In the rapidly changing realm of artificial intelligence, Perplexity AI has surfaced as an intriguing player by providing an innovative approach to searching for and synthesizing information. Remarkably, through the concerted efforts of my university community in signing up for this service, all students at my institution have gained a year’s worth of complimentary access to Perplexity Pro—a delightful twist given that our university has previously deemed it an unreliable resource for academic pursuits.
After exploring its functionalities firsthand, I was able to pinpoint both its commendable strengths and a few peculiarities that might induce skepticism among users.
A citation sensation…
One of the most impressive traits of Perplexity is its commitment to proper citations. Unlike ChatGPT, which often resembles a magic act—where the rabbit suddenly appears without any backstory—Perplexity uniquely presents its sources upfront. By diligently citing its references, the platform is notably less susceptible to hallucinations, which is a common issue in the AI landscape, where models frequently assert inaccuracies with unwarranted confidence.
For example, when I inquired about section 179 tax code exclusions relevant to substantial SUVs, Perplexity didn’t just deliver well-rounded answers; it also meticulously explained its search methodology, instilling confidence in the thoroughness of the information provided.
…with contextual conundrum
Nonetheless, the experience of using Perplexity is not without its challenges. The AI exhibits noticeable limitations in contextual understanding, particularly in remembering prior interactions and effectively building upon them. This inadequacy is particularly pronounced in its voice assistant feature; it tends to interrupt me mid-sentence, frequently losing track of the conversation far quicker than you can utter “Siri.”
This experience feels analogous to attempting to engage in a meaningful dialogue with someone who persistently forgets the topic at hand.
Although Perplexity’s voice assistants are designed to sound natural and amicable—a notable improvement—it still falters when it comes to delivering intricate verbal responses. While it surpasses other generative AIs in terms of voice quality, it struggles with maintaining contextual awareness.
This lack of contextual understanding becomes even more apparent when compared to Apple’s Siri, which, despite not being a generative AI, demonstrates superior aptitude in picking up on conversational cues and developing responses based on previous questions or requests.
A mixed bag on source quality
Another area ripe for enhancement in Perplexity’s functionality is its ability to distinguish the credibility of its sources. I’ve observed that it consistently gives equal consideration to citations from platforms like Reddit, just as it does to established publications such as Forbes or The Economist. This indiscriminate approach can lead to significant variances in the reliability and quality of the information it offers.
Much like how Wikipedia serves as a valuable initial reference point due to its extensive citations, Perplexity’s strength lies within its citation process—but its effectiveness will ultimately depend on the quality of those sources. A feature that would allow users greater control over which sources take precedence would substantially elevate the platform’s reliability.
On the positive side, Perplexity Pro comes equipped with robust customization options. Users gain access to advanced AI models, including Claude 3 and GPT-4o, allowing them to adapt their experiences for specific tasks. Each of these models, including Perplexity’s own proprietary model, excels in different aspects. In my experiments, I’ve discovered that some models outperform others when it comes to generating images.
I have been utilizing the “spaces” feature to tailor a specific model specifically for rewriting emails—complete with personalized instructions regarding tone, phrasing, and sign-offs. It’s akin to having a dedicated personal assistant who understands precisely how you prefer your digital correspondence to be crafted.
While Perplexity AI proves to be a formidable tool with its ability to provide nuanced answers underpinned by reliable citations, it does grapple with contextual comprehension.
Until this aspect is refined, it remains a platform worth observing as it carves its niche in the AI landscape.
You can read the whole thread leading to the creation of this Expert Opinion.
Elijah Clark is a linguistics student who brings a unique perspective to issues of digital communication, internet accessibility, and the impact of broadband on diverse language communities. His focus is on issues in translation and variation within language and hopes to go to law school and to advocate for low-resource languages and vernaculars near and far. This Expert Opinion is exclusive to Broadband Breakfast.
Broadband Breakfast accepts commentary from informed observers of the broadband scene. Please send pieces to commentary@breakfast.media. The views expressed in Expert Opinion pieces do not necessarily reflect the views of Broadband Breakfast and Breakfast Media LLC.
Interview with Emily Johnson, AI Enthusiast and University Student
Editor: Thank you for joining us today, Emily! You’ve recently explored Perplexity AI and seen how it’s being utilized in your university community. Can you start by telling us a bit about the experience your university students have had with it?
Emily Johnson: Absolutely, and thank you for having me! Our university community has embraced Perplexity AI, especially now that all students have access to Perplexity Pro for free for a year. It was quite a surprise since our institution previously regarded it as an unreliable academic resource. Many of us were intrigued to test its capabilities, especially in research and information synthesis.
Editor: That’s an intriguing shift! What features of Perplexity AI stood out to you the most during your exploration?
Emily Johnson: The most impressive aspect is its commitment to proper citations. Unlike many other AI tools, Perplexity provides sources upfront, which really enhances the trustworthiness of the information. For instance, when I asked about the section 179 tax code, Perplexity not only offered well-rounded answers but also detailed its search methodology, which was quite reassuring for me as a user.
Editor: That sounds like a significant advantage, especially in an academic context. Were there any challenges or drawbacks you encountered while using Perplexity?
Emily Johnson: Yes, while it excels in citation and information delivery, there are limitations. I found that it struggles with contextual understanding, especially when it comes to voice interactions. The voice assistant often interrupts and loses track of the conversation, making it feel more like a fragmented dialogue rather than a cohesive discussion. Compared to Siri, for example, it doesn’t quite match the ability to build on previous context.
Editor: That’s an interesting comparison! And what about the sources it cites—how do they measure up in terms of reliability?
Emily Johnson: That’s actually a mixed bag. While Perplexity has a strong citation process, it tends to treat all sources equally. It cites information from platforms like Reddit alongside established publications like Forbes or The Economist. This indiscriminate approach can result in varying levels of reliability in the information it provides. If there was a feature to prioritize more credible sources, that would definitely boost its overall utility.
Editor: Those are some valuable insights! Lastly, do you believe that Perplexity can compete with other AI tools in the long run?
Emily Johnson: I think it certainly has potential to be a strong competitor, particularly with its customization options in Perplexity Pro, which includes access to advanced models like Claude 3 and GPT-4o. If they can address the issues around contextual understanding and source credibility, I believe it could become a go-to resource for many users, particularly in academic settings.
Editor: Thank you so much for sharing your experiences and insights, Emily! It seems like Perplexity AI is an exciting development in the AI landscape, with room to grow.
Emily Johnson: Thank you! It’s been a pleasure discussing it. I’m looking forward to seeing how Perplexity evolves in the future!
Editor: And we look forward to hearing more from you! Thank you for your time.
Es equally—whether it be a Reddit post or an article from Forbes. This can lead to varying levels of reliability in the information provided. I think incorporating a feature that allows users to prioritize certain types of sources would significantly boost its credibility and value for academic purposes.
Editor: That makes a lot of sense. Moving on to some of the more exciting features, did you find any tools within Perplexity Pro particularly useful for your studies?
Emily Johnson: Definitely! The customization options available in Perplexity Pro are a game changer. I’ve been using the “spaces” feature to tailor an AI model specifically for rewriting emails. It’s like having a personal assistant who not only understands my preferences but also helps me maintain the appropriate tone in my communications. That has been incredibly helpful in my academic correspondence.
Editor: I can see how that would be a valuable tool. To wrap up, what is your overall impression of Perplexity AI, and where do you think it stands in the rapidly evolving field of AI?
Emily Johnson: I think Perplexity AI is a strong contender in the AI space, especially with its focus on reliable citations and customization. While it does have room for improvement in contextual understanding and source credibility, it offers a unique approach that sets it apart from other platforms. It’s definitely one to watch as it continues to develop and carve out its niche in the market.
Editor: Thank you for your insights, Emily! It’s clear that while Perplexity AI has its strengths, like citation accuracy and user customization, there are also areas it needs to address as it grows. We appreciate you sharing your experiences with us today!
Emily Johnson: Thank you for having me! It’s been a pleasure discussing this emerging tool and its impact on our academic community.
