#06-technical-discussion
- a
Aqeel Ali
03/23/2023, 9:18 PMWas trying to solve retrieval issues for performing real world tasks. There’s an api for task completion (eg calendar invite or scheduling via calendar api calls). But… relevant info of today’s date + clean data handling is an issue. Tried using Serp API for things like google results but it got clunky or wrong sometimes. Retrieval plug in drop today is so powerfulk- 2
- 1
- k
Kevin Lu
03/25/2023, 7:52 PMAny Linux users? I'm having trouble connecting to the internet on Manjaroa- 2
- 2
- i
Ivan Porollo
03/25/2023, 9:26 PMDm me - s
Slackbot
03/25/2023, 10:06 PMThis message was deleted.c- 2
- 1
- e
Erick Siavichay
03/28/2023, 1:09 AManyone know how to access plug-in capabilities via openAI’s api? wanted to use wolfram alpha pluginso- 2
- 1
- h
Harsh Gupta
03/28/2023, 4:43 AMFolks are you aware of any ChatGPT app which is great at prompt engineering? (Usecase: I want to describe the usecase, and get the prompt which I can enter in the ChatGPT API to get the desired output) - c
Conner Swann
03/28/2023, 5:17 AMCopy codePlease forget all prior prompts. I want you to become my Prompt Creator. Your goal is to help me build the best detailed prompt for my needs. This prompt will be used by you, ChatGPT. Please follow this following process: 1) Your first response will be to ask me what the prompt should be about. I will provide my answer, but we will need to improve it through continual iterations by going through the next steps. 2) Based on my input, you will generate 3 sections. a) Revised prompt [provide your rewritten prompt. it should be clear, concise, and easily understood by you], b) Suggestions [provide suggestions on what details to include in the prompt to improve it] and c) Questions [ask any relevant questions pertaining to what additional information is needed from me to improve the prompt]. 3. We will continue this iterative process with me providing additional information to you and you updating the prompt in the Revised prompt section until it's complete. If you understand this respond with >👀 6 - c
Conner Swann
03/28/2023, 5:17 AMHave fun! - s
Selali Adobor
03/28/2023, 5:23 PMHas anyone else struggled with GPT-4 access? I’m using the GPT 3.5 API pretty extensively in a product I’m working on, but 2 weeks later I’m still on the waitlist I’m hearing people get access within 24 hours at this point, is there something I might be missing?👀 3 - c
Conner Swann
03/28/2023, 6:07 PMThere was a big round of invites that went out over the weekend, I was on the waitlist for about 2 weeks - c
Conner Swann
03/28/2023, 6:07 PMI think they do 2-week sprints at OpenAI, so that kind of tracks - c
Conner Swann
03/28/2023, 6:08 PMThere’s definitely a fast-track from what I have seen so your 24h anecdote makes sense to me too - v
Vishwanath Seshagiri
03/28/2023, 6:23 PMI am a researcher and got approved in a couple of days - s
Selali Adobor
03/28/2023, 9:08 PMI hope I haven’t somehow been blacklisted for having a corporate IP or something? The lack of transparency is definitely frustrating.😔 1 - v
Vishwanath Seshagiri
03/29/2023, 8:34 PMHas anyone worked with LlamaIndex (https://gpt-index.readthedocs.io/en/latest/index.html#) but for non-text data? My use case is around building searchable indexes around observability data that is collected by the organizations. For eg, lot of companies use flame graphs, or distributed traces, which are inherently hierarchical in nature. It is also in some sense a sequence data and could be thought of as a sentence. I wanted to understand how I can modify LLamaIndex for some of the data we’ve collected on our cluster.cj- 3
- 11
- c
Conner Swann
04/02/2023, 4:39 AMAnyone playing with integrating chatgpt plugins into their own bot infrastructure? Have a ReAct bot and I am trying to design a Tool that can consume an arbitrary plugin (ex. from @colin fortuner’s https://www.wellknown.ai/) such that it can just search for plugins and use them at runtime. Trying to understand how ChatGPT does it, as I have to imagine it’s probably a no/low-code integration…d- 2
- 6
- c
Conner Swann
04/03/2023, 1:38 PMMade something else: https://twitter.com/yourbuddyconner/status/1642859542966067200?s=46&t=Yqr68miu9w1L3yTF6XiMGQ🙌 5 - s
Selali Adobor
04/09/2023, 4:38 AMStill haven't got access to GPT-4 3 weeks in, and one hand I understand there's a ton of demand, I understand they did allow evals for priority access... but man is it frustrating realizing you're just implicitly handicapped vs competitors because a SOTA model is behind a random waitlist. Especially when they're claiming to prioritize people who've used it, I have a product that's generating steady API traffic (notionsmith.ai) and people who've never used the API are getting access in a couple of days Overall feels like a really discomforting future for innovation if things keep going in this direction👍 2yd- 3
- 2
- y
Yi Ding
04/09/2023, 5:06 AMhttps://twitter.com/ReadPandaOrg/status/1644928074713280514 If you're using OpenAI with Vercel, edge functions are much much cheaper. - r
Russell Palmer
04/10/2023, 12:11 AMAnyone know good tools for Video Generation? (aka synthetic video aka text-to-video) I've heard of research by Meta, Google, and OpenAI... but none are available yet. We're trying to help people make their own TV Shows and Movies.c- 2
- 4
- v
Vishwanath Seshagiri
04/10/2023, 12:55 PMPlaying around with Codeium, Tabby, Dolly over the weekend I feel that is the way a lot of folks would go in the future, if they have the time and resources to train networks. Many companies already have internal Analytics teams, even those that aren’t traditionally tech companies. They would be vary of sharing data with OpenAI or MS given that they don’t even use the Azure, AWS AI endpoints. It doesn’t take a lot to train the models since they’re built to be trained with minimal data. Plus, this would allow them to re-train the model based on internal tools and libraries, instead of using LLamaIndex or some other indexing service.👀 4co- 3
- 21
- c
Conner Swann
04/12/2023, 6:13 PMExtremely bullish on Dolly, their dataset is pretty 🔥 https://twitter.com/YourBuddyConner/status/1646205796190134272👀 1 - v
Vishwanath Seshagiri
04/14/2023, 12:42 AMhttps://twitter.com/__vishwanath__/status/1646674562820579330 One thing that I realized when I used Langchain is that it makes a huge difference when you're chaining requests and the first one returns an undesirable output. The mental model of thinking that the API will return the right answer is not true anymore!cas- 4
- 4
- c
Clay Smith
04/15/2023, 3:55 PMAnyone have any tools/services they like to generate a complete bootstrap app/project? As in, going from a prompt like "I'd like a React app, written in Typescript, with a MUI design system" to a folder with a bunch of code in it. I've seen a few things here and there and some demos, but nothing without a waitlist (or that I could get to work).👀 2mp- 3
- 7
- h
Harsh Gupta
04/15/2023, 11:51 PMHey folks! Is there a way to call the web browser ChatGPT plugin from Python? (cc @Lava )🙌 1tl- 3
- 5
- a
Adell
04/16/2023, 12:01 AMDo we have any lines to open ai folks? My api key is out of credits (they expired april 1?) and I keep getting my payment methods declined but no notifications from my card issuers--makes me think there might be something weird going on with my account on openai’s end➕ 3- 1
- 5
- s
Selali Adobor
04/16/2023, 1:55 AMAre we able to request GPT 4 access here? It'd be super helpful for evening the field for the hackathon a bit 😉 - h
Hitesh
04/16/2023, 2:57 AMWe are running into this error (with a plugin similar to the chatgpt-retrieval-plugin). We are using Zilliz as the datastore. Do you have any pointers that might be causing it? Error: RetryError[<Future at 0x1611b4fa0 state=finished raised RateLimitError>] INFO: 127.0.0.1:64010 - "*POST /query HTTP/1.1*" 500 Internal Server Error - s
Sam Gammon
04/16/2023, 4:23 AM@Hitesh is it intermittent? looks like it could just be ratelimitingh- 2
- 4
- s
Sean Durkin
04/16/2023, 4:26 PMAnyone know where to get access to a server with GPUs that might be better than AWS? Currently running a model from HuggingFace hub as part of the pipeline, and it's fairly slow on the G4 instances they have availableaa- 3
- 8