Perplexity's New AI Brain: DeepSeek R1 Integration is a Game Changer

Perplexity, a search engine I've been using a lot lately, just added the DeepSeek R1 reasoning model. This is a pretty big deal, and it means you can now use this new DeepSeek model right inside Perplexity. I'm going to show you how it works with some examples.

Perplexity Pro is a paid upgrade, and you get a few free uses of the R1 model each day. After that, you'll need to upgrade, which costs $20 a month. But honestly, I use Perplexity for all my searches anyway, so I think the Pro upgrade is totally worth it. You can even attach files to your searches. Plus, Perplexity also offers OpenAI's reasoning model, which is pretty neat.

Testing the R1 Model: Nvidia Stock Example

Let's say I want to figure out if Nvidia stock is a good buy right now. (Just a heads-up, this isn't financial advice, but it's a good way to see how the R1 model helps with research.) When I send this prompt, the AI starts by explaining its process. It breaks down what it's going to do, like looking at recent prices, comparing good and bad outlooks, and checking predictions for 2025. It pulled information from 22 sources and gave a verdict: Nvidia is a high-risk, high-reward play. Long-term investors might see a dip as a chance to buy in, but short-term traders should expect some ups and downs.

One of the best things about Perplexity is the related prompts it suggests. I asked about the long-term impact of DeepSeek on Nvidia's market share, and with the R1 model turned on, it found 23 sources. It analyzed Nvidia's strategy to deal with competition, noting that while these impacts are big, Nvidia still holds a strong position in the AI chip market thanks to its established ecosystem and ongoing innovation.

Comparing Search Results: Inflation and Currency Trends

Perplexity is a strong search engine even without the Pro features, but let's see how R1 handles trend analysis and predictions. I asked it to find current inflation rates in the US, UK, and Canada and predict which currency might strengthen most by year-end. The regular search gave a quick answer: the US dollar might strengthen the most due to rising inflation. It provided a list and a conclusion.

Now, I'll use the "rewrite" function and ask for the same thing but with the R1 reasoning model. The R1 model broke down the information much more cleanly, giving more data points. The conclusion was clearer, stating the US dollar is expected to strengthen. It's impressive how much faster this speeds up research.

Decision Making: Software Engineering Career Paths

This is where combining search with AI reasoning really shines, something you can't do in ChatGPT right now. Even in the $200/month version of ChatGPT with the 01 model, search isn't available. Perplexity lets you use 01, R1, and search all together, making it a super useful AI tool.

I asked Perplexity to search for the latest job trends for software engineers and recommend a specialization for someone starting in 2025, based on demand and salaries. The R1 model went through several steps, acting like a search agent. It identified AI and machine learning engineering as the top recommendation, with average salaries over $200k and significant growth potential. It advised gaining practical skills, focusing on Python, and staying updated with AI, cloud, and cybersecurity.

Creative Problem Solving: Freelancer Taxes in California

For creative problem-solving, I asked about recent laws affecting freelancers in California and how someone earning $100k should structure their taxes. The R1 model found new laws and provided specific tax strategies: forming an LLC, making an S-corp election to save on self-employment taxes, making estimated quarterly payments, and setting up retirement accounts like a SEP or Solo 401(k).

It pulled information from 25 different sources and provided step-by-step reasoning. The amount of search and analysis it did in such a short time is incredible.

Ethical Reasoning: AI in Healthcare

Finally, I looked at an ethical question: the pros and cons of AI in healthcare and how hospitals should create guidelines for diagnostic AI tools. Using R1, it analyzed 19 sources and broke down the benefits of AI in healthcare, along with risks and challenges like ethical dilemmas, privacy concerns, and clinical risks. It also proposed ethical guidelines, emphasizing transparency. The takeaway was that by balancing innovation with safeguards, healthcare systems can use AI effectively while keeping patients safe.

I'll put some of these prompts in the description below so you can get ideas for your own searches. I think you'll find it really useful. Thanks for watching!