Major outages across ChatGPT and API • Hacker News Board

Re: Major outages across ChatGPT and API

#601

Post by benterix » Fri, Nov 10, 2023, 8:50 AM UTC

nvm0n2 wrote: post #552

zarzavat wrote: post #249

Earlier quoted context omitted.

OpenAI is obviously using libgen. Libgen is necessary but not sufficient for a top AI model. I believe that Google's corporate reluctance to use it is what's holding them back.

How do you know? Has someone created a set of test questions that can definitively prove it one way or another?

I don't believe it's even possible to create such a set of test questions. You would need to specifically test for differences between the versions found on Libgen and those that aren't available there. But even then, the evidence would be inconclusive.

Re: Major outages across ChatGPT and API

#602

Post by sumedh » Fri, Nov 10, 2023, 1:24 PM UTC

GaggiX wrote: post #11

Well I guess this is the best time to say that HuggingFace hosts many open source chat models, one of my favorites is a finetuned version of Mistral 7B: https://huggingface.co/spaces/HuggingFaceH4/zephyr-chat

How do they make money if its free?

Re: Major outages across ChatGPT and API

#603

Post by Kye » Fri, Nov 10, 2023, 1:26 PM UTC

sumedh wrote: post #602

GaggiX wrote: post #11

Well I guess this is the best time to say that HuggingFace hosts many open source chat models, one of my favorites is a finetuned version of Mistral 7B: https://huggingface.co/spaces/HuggingFaceH4/zephyr-chat

How do they make money if its free?

They don't. They're on several rounds of VC.

Re: Major outages across ChatGPT and API

#604

Post by swingingFlyFish » Fri, Nov 10, 2023, 1:33 PM UTC

It's crazy to me how people have hopped on using AI in production and it's proven time and time again it's just not ready. This outage is the least of my concerns about it. It's just too immature.

Re: Major outages across ChatGPT and API

#605

Post by ho_schi » Fri, Nov 10, 2023, 1:59 PM UTC

seattle_spring wrote: post #457

ho_schi wrote: post #105

Never used ChatGPT. I couldn’t care less.

This seems like one of those "oh I don't own a TV" moments.

Is that bad?

My impression is that people without a TV or relying on a feature phone seem rather happy. Does something provide a necessary purpose? Is there an alternative? Is it working autonomously? Bidirectional?

I also assumed you need this in that in the past. Things anre obligatory and so on. And I’ve changed my opinion.

Re: Major outages across ChatGPT and API

#606

Post by zamadatix » Fri, Nov 10, 2023, 2:35 PM UTC

dragonwriter wrote: post #493

zamadatix wrote: post #479

Earlier quoted context omitted.

DALL-E 3 isn't available via API or labs. If you don't use ChatGPT you get the (significantly lower quality) DALL-E 2 at the moment. That's supposed to change by the end of fall. One of the most frustrating things with "Open"AI is you can't just use what they announce as available, you have to wait for an A/B rollout (as a paying customer!) or for it to be accessible in direct way instead of going through multiple mo…

> DALL-E 3 isn't available via API or labs. Yes, it is available in the API (just tested to make sure the docs weren't misleading.)

Ah, looks like it released a couple of days prior! https://techcrunch.com/2023/11/06/openai-launches-dall-e-3-a...

Re: Major outages across ChatGPT and API

#607

Post by seattle_spring » Fri, Nov 10, 2023, 4:16 PM UTC

ho_schi wrote: post #605

seattle_spring wrote: post #457

Earlier quoted context omitted.

This seems like one of those "oh I don't own a TV" moments.

Is that bad? My impression is that people without a TV or relying on a feature phone seem rather happy. Does something provide a necessary purpose? Is there an alternative? Is it working autonomously? Bidirectional? I also assumed you need this in that in the past. Things anre obligatory and so on. And I’ve changed my opinion.

It's not bad, but many people wedge that fact into unrelated conversations simply to assert some level of smug superiority.

Re: Major outages across ChatGPT and API

#608

Post by seanhunter » Sat, Nov 11, 2023, 5:21 AM UTC

mst wrote: post #583

seanhunter wrote: post #582

Earlier quoted context omitted.

I have a slightly different take. Not all use cases are narrow use cases. OpenAI crushes the broad and/or poorly defined use cases. On those if you tried to train your own inhouse model it would be very expensive and you would produce a significantly inferior model.

I'm not sure how my "quite a lot of use cases" and your "not all use cases are narrow use cases" are meaningfully different (slightly) to you. This isn't a snipe, mind, it's me being unsure if we even disagree, especially given the latter part of your comment seems entirely correct (so far as my limited understanding goes ;).

That's not the part that's different. The part where I feel we perhaps differ is rather than being "premium mediocre" I think that openAI is really excellent where the problem space is very broad or is poorly specified. Then we both agree there are better choices where it is narrow and well specified.

Re: Major outages across ChatGPT and API

#609

Post by satvikpendem » Sat, Nov 11, 2023, 7:26 AM UTC

still_grokking wrote: post #554

satvikpendem wrote: post #491

Earlier quoted context omitted.

> it's that the dataset really is the 'secret sauce' alwayshasbeen.jpg There have been articles about how "data is the new oil" for a couple of decades now, with the first reference I could find being from British mathematician Clive Humby in 2006 [0]. The fact that it rings even more true in the age of LLMs is simply just another transformation of the fundamental data underneath. [0] https://en.wikipedia.org/wiki/Cl…

2006? https://en.wikipedia.org/wiki/Scientia_potentia_est

> There have been articles about how "data is the new oil" for a couple of decades now, with the first reference I could find being from British mathematician Clive Humby in 2006

I am specifically referring to the phrase I quoted, not some more abstract sentiment.

Re: Major outages across ChatGPT and API

#610

Post by piersolenski » Sat, Nov 11, 2023, 12:09 PM UTC

sibit wrote: post #458

piersolenski wrote: post #446

Earlier quoted context omitted.

I made a NeoVim plugin that helps debug diagnostic messages, providing explanations and solutions for how to fix them, custom tailored to your code. It's a bit different from other plugins which only act on the text in the buffer in that it also sends the diagnostics from the LSP to ChatGPT too. https://github.com/piersolenski/wtf.nvim

That's a really neat plugin! Couple of questions for you: 1. How many times a [day/month] do you use it? 2. In your experience how often does GPT 'hallucinate' an explanation?

Thanks! I use it every day, especially with Typescript where the diagnostic errors and notoriously hard to read. The instructions it sends off the GPT are quite strict and fine tuned at this point, so the responses tend to be good. The solutions are either bang on the money or at least the explanations point me in the right direction. You can pass additional instructions and configure the model as well to further fine tune it.