Immediately after finishing a Chinese drama, my brain has been split between languages. Somehow @OpenAI knows this and asked o1 to throw in some surprises for me.
There is definitely still prompting, tuning, guardrailing etc. that makes one company's model perform - but the entrance of open source SOTA is more about accessibility in this context. Slightly different moat.
It feels like the "AI moats" that are "falling" are a misnomer, because it's not quite the same as the traditional "tech moat" - it's more like willingness / ability to pay for SOTA models when they are expensive vs. creating proprietary tech
Many thanks to @alex for writing a deep-dive into our 2023 #OpenCloud report (and why we're so optimistic about the AI-driven path forward for the software industry).
Check out the full article in @TechCrunch here: https://t.co/Er567OwRlY
We’re excited to release the 2023 State of the #OpenCloud report, which explores the cloud landscape’s growth, driven by AI, with insights from @Dthakker02@DanelDayan and team.
For the full report, please visit: https://t.co/ijzZ2RDDIM
Interesting to see explainability in DL come full circle now that LLMs are top of mind - except we've gone from understanding the model to just pointing at an audit trail of the input data.
Lots of applications use the RAG style framework with @langchain and @pinecone
But how can you trouble shoot those applications?
Enter @arizeai
Excited to be doing a LangChain x Pinecone x Arize webinar on all this (feat @RLanceMartin)
Sign up 👇
https://t.co/OMXQvoLAQA