blog

2025 - DeepSeek r1 Reminds the World we don't need Expensive Data Centres and per-transaction fees to benefit from AI

Written by Jason W. D. Cassidy | Jan 29, 2025

Wanna see DeepSeek R1 in action?

Private and secure data behind your firewall.
Zero per-API call cost
Scalable to millions of documents at no extra cost
Scalable to thousands of users at no extra cost

Here's a fun example.

Imagine the productivity gains if you didn't have to do teams calls or meetings to collect information.  You could just ask your computer network a question, and have it give you the answer without even reading the source document.

Problem Statement: "What Demographic Donates the Most Money to Charity"?

Steps  

1.  Use Shinydocs Search to narrow document context to charity related documents in our data set.

simple search for "charitable giving"

We've got some search results for "charitable giving", click "Ask AI" to ask the documents a question...


Type in "Who donates the most?"


The Shinydocs search engine quickly loads the document information into the DeepSeek R1's model's brain and asks the "Who donates the most?" question.  The model does some computation (thinks about it).


This is where DeekSeek R1 shines - it gives its thinking behind its answer very clearly.


And here's the answer it gives.


For reference - here's where it found the info it gave in the answer. When I look at the original document(s) and from the clear descriptions of its thought process, we can verify where it got this information.

Video: The whole process start to finish in real time...

 

How did we do this?

Use Shinydocs Pro (on prem, your cloud, or public cloud) to privately and securely classify and make searchable all documents and files on your network for you and/or your users.

Use the optional Shinydocs Pro feature "Ask AI".  This requires a little extra server side hardware (at least one decent video card for AI processing).

That's it.