[HN Gopher] Ask HN: Is ChatGPT down?
       ___________________________________________________________________
        
       Ask HN: Is ChatGPT down?
        
       Author : franze
       Score  : 76 points
       Date   : 2024-12-26 18:58 UTC (4 hours ago)
        
       | alex_young wrote:
       | Seems that way. Status page says we're wrong though.
       | 
       | https://status.openai.com/
        
         | transcriptase wrote:
         | Wish I could find where someone at AWS described in great
         | detail why their status page was so useless for so long. It
         | basically needed so many approvals to change it (which was
         | disincentivized by SLAs) that issues were usually resolved
         | first.
         | 
         | That's setting aside when they hosted the status
         | red/yellow/green indicator images on s3, so the surest sign
         | that s3 was having issues was that the status indicators didn't
         | load at all.
        
           | jiggawatts wrote:
           | You answered your own question!
           | 
           | "Boss, can I change this indicator from green to red?"
           | 
           | "How much will that cost us?"
           | 
           | "About a million dollars an hour."
           | 
           | "No."
        
             | jsiepkes wrote:
             | Besides, we are not fully down. Some services are just
             | degraded. Only 50% of the requests take 40 seconds to
             | complete. Never mind the fact you'll never be able to load
             | a complete page due to all the timeouts.
        
           | LeoPanthera wrote:
           | "Step up to red alert."
           | 
           | "Sir, are you absolutely sure, it does mean changing the
           | bulb."
           | 
           | https://youtu.be/jZn1fhMdrbQ
        
         | fcmgr wrote:
         | It got updated, they have acknowledged there is a major outage.
        
       | albeebe1 wrote:
       | It is for me
        
       | slater wrote:
       | Yup, same here. Loads parts of the UI, but no history or answers
       | from server
       | 
       | edit: downdetector confirms:
       | 
       | https://downdetector.com/status/openai/
        
       | FGTN5cPn8fVHKAV wrote:
       | Yes it is. Status page is lying to us all.
        
       | FGTN5cPn8fVHKAV wrote:
       | New intern did something wrong again I guess.
        
         | alwaysrunning wrote:
         | DNS change...
        
       | nozzlegear wrote:
       | I think Azure itself is having issues right now - that's the
       | story all of my services deployed there are telling at least.
        
       | slater wrote:
       | downdetector comments with the jokes: "This signals the workday
       | is over"
        
         | jacobmarble wrote:
         | But there are dozens of us working today! Dozens!
        
       | delduca wrote:
       | Yes, it is down for me (desktop & mobile apps)
        
       | prathamkumar11 wrote:
       | ChatGpt down since an hour!
        
       | m4rc3lv wrote:
       | Yes, it is down
        
       | rvz wrote:
       | So many major outages and little gets fixed. Almost as if nothing
       | was learned since the previous postmortem from the last outage.
       | 
       | Can't wait for this postmortem to be released.
        
       | 0xFEE1DEAD wrote:
       | Looks like Azure is experiencing a major outage, but I cannot
       | find anything about it. If you look at downdetector.com you'll
       | notice reported outages from OpenAI, Microsoft 365, XBox Live,
       | Walmart, the list goes on.
        
         | colinbartlett wrote:
         | If nobody minds a plug: My own product, StatusGator, which was
         | launched here on HackerNews 10 year ago, notifies IT teams
         | about outages before they are acknowledged by official status
         | pages.
         | 
         | - This OpenAI outage[1] we notified 4 minutes before they
         | acknowledged.
         | 
         | - The last AWS outage[2], we notified 28 minutes before they
         | acknowledged
         | 
         | - There is def an Azure outage[3] now yet they have still not
         | updated their status page. We notified 35 minutes ago.
         | 
         | 1. https://statusgator.com/services/openai
         | 
         | 2. https://statusgator.com/blog/amazon-cognito-outage-
         | december-...
         | 
         | 3. https://statusgator.com/services/azure
        
         | belter wrote:
         | > Looks like Azure is experiencing a major outage...
         | 
         | Does Azure have any other state? :-)
        
         | sunaookami wrote:
         | Status page is updated now: https://azure.status.microsoft/
         | 
         | >Impact Statement: Starting at 18:44 UTC on 26 Dec 2024, you
         | have been identified as a customer who was impacted by a power
         | incident in South Central US and may experience a degraded
         | experience.
         | 
         | >Current Status: There was a power incident in the South
         | Central US AZ03 which affected multiple services. We have
         | applied mitigation and are actively validating recovery to the
         | impacted services. Further updates will be provided in 60
         | minutes, or sooner as events warrant.
         | 
         | The times are the same for OpenAI - first notice from 11:00 PST
         | (19:00 UTC)
        
       | readyplayernull wrote:
       | Didn't notice, Claude is my goto bot for programming.
        
       | TZubiri wrote:
       | yes, meanwhile I'm trying out Google Gemini
        
       | fcmgr wrote:
       | Yes, status page got updated: https://status.openai.com/. It says
       | "This issue is caused by an upstream provider and we are
       | currently monitoring."
        
       | novaRom wrote:
       | Don't put all your eggs into one basket - what are some good free
       | alternatives? I mean comparable "usefulness" including image
       | understanding, voice interface, and general knowledge?
        
         | porphyra wrote:
         | Grok 2 is available for free for all X users [1]
         | 
         | [1] https://x.ai/blog/grok-1212
        
           | manmal wrote:
           | Is it comparable with OpenAI's and Anthropic's models? I have
           | a strong resistance towards using Musk's products nowadays,
           | but maybe I should take a look over the fence.
        
             | porphyra wrote:
             | In terms of general purpose usage and image understanding,
             | I think Grok 2 is pretty good and roughly on par with
             | ChatGPT 4o. Grok 2 will also look for online sources
             | similar to ChatGPT search which is nice. I've occasionally
             | had it report irrelevant results when doing so though ---
             | for example when I was looking up the price of a car it
             | once found some random webpage with the price in rupees
             | even though I'm in the US.
             | 
             | For logical reasoning of course chain-of-thought models
             | like the O1 family are better.
        
         | KennyBlanken wrote:
         | General knowledge?
         | 
         | Five times in two weeks I've asked OpenAI some basic factual
         | information and it didn't get even close on any of them.
        
           | franze wrote:
           | Dont ask GPTs for facts, thats a knowledge problem, they dont
           | habe any.
           | 
           | Ask it for reasoning, u habe to bring the facts.
        
         | Mond_ wrote:
         | If you're using ChatGPT actively, then surely you have heard of
         | Gemini and Grok (no clue of how far Grok gets you nowadays, but
         | Gemini should. Not sure how good the voice interface is).
        
           | Yawrehto wrote:
           | I wouldn't use Grok. Google is a big company that has to be
           | reasonably objective, but Elon seems to be the sort of guy
           | who would pettily include jabs at people he doesn't like into
           | the data.
           | 
           | Also he's a horrible human for many reasons and I'd prefer
           | not to support him if I can avoid it. (You know it's bad when
           | _Google_ is ethical in comparison.)
        
             | llm_trw wrote:
             | Ah yes, google a company known for not injecting an
             | ideology into their AI:
             | https://www.telegraph.co.uk/news/2024/02/23/google-gemini-
             | ai...
             | 
             | Nor building AI tools for the pentagon to bomb Yemeni
             | weddings more efficiently:
             | https://en.wikipedia.org/wiki/Project_Maven
        
             | sandspar wrote:
             | Do you sincerely believe that Gemini is not very biased?
        
         | manmal wrote:
         | Not free, but something I'm doing right now: Using Cursor's
         | chat mode with my prepaid Anthropic key. Works pretty well as a
         | stop gap.
        
           | jascination wrote:
           | What's the difference when using your own anthropic key vs
           | vanilla? Are you hitting limits with Claude in cursor that
           | they key unlocks?
           | 
           | I ask cos I use it very liberally and haven't had any issues
           | that have made me consider adding a key, except when I made
           | it read my whole codebase on every request
        
         | s1gsegv wrote:
         | I've been loving having mistral-nemo on my laptop. Not
         | comparable for images, voice, etc, however it is very nice when
         | you're away from the internet. For the cost of a couple
         | gigabytes you get to keep a good amount of info at the ready.
         | Very easy to run models these days, install Ollama and then do
         | `ollama run mistral-nemo`.
         | 
         | Plus mistral-nemo in particular has a large context window, so
         | you can cook up some shell scripts to throw a bunch of context
         | into the buffer before your question. One I use a lot takes the
         | name of a manpage and a question about it, then the LLM has the
         | whole manpage to reference.
        
       | bobjordan wrote:
       | Wow, I subscribed to pro-mode and have been using it like crazy
       | since release to assist my firmware development. Extremely
       | helpful. Now it's changed my workflow so much, I don't even want
       | to work when it's down.
        
         | projectileboy wrote:
         | Would you be willing to elaborate on the ways in which the
         | $200/month subscription is better than the $20/month
         | subscription? I'm genuinely curious - not averse to paying
         | more, depending on the value.
        
       | hbamoria wrote:
       | Working using API, though
        
       | deadbabe wrote:
       | Don't understand these people who can't work without ChatGPT.
       | Just look stuff up on stack overflow.
        
         | kshacker wrote:
         | ChatGPT does much better when it gets it right.
         | 
         | Stackoverflow will have duplicates, approximates and what not
         | and sometimes that works. But at other times, you hunt for a
         | half hour before you figure it out.
         | 
         | You can throw the problem at ChatGPT, it may go wrong but your
         | course correct it with simple instructions and slowly but
         | steadily you move towards your goal with minimal noise of the
         | irrelevant discussions.
         | 
         | What stands between the solution and you then is your ability
         | to figure out when it is hallucinating and guide it to the
         | right direction. But as a solution developer you should have
         | that insight anyways
        
           | disqard wrote:
           | I'm with you (I use Claude Sonnet, but same difference...).
           | 
           | I do wonder if we're the last generation that will be able to
           | effectively do such "course correct" operations -- feels like
           | a good chunk of the next generation of programmers will be
           | bootstrapped using such LLMs, so their ability to "have that
           | insight" will be lacking, or be very challenging to
           | bootstrap. As analogy, do you find yourself having to "course
           | correct" the compiler very often?
        
             | kshacker wrote:
             | Many a times.
             | 
             | I asked it a simple non programming question. My last
             | paycheck was December 20, 2024. I get paid biweekly. In
             | which year will I get paid 27 times. It got it wrong ...
             | very articulately.
             | 
             | I run into this every single day.
        
               | CharlesW wrote:
               | It helps to know how LLMs work. They're not "good" at
               | math ("good" being anthropomorphization) because they
               | just predict text patterns based on training data rather
               | than perform calculations based on logic and mathematical
               | rules.
               | 
               | To do this reliably, prepend your request to invoke a
               | tool like OpenAI's Code Interpreter (e.g. "Code the
               | answer to this: My last paycheck was December 20, 2024. I
               | get paid biweekly. In which year will I get paid 27
               | times.") to get the correct response of 2027.
        
         | mrweasel wrote:
         | You joke, but development is getting weird and perhaps to
         | dependent on online services. People can't work when, ChatGPT,
         | Github, AWS, NPM (add any other package manager you'd like) or
         | Stack Overflow (this seems less important these days) is down.
         | 
         | These services are accumulative and the chance that any one of
         | of those services being down at any given time is increasing
         | every time we add a new online dependency.
        
         | Sohcahtoa82 wrote:
         | Closed as duplicate. Here's a link to a question that is only
         | marginally related to what you were actually asking and has an
         | answer that is horribly out of date and doesn't even work
         | anymore.
        
           | ok_dad wrote:
           | Guess what ChatGPT was trained on!
        
       ___________________________________________________________________
       (page generated 2024-12-26 23:00 UTC)