Skip to content

The Race to Customize Large Language Models: Insights from Databricks’ State of Data+AI Report June 2024

Databricks just published a report on the State of Data + AI. I found it an interesting read so I wrote this summary for you to read if you’re interested.

You have heard about the super cool artificial intelligence called ChatGPT. It’s like having a knowledgeable assistant that can help with all sorts of tasks just by asking it questions. 

Well, companies across many different industries are really excited about this new AI technology too. They want to use it to create better products, services, and experiences for their customers. But here’s the thing – the regular ChatGPT doesn’t actually know anything about a specific company’s business or data. It’s like a super smart student who hasn’t specialized in a particular subject yet.

That’s where the findings from Databricks’ new State of Data+AI Report come in. This report looked at data from over 10,000 companies, including huge names like Comcast, Shell, and many others. And it reveals how these companies are working to make AI like Chat GPT even smarter and more useful by feeding it their own data and information.

One of the biggest trends is something called retrieval augmented generation or RAG for short.

Basically, it’s a way for companies to give AI assistants access to their databases and documents. So instead of just general knowledge, the AI can directly reference a company’s proprietary data to provide way more accurate and tailored responses. It’s like giving the smart student all the textbooks for a particular class!

The report found that the use of special databases called vector databases, which are key for RAG, shot up by a whopping 377% in just one year! 

Companies are clearly seeing major benefits in customizing AI this way.

Interestingly, most companies actually prefer using open source AI models that anyone can access and modify, rather than proprietary ones from big tech companies. Open source allows more flexibility to tweak and customize the AI to fit their specific needs.

And get this – some of the industries most eager to use this customized AI are ones you might not expect; like finance, healthcare, and others with lots of rules and regulations. 

Companies are now recognizing how powerful and useful AI can be when it deeply understands their unique data and operations.

The report also showed that in general, companies are getting much better at taking the artificial intelligence models they experiment with and actually putting them into real-world production uses for their customers. And industries like retail and consumer goods are leading the way in efficiency here.

So in summary, the AI revolution is being fueled by companies that realize their own data is the key ingredient. By feeding it to AI assistants through methods like RAG, they can create personalized, smart experiences that raise the bar for customers across all kinds of industries and businesses.

It’s an exciting time where human intelligence is enhancing artificial intelligence in creative new ways.

I don’t know about you, but I can’t wait to see what these supercharged AI assistants will be able to do next!

Let me know if you have any questions.

Leave a Reply

Your email address will not be published. Required fields are marked *