I tested NotebookLM vs. Perplexity for deep research with 5 difficult prompts — here’s the clear winner
Perplexity AI and NotebookLM are two AI tools that are changing the way we dig into research, but they do take slightly different approaches.
Perplexity’s Deep Research feature is like having a super-efficient research assistant. You ask it a question, and it scours multiple sources, summarizes the findings, and even links you to the original content so you can dig deeper if you want. It’s great for quick answers accompanied by context and citations.
Google’s NotebookLM, on the other hand, is more of an overly keen teacher. It pulls out documents, makes videos and podcasts, and fires off twenty different versions of the information you need.
1. How to keep houseplants healthy
Prompt: Identify the best proven methods for keeping common houseplants healthy in low-light apartments. Compare plant species, care routines, watering strategies and any credible horticultural research on light thresholds.
NotebookLM: Here, deep research feels more like sitting at a digital school desk rather than asking a quick question online, which can be impressive but a bit much for a first-time user.
It didn’t just answer the question; it consolidated a list of PDFs and web sources available for personalized selection to inform the pending report, offering 23 sources and even a ‘five pillars of care’ diagram.
It’s overwhelming, but the detail is remarkable, especially with the AI-generated video.
Perplexity: On the other hand, Perplexity delivered concise and digestible results.
Though not as comprehensive as NotebookLM, it was more manageable and actionable. Over on the Links tab, it provided direct sources to further information, inviting me to explore the subject at my own pace.
Though this required more effort on my part through some light reading, I appreciated the source-backed list and came away with an answer far quicker than I would have with NotebookLM.
Winner: NotebookLM
While it can be overwhelming, NotebookLM’s response went the extra mile to answer my question, offering additional learning resources to guide me.
2. Comparing breakfast cereals
Prompt: Compare three major breakfast cereal brands for taste, nutrition and price. Cite independent nutrition analyses, consumer taste tests and recent market pricing data, and provide a clear ranking supported by evidence.
NotebookLM: Once again, NotebookLM immediately feels like a guided workspace for multi-source digging. When I asked it to analyze the top three breakfast cereals, it first introduced its findings.
Then, it provided a detailed breakdown combining sales rankings, nutritional criteria like fiber and sugar, and even taste, drawn from 22 sources.
Perplexity: Perplexity’s response, meanwhile, takes a very different approach. Instead of building a workspace around the topic, it cuts straight to a structured answer with tables, charts and sourced evidence all in one scroll.
It also generated simple diagrams, showing the key pieces of information in easy-to-digest forms.
It even calculated monthly cost differences and added health-based considerations like fibre tolerance for sensitive digestion. Everything is neatly tied together and easy to follow, with citations linked out and generated files clearly labelled at the top.
Winner: Perplexity
Perplexity created a simple yet effective response here, gathering all of the information into one clean format. While NotebookLM delivered as well, it felt like overkill, offering so much extra information that it became overwhelming.
3. Evaluating public transit apps
Prompt: Evaluate which public transit app provides the most accurate real-time information for Bristol City in the UK. Assess data sources, update frequency, user-reported accuracy, privacy practices and any third-party audits.
NotebookLM: After waiting a few minutes for the report on NotebookLM, I was suddenly confused: “These documents collectively outline the landscape of public transport data and real-time information systems, primarily focusing on bus services in the UK and Singapore.” My prompt was clear about the city in question, so I am not sure why it mentioned Singapore.
The infographic itself shows how open data standards boost the accuracy and frequency of real-time bus information for all transit apps using UK sources. However, it doesn’t pick a ‘best’ app or address user reviews, privacy, or app-specific audits; so you may need extra research for that.
Perplexity: Perplexity, meanwhile, immediately provided a straightforward answer in its summary: “For Bristol residents, Bustimes.org emerges as the most reliable source for real-time bus information, combining data from multiple official sources with strong transparency practices…”
It followed up with Key Transit Apps in Bristol, Data Sources and Update Frequency, User-Reported Accuracy Issues, Data Quality, Privacy Practices, and other related sections.
Winner: Perplexity
While text-heavy, Perplexity ultimately delivered the answer I was looking for, which I was able to extract from the summary alone. In this instance, NotebookLM seemed to be all show, no tell.
4. 10-min healthy snacks
Prompt: List healthy snacks I can make at home in under 10 minutes. Prioritize evidence-based nutrition guidance, include ingredient lists, and explain the nutritional trade-offs among the options.
NotebookLM: Citing 22 sources, NotebookLM produced a summary titled ‘Satiety and Strategic Snacking: Protein, Fat, and Weight Management’.
It immediately seemed more convoluted than it needed to be, and once again did not actually deliver the response I was looking for. It offered pages and pages of information… basically everything but the actual healthy snacks suggestions I mentioned.
Perplexity: Perplexity did not disappoint in this area, as it delivered precisely what was requested and proved it: “Here are several healthy snacks you can make at home in under 10 minutes, each with their ingredient lists and evidence-based nutritional guidance. The options provided prioritize whole foods and balance macronutrients, with an explanation of their nutritional trade-offs.”
The Deep Research result produced 10 recipes, followed by their respective ingredients, nutrition, and trade-offs.
And over in the Links and Images tabs, I was directed to a variety of credible sources that complemented the summary findings in the Answer tab.
Winner: Perplexity
Perplexity actually completed the task I asked of it. Once again, NotebookLM was all style, but lacking in some of the key substance that was needed.
5. Reducing household electricity in winter
Prompt: Explain the most effective ways to reduce household electricity use in winter. Focus on data-backed strategies, cost savings, insulation improvements, appliance efficiency and behavioral changes supported by credible studies.
NotebookLM: NotebookLM’s summary offers a comprehensive overview of “residential energy consumption reduction strategies, emphasising both technological upgrades and behavioural changes,” which is on track for delivering a relevant result.
It also provided a slide show and even an engaging podcast, which was remarkable. Concerning household energy efficiency, NotebookLM really delivered.
Perplexity: Perplexity’s result was notably more dull in terms of layout, interactivity, and readability. Given the complex topic of energy efficiency itself, I was at least expecting some charts or graphs to help illustrates the points.
However, it leaned heavily on text and scrolling that seemed endless. There wasn’t much to grab your attention and keep you hooked or engaged, unless you consciously navigated to the Links or Images tabs, which provided some blogs and websites to corroborate its findings, and the Images revealed source-backed charts.
Winner: NotebookLM
Here, Notebook’s extra detail pays off, fully answering my query, but also doing it in a natural and interesting way.
Overall winner: Perplexity
Both models went above and beyond to answer my questions here. While it was a close battle, Perplexity just about took the win.
This was thanks to Perplexity’s focus on the details, compared to NotebookLM’s tendency for flare. Sometimes this works great, developing a detailed and interesting report, other times it’s focus led it in the wrong direction.
This could also end up being information overload at certain times. While Perplexity wasn’t quite as interesting, it was overall more helpful.
Follow Digitpatrox on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Digitpatrox
Back to Laptops
Source link