Knowledge Intelligence and AI Tendencies: High merchandise, RAG and extra

Generative AI fever reveals no indicators of cooling off. As stress and pleasure construct to execute sturdy GenAI methods, knowledge leaders and practitioners are looking for the most effective platform, instruments and use instances to assist them get there.

How is that this taking part in out in the actual world? We simply launched the 2024 State of Knowledge + AI, which leverages knowledge from our 10,000 international clients, to grasp how organizations throughout industries are approaching AI. Whereas our report covers a broad vary of themes related to any data-driven firm, clear traits emerged on the GenAI journey.

Right here’s a snapshot of what we found:

High 10 knowledge and AI merchandise: the GenAI Stack is forming

With any new know-how, builders will experiment with a lot of totally different instruments to determine what works finest for them.

Our High 10 Knowledge and AI Merchandise showcase essentially the most widely-adopted integrations on the Databricks Knowledge Intelligence Platform. From knowledge integration to mannequin growth, this listing reveals how corporations are investing of their stack to assist new GenAI priorities:

Hugging Face transformers leap to No. 2

In simply 12 months, Hugging Face jumps from spot #4 to identify #2. Many corporations use the open supply platform’s transformer fashions along with their enterprise knowledge to construct and fine-tune basis fashions. 

LangChain turns into prime product months after integration

LangChain, an open supply toolchain for working with and constructing proprietary LLMs, rose to identify #4 in lower than one yr of integration. When corporations construct their very own trendy LLM purposes and work with specialised transformer-related Python libraries to coach the fashions, LangChain permits them to develop immediate interfaces or integrations to different techniques.

Enterprise GenAI is all about customizing LLMs

Final yr, our knowledge confirmed SaaS LLMs because the “it” instrument of 2023, when analyzing the most well-liked LLM Python libraries. This yr, our knowledge is exhibiting using general- objective LLMs proceed however with slowed year-over-year progress.

This yr’s technique has taken a serious shift. Our knowledge reveals that corporations are hyper-focused on augmenting LLMs with their customized knowledge vs simply utilizing standalone off-the-shelf LLMs.

Companies need to harness the facility of SaaS LLMs, but in addition enhance the accuracy and mildew the underlying fashions to raised work for them. With RAG, corporations can use one thing like an worker handbook or their very own monetary statements so the mannequin can begin to generate outputs which can be particular to the enterprise. And there’s large demand throughout our clients to construct these personalized techniques. Using vector databases, a significant element of RAG fashions, grew 377% within the final yr – together with a 186% leap after Databricks Vector Search went into Public Preview.

The enterprise AI technique and open LLMs

As corporations construct their know-how stacks, open supply is making its mark. In reality, 9 of the ten prime merchandise are open supply, together with our two large GenAI gamers: Hugging Face and LangChain.

Open-source LLMs additionally provide many enterprise advantages, resembling the power to customise them to your group’s distinctive wants and use instances. We analyzed the open supply mannequin utilization of Meta Llama and Mistral, the 2 greatest gamers, to grasp which fashions corporations gravitated towards.

With every mannequin, there’s a trade-off between price, latency and efficiency. Collectively, utilization of the 2 smallest Meta Llama 2 fashions (7B and 13B) is considerably larger than the most important, Meta Llama 2 70B.

Throughout each Llama and Mistral customers, 77% select fashions with 13B parameters or fewer. This implies that corporations care considerably about price and latency.

Dive deeper into these and different traits within the 2024 State of Knowledge + AI. Contemplate it your playbook for an efficient knowledge and AI technique. Obtain the total report right here.