We went from only one GPT-4 class model to four this year. Next move is OpenAI’s, which will tell us what the future looks like- plateauing technology or continued rapid development.
AI
-

Open-Source GPT-4-Level Models Now Freely Accessible Globally
By
–
We're entering a new world where GPT-4-level models are open-source and freely accessible. Absolutely massive.
-
ChatGPT vocabulary bias traced to Nigerian RLHF annotation workers
By
–
Love this theory by @alexhern that the reason ChatGPT uses words like "delve" a lot is that OpenAI outsource a lot of their RLHF annotation to workers in Nigeria, and Nigerian formal English has a slightly different vocabulary
-
Training 400B+ Parameter Model to Outperform GPT-4
By
–
They're training a 400B+ parameter version that should outperform GPT-4.
-

LangChain Mistral RAG Agent Cookbooks Released
By
–
LangChain x Mistral RAG Agent Cookbooks + With the release of new Mixtral 8x22B, there's high interest in building agents with open source LLMs. We partnered with @MistralAI to release a set of cookbooks and a video that shows how to do this using control flows in
-

LLaMA 3 400B Matches Claude 3 Opus Performance
By
–
The craziest LLaMA 3 reveal: The 400B+ version of the model is **on par with Claude 3 Opus**, and it's still training. Soon, we'll have a better-than-Opus, fully open-source model. The implications are huge.
-
Charging AI Bots at $100/mo with Cost Savings
By
–
Means we can charge AI bots at even 100$/mo and provide cost savings! Crazy.
-
SAS Innovate Day 2: Vision AI Innovation Showcase
By
–
Day 2 at #SASInnovate. Follow us to learn more about #sasvisionay pic.twitter.com/pekhpeAVzi
— Helen Yu (@YuHelenYu) 18 avril 2024Day 2 at #SASInnovate. Follow us to learn more about #sasvisionay
-

LLaMA 3 70B Outperforms Claude 3 Sonnet Cost-Effectively
By
–
Holy shit. LLaMA 3 70B cleanly beats Claude 3 Sonnet. Small enough to host at scale without breaking the bank.
