hostage because no one can code properly anymore by hand and demands paying 200% higher price for the usage. What can your company do?

>> What can your company do? Hire some Developers?

Developers who can code without LLMs will go extinct in couple years and there will be legends about them, you should at least have some decent open weight model as a backup

Re: Elevated error rate across multiple models

#82

Post by mguerville » Tue, Jun 23, 2026, 2:53 PM UTC

brookst wrote: post #46

spiderfarmer wrote: post #9

Earlier quoted context omitted.

If that 0.07% downtime was holding me back I wouldn't publicly admit that.

Claude is 0.89% downtime. Getting close to one nine. There aren't many tools that remain useful at that rate.

Business/Office productivity tools can be productive at that rate. Core systems like ERP or arguably CRM can't, but MS Teams is probably already that low, Figma, Canva and several others could absolutely afford to be one nine before it affects their churn materially. I suspect OpenAI and Anthropic make most of their profit on business use cases rather than dev use cases (likely higher revenue but less profit) so this may be what sets the standard of uptime.

Re: Elevated error rate across multiple models

#83

Post by kordlessagain » Tue, Jun 23, 2026, 2:54 PM UTC

yanis_t wrote: post #18

I suppose it's a good time to encourage people trying out pi[1] with any cheap model from the openrouter rankings page[1]. [1] https://pi.dev/ [2] https://openrouter.ai/rankings

I just did a build in Nemesis8 (containerized agents) and Pi appears to be working fine. Opencode is a good choice too if you're interested in checking out GLM 5.2 from z.ai.

https://github.com/deepbluedynamics/nemesis8

Re: Elevated error rate across multiple models

#84

Post by throwaw12 » Tue, Jun 23, 2026, 2:54 PM UTC

blourvim wrote: post #49

throwaw12 wrote: post #31

Imagine a future where Anthropic holds your company hostage because no one can code properly anymore by hand and demands paying 200% higher price for the usage. What can your company do?

I would guess that they would want to at the very least 10x their prices. Remember they need to make up for training, marketing, etc.. and make a big chunk of profit on top of that to justify their trillion dollar evaluation

doing 2x 3 times already gets you almost 10x increase

Re: Elevated error rate across multiple models

#85

Post by whh » Tue, Jun 23, 2026, 2:54 PM UTC

I was going to say the modern day equivalent of Github is down, but it's always down.

Re: Elevated error rate across multiple models

#86

Post by secretslol » Tue, Jun 23, 2026, 2:54 PM UTC

efficax wrote: post #74

secretslol wrote: post #33

Earlier quoted context omitted.

'Claude for Government' is the only one with 0.07% downtime, claude.ai has 0.89% downtime and claude code 0.74% - imo, that's a lot of downtime!

over a year, 0.89% is around 3 whole days of downtime

Works out at even more days when you consider working hours. And these downtime events never happen when I'm sleeping, always smack in the middle of the afternoon when I'm working.

Re: Elevated error rate across multiple models

#87

Post by dpedu » Tue, Jun 23, 2026, 2:54 PM UTC

deaton wrote: post #75

dpedu wrote: post #60

Earlier quoted context omitted.

I don't understand this comment. At worst, we're just back to the baseline - working without AI help.

With atrophy to our not-AI ability to do things

I don't buy it. Literacy rates have been increasing even after the invention of text to speech.

Re: Elevated error rate across multiple models

#88

Post by selectodude » Tue, Jun 23, 2026, 2:55 PM UTC

brookst wrote: post #46

spiderfarmer wrote: post #9

Earlier quoted context omitted.

If that 0.07% downtime was holding me back I wouldn't publicly admit that.

Claude is 0.89% downtime. Getting close to one nine. There aren't many tools that remain useful at that rate.

Heh, I’m 5x more productive 99 percent of the time. That is still a very, very useful tool.

Re: Elevated error rate across multiple models

#89

Post by bluefirebrand » Tue, Jun 23, 2026, 2:55 PM UTC

exe34 wrote: post #52

root-parent wrote: post #41

Earlier quoted context omitted.

>> What can your company do? Hire some Developers?

What exactly is a developer in a scenario where no one can code?

As long as I'm alive (and not senile) there will always be at least one developer who can code

I'm not using AI coding tools yet, and even if they force me at gunpoint to use them at work no one can force me to in my spare time

I'm not too worried about the case where no one can code anymore because that will be after I'm dead

Re: Elevated error rate across multiple models

#90

Post by DaSHacka » Tue, Jun 23, 2026, 2:56 PM UTC

dpedu wrote: post #87

deaton wrote: post #75

Earlier quoted context omitted.

With atrophy to our not-AI ability to do things

I don't buy it. Literacy rates have been increasing even after the invention of text to speech.

> Literacy rates have been increasing

uh

https://news.harvard.edu/gazette/story/2025/09/whats-driving...

https://literacybuffalo.org/2025/01/23/adult-literacy-rates-...