26/03/2023
As large language models like ChatGPT continue to push the boundaries of what AI can do, a lot of people are struggling to keep up with the latest prompts, models and open-source alternatives. Here's a quick run down of some of the tools, communities and resources to help you experiment with or build on the latest options in conversational AI.
1- ChatGPT prompt templates repository:
A notion repository with 500 templates of chatgpt prompts that might come in handy
https://ignacio-velasquez.notion.site/ignacio-velasquez/500-ChatGPT-Prompt-Templates-d9541e901b2b4e8f800e819bdc0256da
2- Microsoft Edge dev Channel:
Microsoft is testing a ChatGPT-based tool for its Edge browser, that taps into web pages to generate content. The AI assistant can summarize, answer questions, draft emails or articles on the page's content (even PDFs) and in different styles.
https://www.microsoftedgeinsider.com/en-us/download/dev
3- ChatGPT Plugins ecosystem:
OpenAI wants to create an ecosystem of thirdy-party add-ons that give ChatGPT new extended functionality. If you're a builder and interested in building around this, join the waitlist:
https://openai.com/blog/chatgpt-plugins
4- Claude:
Claude is an AI assistant by Anthropic, a Google-funded startup, meant to rival ChatGPT. It lacks some of ChatGPT's capabilities (can't code) but is good at conversation and reasoning. Claude sounds more mature than ChatGPT, comes in 2 versions, and is cheaper than some of openAi's models. If you're looking for an an alternative to OpenAI's models, Check out Claude:
https://www.anthropic.com/product
5- Dolly:
Databricks just released Dolly, an open-source alternative to ChatGPT. Dolly can do Q&A, writing, and reasoning.
Here's the link to the original post:
https://www.databricks.com/blog/2023/03/24/hello-dolly-democratizing-magic-chatgpt-open-models.html
Here's the github for builders:
https://github.com/databrickslabs/dolly
6- OpenChatKit:
TogetherCompute released OpenChatKit, an open-source alternative to ChatGPT. Built using the GPT-NeoX-20B language model with 20 billion parameters, OpenChatKit was trained on over 43 million examples to handle NLP tasks.
Test the chat agent: https://openchatkit.net/
The complete source code is available here:
https://github.com/togethercomputer/OpenChatKit
For community support, join their discord channel https://lnkd.in/ewZXNm3P
7- Haystack, Langchain, Llama-index
If you want to build on recent progress in large language models, these top three open-source frameworks can help you integrate LLMs into applications, connect them with new data sources, or develop new capabilities using plug-ins and tools.
Haystack by deepset-ai:
https://haystack.deepset.ai/
Langchain:
https://python.langchain.com/en/latest/
Llama-index:
https://cloud.google.com/translate/docs/advanced/translating-text-v3
We’ve implemented initial support for plugins in ChatGPT. Plugins are tools designed specifically for language models with safety as a core principle, and help ChatGPT access up-to-date information, run computations, or use third-party services.