I wanted to try DeepSeek today to explore its reasoning capability, but it says that it's too busy to help little old me. Maybe later?
Then I recalled that I had seen an article on Reddit about Perplexity using DeepSeek's second-best reasoning model for free, so I tried that. The article noted that Perplexity said:
- We launched Sonar Reasoning – Perplexity's new API that’s powered by DeepSeek's reasoning models.
- With Sonar Reasoning, you can build products with chain-of-thought reasoning, real-time web search, and cited answers–all backed by our commitment to privacy and data protection.
- A few important details to note about Sonar Reasoning:
- Our models are hosted in US data centers
- Our models are uncensored
- Zero data collection or sharing. Period.
- Ready to build? Check out our documentation and start integrating today.
1) The process to use the DeepSeek reasoning model on Perplexity is:
1) Go to https://labs.perplexity.ai.
2) Click the "Sonar" link in the upper right-hand corner of the screen.
3) In the lower right-hand corner, select "Sonar-Reasoning" and then ask your question.
Here is the screen (showing the "Sonar" and "Sonar-Reasoning" fields after I asked my question:
2) My prompt was: what technology improvements in genealogy research, family history stories, and autosomal DNA matching and analysis are expected in the next 5 years?
You can see the reasoning start in the graphic above. Here is the complete list of reasoning and then the answer to my prompt.
<think>
Okay, so I need to figure out what technology improvements are expected in genealogy research, family history stories, and autosomal DNA matching and analysis over the next five years. Let me start by going through each of the search results provided to gather relevant information.
First, looking at result [1], it talks about FamilySearch's plans for 2025. They mention AI advancements, specifically in areas like reading old handwriting in more languages and improving lineage-linked data suggestions. They also have a feature called Full-Text Search that can look through unindexed, handwritten records. Additionally, they're expanding their AI for indexing historical records, which should make more records available faster. They're also adding more records from various countries and working on oral genealogies, making them searchable. Plus, they're digitizing more books. So, this seems relevant for genealogy research improvements.
Result [2] introduces a platform called Storied, which focuses on family history stories beyond just family trees. It emphasizes connecting people through shared stories, including non-family influences. They use a graph architecture to allow more connection types, which could help in creating richer, more detailed stories. This seems important for the family history stories part of the query.
Result [3] is about genealogical DNA tests, specifically autosomal DNA. It explains how autosomal DNA works, the testing process, and how matches are identified. It also mentions the use of centimorgans (cM) to measure DNA segments and the challenges in interpreting matches, especially for distant relatives. This is directly related to autosomal DNA matching and analysis.
Result [4] discusses AI in genealogy, highlighting MyHeritage's use of AI for DNA matches, analyzing vast data, and tasks like translation and transcription. It also predicts future AI uses, such as DNA triangulation tools, converting handwriting to text, creating source citations, and identifying ancestors in photos. This covers both genealogy research and DNA analysis.
Result [5] talks about the future of genealogy with GPTs, including virtual recreations of ancestral environments and improved accuracy in research through advanced algorithms. It also mentions AI-generated narratives and accessing a wealth of information. This ties into both research and storytelling.
Putting this all together, I can structure the answer into sections: Genealogy Research, Family History Stories, and Autosomal DNA. Each section will pull relevant points from the search results. For Genealogy Research, AI advancements, more records, and better tools are key. For Family History Stories, platforms that focus on storytelling and connections beyond direct relatives are important. For DNA, improvements in matching, analysis tools, and AI applications will enhance the field.
I need to make sure each point is cited correctly and that the answer is concise, using the information from the search results without copying verbatim. Also, the formatting should be clear, using headers and lists as needed, but not starting with a header. I'll avoid any markdown in the thinking part but structure the final answer accordingly.
</think>
## Technology Improvements in Genealogy Research, Family History Stories, and Autosomal DNA Matching and Analysis
### Genealogy Research1. **AI Advancements**: FamilySearch is expanding AI capabilities to read old handwriting in multiple languages and improve lineage-linked data suggestions. Their Full-Text Search feature will enable searching of unindexed, handwritten records, enhancing research efficiency [1].### Family History Stories
2. **Increased Record Availability**: Expect significant additions to record collections for countries like France, Germany, and Italy, along with oral genealogies from African countries, making these searchable for the first time [1].
3. **Digitization of Historical Books**: FamilySearch will continue digitizing genealogical books, expanding its Digital Library with full-text searchable publications [1].
4. **AI-Driven Data Analysis**: MyHeritage leverages AI to analyze vast data, discovering new connections and understanding migration patterns, enhancing research depth [4].1. **Story-Centric Platforms**: Platforms like Storied focus on creating connections through shared stories, including non-family influences, using graph architecture for diverse connections [2].### Autosomal DNA Matching and Analysis
2. **AI-Generated Narratives**: GPTs are expected to generate detailed narratives of ancestors' lives, offering virtual recreations of ancestral environments for immersive experiences [5].1. **Improved Matching Algorithms**: MyHeritage uses AI to enhance DNA matches, potentially leading to advanced triangulation tools for precise relationship determination [4].These advancements promise to revolutionize genealogy, making research more efficient and enriching family history narratives with technology.
2. **Predictive Capabilities**: AI algorithms will predict familial connections and spot errors, aiding in filling gaps in family trees [5].
3. **Enhanced Analysis Tools**: Future tools may include chromosome browsers and triangulation features for deeper DNA analysis, improving match interpretation [3].
You can see the "chain of thought," real-time web search and the sources that it used to generate the answer, including what it learned from each source. Then it wrote short summaries to answer my query. The source numbers are also used in the short summaries.
The five sources that it used are:
[1] https://www.thechurchnews.com/living-faith/2025/01/06/what-to-expect-from-familysearch-in-2025/
[3] https://en.wikipedia.org/wiki/Genealogical_DNA_test
[5] https://coachcaroleonline.com/exploring-the-future-of-genealogy/
3) So how did this reasoning model on Perplexity do? My impression is that it took five (only 5?) web search results and summaried them. The first article talked about RootsTech 2025 expectations for FamilySearch. The next two articles are somewhat dated, from 2023 and don't discuss the future. The 4th article discusses the state-of-the-art in 2024, and the fifth article discusses future applications.
However, because it relies on some published articles, it can only report what they say. Therefore, articles about future technology and genealogy have not been written yet!
6) I asked the same question in the free Perplexity and it provided different and more pertinent answers to my query, including from the above source list and these additional articles:
- https://aimresearch.co/ai-startups/storii-uses-ai-to-turn-family-memories-into-stories-that-last-forever
- https://pmc.ncbi.nlm.nih.gov/articles/PMC4580997/
- https://www.reddit.com/r/Genealogy/comments/1753lcc/how_will_tech_impact_genealogy_over_the_next_10/
7) This certainly seems to be safer and easier to use than DeepSeek, although it is not the best DeepSeek model.
This Perplexity reasoning model would not allow me to chat with it for some reason - I wanted to ask it to add more descriptive information about each numbered point.
8) Now I will try to do some real genealogy work with the Perplexity reasoning model and will report on it soon.
NOTE: I use only freely available AI tools at this time. If you subscribe to AI tools, your results may be better.
My blog posts, and those of other genealogists using Artificial Intelligence tools are on my AI and Genealogy Compendium page.
The URL for this post is: https://www.geneamusings.com/2025/01/ask-reasoning-ai-using-perplexitys.html
Copyright (c) 2025, Randall J. Seaver
Please comment on this post on the website by clicking the URL above and then the "Comments" link at the bottom of each post. Share it on Twitter, Facebook, or Pinterest using the icons below. Or contact me by email at randy.seaver@gmail.com. Please note that all Comments are moderated so they may not appear immediately.
Subscribe to receive a free daily email from Genea-Musings using www.Blogtrottr.com.
No comments:
Post a Comment