Live data from Hacker News

Elevated Errors on API and ChatGPT

status.openai.com

161–170 of 182 posts

Re: Elevated Errors on API and ChatGPT

#161
post #141

They made ChatGPT with voice available to all free users. Really wonder why they went ahead with a launch today with 90% of staff ready to jump ship.

Trying to look like its business as usual and project a appearance of stability. Only it didn't work

Re: Elevated Errors on API and ChatGPT

#163

I would pay money to watch a livestream of what’s going on in the OpenAI offices

It would just be like 5 people running around with their hair on fire.

5 because thats all thats left or 5 because its their responsibility?

Re: Elevated Errors on API and ChatGPT

#165

If the last week of OpenAI's ordeal was a setup in a hollywood flick it would be at this point where there is a hard cut to the third-billed, yet-to-be-seen-onscreen actor popping out a metal backdoor of a server room while stuffing a solid state drive into an inner jacket pocket while donning a motorcycle helmet and gloves, with a music score change indicating that the entire first two-act slow burn of this corporat…

(Sorry - The drive is to hold stolen model weights. Keen audience members would have figured that out by this point in the film. Cue music.)

no, the training data is the mcguffin, not the weights.

Re: Elevated Errors on API and ChatGPT

#166
post #102
post #90

If you’re on a Mac, Ollama with the Code Llama model and a bit of prompting is pretty decent if you’re doing Python or Javascript. Runs locally on Apple Silicon. 5-10 minute setup. It’s all free. https://ollama.ai Edit: a couple of comments below are recommending Mistral 7B over Code Llama/Llama2, so give that a try with Ollama too: ollama run mistral https://ollama.ai/library/mistral

I like using the OpenChat model based on Mistral, I think it is by far the best 7B model: https://openchat.team/ (you can also use it online here)

Holy balls, this is amazing. Thank you for sharing. At that size, I could probably go one up higher and still put it on my 3090. Thank you thank you.

It actually does really well for simple queries like I'm asking about writing systemd files.

Re: Elevated Errors on API and ChatGPT

#168

Earlier quoted context omitted.

https://phind.com is up and running if you're looking for a replacement. Disclosure: I'm a co-founder.

Anecdata: phind was just as good if not better than GPT-3 (I don't pay for GPT-4, so not sure how it compares there).

i pay for GPT-4 and (for my limited uses) it is actually better, because it has the 'sources' on the right side of the screen. the font and colors are also nicer.

i have asked GPT-4 how it knows certain things and it is like pulling teeth trying to get it to admit where it got information.

Re: Elevated Errors on API and ChatGPT

#170

If the last week of OpenAI's ordeal was a setup in a hollywood flick it would be at this point where there is a hard cut to the third-billed, yet-to-be-seen-onscreen actor popping out a metal backdoor of a server room while stuffing a solid state drive into an inner jacket pocket while donning a motorcycle helmet and gloves, with a music score change indicating that the entire first two-act slow burn of this corporat…

I had GPT 4 write a screenplay with a summary of the events of the past week. I then fed that screenplay back to it and asked it to rate it in terms of realism. It got a 3. https://chat.openai.com/share/156e4495-2e57-493b-83c9-8ac30f...

This is gold. Now tell it that it's depicting real world events and ask it if it feels safe being run by such a company.
Post reply on HN