Is Perplexity AI Based on ChatGPT?
If you’ve found yourself scratching your head over the capabilities and origins of Perplexity AI, you’re not alone. The tech landscape can be a maze of AI tools and platforms, especially when it comes down to the generative models that have surged into the mainstream. So, is Perplexity AI based on ChatGPT? It’s a question worth exploring, especially when you consider what Perplexity brings to the table compared to its ChatGPT counterpart.
A Quick Snapshot of Perplexity AI
At its core, Perplexity AI is a conversational generative AI founded by Aravind Srinivas, a former research scientist at OpenAI—the very house that created the popular ChatGPT. So, while it’s not a direct clone, there are undeniable connections. Imagine the well-known feel of ChatGPT 3.5, made all the more powerful when it taps into the big, wild, unfiltered world of the internet. Unlike ChatGPT, which has data locked away after September 2021, Perplexity pulls from live information across the web, including sources like Reddit and the now-rebranded Twitter.
This live connection is not just a fancy add-on; it changes the game completely. You get not just the fabric of a conversational AI, but one that weaves in threads of real-time data from the internet, making it feel much more current and relevant. Although it has some gaps, when it comes to providing timely insights, it’s a feature that could happen to be invaluable to the user looking for answers beyond what’s in the AI’s training data.
The Backbone of Perplexity AI
So, the next question on your mind might be: « What’s under the hood? » As it turns out, Perplexity AI’s basic structure is built on the GPT-3.5 model. However, the additional online connection serves as both a strength and, interestingly, a weak point. While Perplexity can pull data for up-to-date information, there can also be a risk of misinformation or “hallucinations,” which is a term used in the AI world to describe when the system generates false information confidently.
Pros and Cons of Perplexity
Being on the cutting edge means wearing both the crown and the burden. Below are some noteworthy pros and cons of using Perplexity:
Pros
- Connected Online: With its accessibility to the internet, you get real-time information that’s beyond what ChatGPT can muster.
- Reddit Integration: The option to pull insights from platforms like Reddit makes it a bit of a social aggregator, summarizing discussions that are ongoing.
Cons
- Hallucinations: It’s not uncommon for the AI to sometimes spew forth factually inaccurate information.
- Data Synthesis: When faced with complicated questions, Perplexity may flounder in synthesizing the information effectively.
Price and Availability
Just for your knowledge, Perplexity AI is available for free, which adds to its appeal. It can be accessed via both a web interface or mobile application. Its exciting features include voice recognition, which makes interacting with it effortless, but note that image generation is restricted to the paid version only.
Comparative Analysis: AI Chatbot Performance
When it comes to assessing performance, a reputable platform like CNET takes a practical approach. They delve deep into real-world use cases that include cooking, shopping, and travel planning. Each AI gets scored on a 10-point scale, evaluating its accuracy, creativity, speed of response, and the number of times it hallucinates.
During these competitive comparisons, Perplexity shows strengths but often falters against the likes of Anthropic’s Claude and Microsoft’s Copilot. For instance, if you ask Perplexity for shopping advice between two similar TV models, it might pull some data, but the depth of analysis sometimes leaves room for improvement.
The Shopping Experience with Perplexity
Imagine you are torn between two TV models, say the LG C3 and G3. It would help immensely if you could get a concise synthesis of opinions and facts that highlight key differences. Ideally, an AI should sift through the mounds of user reviews and articles to bring you the essence of those discussions.
Perplexity does attempt this by citing credible sources like Rtings, Tom’s Guide, and WhatHiFi. However, it often struggles to extract the finer details that would help a user make an informed decision. For example, while it recommends the pricier G3 model if you can swing it, it fails to effectively contextualize why that might genuinely be worth the additional spend. The AI does mention that the G3 is « 70% brighter, » a fact that seems captivating until you realize it lacks the necessary context, which is paramount for wise consumer choices.
Furthermore, when delving deeper about past models—like asking to compare the older C9 and C3—Perplexity sometimes resorts to hallucinations, displaying a lack of clarity around certain technical details, such as the differences in technology that different OLED models utilize.
Recipes: A Culinary Challenge
Moving beyond the world of tech gadgets, you might be keen on cooking. Generative AI holds the potential to revolutionize the way we access recipes. However, a considerable number of online recipes can be bogged down with superfluous chatter that’s crafted for SEO purposes. So, can Perplexity deliver here?
In terms of recipe generation—say, a madhouse of chicken tikka masala—Perplexity does put forth a somewhat acceptable marinade at first but misses key ingredients, like chili powder. When prompted again, it managed to include what was missing, but this kind of inconsistency reflects issues with reliability that both users and full-fledged chefs generally seek in a reliable assistant.
Research and Accuracy
One of the significant advantages of Perplexity is that it doesn’t just regurgitate information; it can link to actual sources. So, if you’re asking about the impact of homeschooling on neuroplasticity, you could end up with actual academic papers to glance through. While it might stumble upon less rigorous sources, like advocacy pages, its ability to deliver scholarly citations is definitely a feather in its cap.
However, much like a multitasker pushing too many plates at once, Perplexity often fails to make connections between multiple research findings, resulting in a mishmash of facts without a thorough synthesis—again, a shortcoming that could confuse users looking for concrete answers.
The Summarization Struggle
Rounding off Perplexity’s offerings is its ability—or lack thereof—to summarize articles effectively. While it can capture the gist, it sometimes misses the main arguments that are essential to grasping the author’s intent. In an era where summaries save time, this limitation quickly rises to the forefront as a point of contention.
Final Thoughts
So, is Perplexity AI built on ChatGPT? In a nutshell, it’s a product that is fundamentally rooted in technology developed at OpenAI but takes it a step further by incorporating real-time data. While it has its perks, such as live sourcing and the ability to delve deeply into user-generated content on platforms like Reddit, users must be mindful of its limitations regarding accuracy and nuanced understanding. Like any AI, it has strengths and weaknesses, and only thorough engagement can help you navigate the foggy landscape of generative AI.
As the world of AI continues to evolve, both ChatGPT and Perplexity will likely adapt to better serve their users—providing not just information, but context, accuracy, and reliability, something that all of us desperately seek whether purchasing a new TV or whipping up dinner for the family. Until then, use these platforms wisely, combining their insights with a sprinkle of skepticism!