Funny Joe used the phrase ' garbage in, garbage out' without realizing that is the truth, his query was garbage, so Perplexity gave him a garbage response.
Presenter is not well informed. However when he says "garbage in" he is not referring to his query, but referring to source of AI data, the web and well there is a lot of garbage in it, no?
He’s clueless. ChatGPT gives a well rounded response to the origins of Covid, far more balanced than him: “COVID-19, caused by the coronavirus SARS-CoV-2, was first identified in Wuhan, China, in December 2019. The exact origins of the virus are still under investigation, but it is widely believed to have a zoonotic origin, meaning it was transmitted from animals to humans. The initial cluster of cases was linked to the Huanan Seafood Wholesale Market in Wuhan, which also sold live wild animals. This has led to the theory that the virus might have originated in bats and possibly passed through an intermediary animal host before infecting humans. However, conclusive evidence regarding the specific animal source and the pathway of transmission to humans has not yet been established, and research continues in this area.”
@@montymonto6430 but that is only if you are quering about things that are opinion. At the end he revealed the truth. He was asking about a COVID conspiracy theory, and then was mad when Perplexity shot down his conspiracy theory and gave the real facts. That was his entire reason for shooting down Perplexity. Because it didn't reinforce his false beliefs, he wanted Newsmax and didn't get it.
This was my take away from this "interview" as well. Both Wired and CNBC are scared of becoming obsolete due to AI. Why do I need a journalist to gather information and give me their biased take on it? I can use AI and ask as many follow up questions as I like to understand what is happening. I also noticed that she said that AI is crawling parts of their site that they don't want it to look at. So this is 100% a hit piece to stop AI from sharing their information. And they don't want AI crawling parts of the web that they decide we shouldn't know about. GTF out of here.
Disagree. This is almost hit-piece level unbalanced. Perplexity has been super useful for me. And it footnotes all sources, doens't just scrape reddit. It's a great product.
The issue is that he works for a media company and Perplexity has not been as kind as google when it comes to addressing the established patterns related to crawling copyrighted content. Media companies are not happy Perplexity will draw content from their platforms without paying or crediting them to the degree they would like.
WIRED THEMSELVES ARE GETTING SCREWED..........WHY WUD I GO ON THERE WEBSITES IF I USE PERPEXITY...........THEY DNT NEED ANYONE TO TELL THEM TO GO AFTER THEM............PERPEXITY IS COMING AFTER THERE JOBS N THEY ARE SCARED
If it’s “garbage in garbage out”, then if you go directly to the information source, it’s still garbage. Only now you’ve spent precious brain cycles having to look through ads to get to it. I don’t use Perplexity to find the right answer immediately, I use Perplexity to help me get closer to the right answer more efficiently without having to do manual searches and looking at irrelevant content.
Exactly! I think that both Wired and CNBC are scared of becoming obsolete due to AI. Why do I need a journalist to gather information and give me their biased take on it? I can use AI and ask as many follow up questions as I like to understand what is happening. I you're a smart person, then you can ask AI the right questions and sift through the garbage. If you ask a 4 or 5 word question and expect a perfect answer without any further research or questions, then you're just an idiot. This CNBC guy is an idiot.
Joe has given us a demo of how a discussion between Luddites would look like. There is an ongoing bet in the AI community on whether Joe will be replaced by a bot before his natural retirement.
Exactly, rather than embracing it and making it work for news journalist to amplify their reporting, they are totally threatened and lying about it - missing the point completely. He just said everything that Perplexity is not, and gave every user reason to continue utilizing it. AI fears and negative messaging will be their own demise.
It is odd that attack on Perplexity by CNBS. The anchor admits he doesn't know much, then jumps to the valuation that has nothing to do with the criticism on quality of the product. I use many of the LLMs, and I like Perplexity because it reads searched web pages and present a summary of the sources. If you don't trust or like a source, you can tell Perplexity to ignore it. The additional followup questions at the bottom is quite helpful too.
These tech experts can't even explain what is wrong with the product. He probably asked what color is the shirt that i'm wearing right now? and it gave him 10 different colors and told him where to buy it.
He said the truth at the end. He is a conspiracy theorist and gets angry when AI doesn't support his conspiracies. That is what the whole segment was about.
The fellow's description doesn't compute. Perplexity was recommended to me this morning by someone with a very wide grasp of the digital world and with my very brief experience today I'm finding Perplexity very different from the fellow's description in this YT video above. Unrecognizable, actually.
Exactly! It doesn't compute because it's a smear campaign - a hit piece. Everything these news media journalist lie about will only be verified and proven wrong, anyways, they're making themselves look like idiots!
If you ask it opinion questions you are going to get opinion answers. Joe definitely asked it " do people like me "... For technical or fact based stuff Perplexity is amazing..
Note to CNBC: Your viewers expect relevant , accurate information that keeps them up to date on new Technology not misinformation spewed by an out of date out of his element host who seems to fit the mold of a FOX NEWS cast member. I was expecting to hear him say "What in tarnation is this ?" at any moment .
the fact he also says that whats the point of AI if you have to check it... well Joe, if something used to take analysts an hour to do but takes the AI seconds and analyst now spends 5 minutes checking the work so that they can use 55 minutes to do more higher value stuff, that's pretty valuable to a lot of businesses
So: wired lady agree that perplexity is "garbage in garbage out" because perplexity scanned wired magazine? Does it mean that the wired thing is garbage?
Dear CNBC, this interview was a catastrophe. The host interrupted the guest on numerous times to make irrelevant points that do not reflect the Perplexity product. He should not speak on technology. Even at its base use of “summarizing ‘Google’ results” that is a huge time saver and the host is missing the value. Would you rather run 20 Google searches and read 50 articles or have an AI do it for you? As alluded to by the guest, Wired (Conde Nast) has a bone to pick against perplexity as they do not want their content being used to train an AI. With an interview like this, you should interview someone from Perplexity as well to defend against the accusations. It could be that the wired site was misconfigured or used as a honeypot to wield against Perplexity. Please work on the quality of your journalism.
Perplexity has a free version and it has a pro version. When you use the pro version, you get to choose what premium AI model will be used to answer your question. The fact that this report doesn’t say which version of even which AI model they were using makes it absolute junk journalism
I wouldn't trust CNBC at this point. If they couldn't get it right the first time, considering their legacy branded reputation, which proves why mainstream media is dying. I paused my workflow to watch the CNBC just because of Perplexity reporting and now I will never watch another reporting. Although I do like Andrew Sorkin. But the news reporter and Wired magazine, which seems to be racing to the bottom, have each lost my viewership. I wouldn't trust any mainstream media reporting about AI because they're full of negativity, lies and hatred. AI is supposed to enhance, support and increase productivity, our lives, etc. Perplexity is wonderful - CNBC birthed out of the internet, one would think it would transition with AI. Reporters with wisdom and are needed but ones with negative attitudes and anger like this guy are the reason why mainstream media journalist like this man are not worth listening to. He's 100% wrong and disgruntle.
Perplexity is amazing and almost always gives accurate information in a much better way than any other product. This analysis of Perplexity is garbage in garbage out.
Gotta love these hit pieces. He said all you need to know at 2:35 when he gushed over Googles' brilliance. Perplexity absolutley is a useful AI tool that is threatening Google because it answers questions with cited sources vs Google just providing search reaults.
Use a search engine if you want to find websites. Use Perplexity if you want to answer questions. Perplexity is incredibly good at sorting through information and bringing you all the relevant data. It's a time saver. You can try to sort through the information yourself, but it will take you ages in comparison. Perplexity is not an oracle or a super intelligence. It's limited by the information that's available on the internet just like you are. I don't think this news anchor understands that.
Joe Kernan has turned into a right-wing mouthpiece over the last 8 years, since Trump first ran. Thank god Bloomberg TV exists as an unbiased source of financial news. Whenever Kernan is on CNBC, I change the channel to Bloomberg. Kernan seems to forget that Democrats buy stocks too. I just wish that the producers at CNBC who keep him on the air would understand that too.
Perplexity is advertised as an AI search engine so of course all it does is search the internet. It does all the searching for you and then compiles the information it finds. I use it often for basic things I'm curious about or for gaming information. I agree it does make mistakes sometimes, but if you're not sure about something it says, or if you want to double check something it provides all of its sources.
I just tried Perplexity. I tried asking it very basic engineering questions and it consistently gave me wrong answers and presented those wrong answers very convincingly.
Meta is essentially doing the same thing . Ask it any question and it will scrape various websites and formulate a response quoting their source at the bottom. Yesterday Meta AI actually gave me two wrong answers.
Would Joe feels to understand is that AI is incredibly powerful but far from where it should be. It's fantastic at writing recipes, even writing code for my Reaper DAW. I threw in some HTML today and asked it to parse it out into tab delimited text and it just ate it up and gave out the data correctly. Writing code is quite difficult but it can still have good results after many many many failed attempts. It is able to work on relatively complex things. It also lies so you have to know exactly what kinds of questions you ask it because if the questions are consequential at all you may be in big trouble if it lies to you which it can do easily and has. It's easier just to call AI Google 2.0 which is what it really works as for me
@@KK-pm7ud hiQ v. LinkedIn, The Ninth Circuit court of appeals. It was brought before the Supreme Court but they ordered the appeals court to hear the case again.
Perplexity is an excellent tool for addressing the problem of hallucinations, contrary to what this hit piece suggests. How? It keeps the "human in the loop" by allowing users to instantly check the results because it provides citations. It’s a great service, and you can try it yourself. Just be sure to craft the prompt properly by directing it to look at authoritative sources. I use it as part of my prompting sessions to illustrate how CPAs and other business professionals can get preliminary research done fast.
I like how solid and accurately Katie responds, even while the CNBC guy is extremely biased and also seems to be the “I’m shouting the loudest, so I’m important” kinda guy. He’s not knowledgeable enough to lead.
Train of thought: - What if i searched covid? - What if it says covid jumped from one animal? - It could do that - Are any of them fair at this point Epic logical reasoning from this genius.
It is obvious that the moderator does not truly understand how Perplexity works. If he did, he would recognize that with Perplexity, you can search using many different models to get a more nuanced response. If he is just using the default model, he is really not using Perplexity in the best way possible. Of all the AI projects out there, you are more likely to get a balanced response from Perplexity because you have so many different models to pick from. Legal issues will be hashed out in court, but as for me, Perplexity gets a big thumbs up.
AI needs to be empircally validated. It hasn't achieved independent measurement validation, He is just responding the way any skeptic would reasonably respond.
Perplexity is great. It does great research and hardly ever gets things wrong. I suspect these people did not even subscribe and were using the free version.
I find Perplexity to be an excellent tool for research, for grasping deep profound philosophical concept, making connections. Just everything you want in an AI toll, the collection feature to keep your past thread organized for easy reference is a great addition!
Let me do a step by step rebuttal. 1. First, Google has been in the search business for far longer than perplexity but their AI-enabled previews are qualitatively worse than Perplexity. There is hallucination in Perplexity but it hallucinates to a lesser extent than Google's feature which I must say was so that they can compete with Perplexity. 2. Isn't what Joe is saying and the Wired piece actually opposite. Joe says Perplexity uses garbage data to train on stuff like reddit posts. If that is true, then it is not a threat to websites owned by Wired's owner Condé Nast as they produce legitimate websites I assume or at least what the Wired reporter would like to say. So, if Perplexity is using genuine websites, then Joe's argument is wrong. If Perplexity is only using "junk websites" as according to Joe, then the Wired lady does not have anything to worry about. They should first discuss what they are agreeing on before ranting. Is the topic: Perplexity is bad and hallucinates, or is it Perplexity broke the robots.txt norm of scraping/crawling. 3. For Joe, it seems like he is feeling personally threatened that Perplexity is good. If Perplexity is bad, then like all startups which give unsatisfactory results, it will perish irrespective of who funds them. 4. So Wired has a bone to pick with Perplexity so their piece is not an objective piece then. If this was coming out of another company (TechCrunch or Verge), one can listen to their argument. But if it comes from Ars Technica or Wired both owned by Condé Nast, then there is conflict of interest in how they will present the case for Perplexity. 5. Last but not the least, robots.txt is still seem as proper etiquette and in the US, the 9th circuit court had considered following robots.txt as terms of service/agreement of web use but had not sided with the aggrieved website in a past case. So, there is no clear legal precedence in US law which would guarantee that Perplexity is 100% legally wrong even if it is proven that they did do what Wired is alleging. Their investigation at this point has no meaning to any neutral party.
My experience has been that Perplexity always supplies source links and also almost always warns where warnings seem to be needed - either the commentator is out of his depth big time (most likely) or less likely it's hit job..
Fascinating... I am a boomer software engineer since the 80's, I successfully use Perplexity with prompts on technical topics (coding, science, health). Am knowledgeable in a wide variety of subjects so can discern when Perplexity is wrong, no TLDR here. The host might have a basic non-technical understanding but his presentation is a bit unhinged. Granted, if your prompt is about non-technical info eg politics, social issues it will reflect the wide chasm of data in these areas. Market valuations for "AI" reflect outsized investment in infrastructure not based on revenue generated by "AI" products and services.
Garbage in garbage out, very true for all AI models. But the input is what the user is giving in the search prompt of perplexity. So is he inputting garbage? Arvind and Co are undoubtedly building the best answer engine.
Watson seriously made a mistake that big? I was possibly looking into Watson for my company but after seeing how much they are charging I am shocked and even more shocked to learn they are charging that kind of money for errors! These models need to have a logic engine at the last mile to the user to verify and confirm the information. Hell it could write its own logic code. I think the problem is our RAG methods we are using, cutting things into chunks causes context loss.
I’m perplexed at their take on Perplexity Ai. So far, I find it to be the best of all of them out there. It hallucinates way less than any of the ChatGPT models.
This fails to note Perplexity cites all sources used to generate answers , it’s not garbage when the sources are cited from reliable sources including medical journals for example
Joe is great example of just how bad you can be at a job, yet still not quite bad enough to get fired. What a wasted opportunity to interview the guest so that the viewer is provided with further useful information (regarding use of secret IP addresses and scraping websites even when they say no, and exploring that issue further etc.)
I used free version of perplexity and it was amazing. I asked it many questions that only some one with deep knowledge of machine learning could answer. Perplexity not only answered it but gave me links to the material for further detail study. This looks like a jealous competitor doing a hit piece and this will paradoxically make perplexity a bigger hit.
I understand that perplexity might not be a breakthrough product but it’s a good product nevertheless. I have used it a few times but the value of the product lies in it being a type of copilot.
This CNBC guy is the absolute worst. Whenever I hear him talk about Tech he is the most consistently uninformed person and asks the dumbest questions.
Amen.
That's true about literally any topic. He's a well paid fool. People like him are exhibit A for our system not being a meritocracy at all.
This idiot wants AI to give unbiased answers
he got your click tho
He reflects his generation. It’s important he is there
Funny Joe used the phrase ' garbage in, garbage out' without realizing that is the truth, his query was garbage, so Perplexity gave him a garbage response.
Joe is like Biden in CNBC
Presenter is not well informed. However when he says "garbage in" he is not referring to his query, but referring to source of AI data, the web and well there is a lot of garbage in it, no?
Haha he said the AI had ego. This can only be projection because AI has no ego 😂
He’s clueless. ChatGPT gives a well rounded response to the origins of Covid, far more balanced than him:
“COVID-19, caused by the coronavirus SARS-CoV-2, was first identified in Wuhan, China, in December 2019. The exact origins of the virus are still under investigation, but it is widely believed to have a zoonotic origin, meaning it was transmitted from animals to humans. The initial cluster of cases was linked to the Huanan Seafood Wholesale Market in Wuhan, which also sold live wild animals. This has led to the theory that the virus might have originated in bats and possibly passed through an intermediary animal host before infecting humans. However, conclusive evidence regarding the specific animal source and the pathway of transmission to humans has not yet been established, and research continues in this area.”
@@montymonto6430 but that is only if you are quering about things that are opinion. At the end he revealed the truth. He was asking about a COVID conspiracy theory, and then was mad when Perplexity shot down his conspiracy theory and gave the real facts. That was his entire reason for shooting down Perplexity. Because it didn't reinforce his false beliefs, he wanted Newsmax and didn't get it.
Perplexity is a great product.
Sounds like the presenter didn’t like the answer he got when he searched for himself.
I just used it and it’s pretty cool
totally agree, the host is a total idiot... I use perplexity and it works great! The guy don't know what he's talking about
Garbage in, garbage out 🤣
It's a hit piece by typical media lies and censorship.
Perplexity is amazing. This is a hit piece funded by a competitor.
This was my take away from this "interview" as well. Both Wired and CNBC are scared of becoming obsolete due to AI. Why do I need a journalist to gather information and give me their biased take on it? I can use AI and ask as many follow up questions as I like to understand what is happening.
I also noticed that she said that AI is crawling parts of their site that they don't want it to look at. So this is 100% a hit piece to stop AI from sharing their information. And they don't want AI crawling parts of the web that they decide we shouldn't know about. GTF out of here.
I'm not sure how it's different than Bing copilot
Lol. They sound butthurt. It’s a made up issue.
Disagree. This is almost hit-piece level unbalanced. Perplexity has been super useful for me. And it footnotes all sources, doens't just scrape reddit. It's a great product.
Agreed. He's being paid under the table by ChatGPT.
Ok bot
Aravind is that you
Def a hit piece. That much is obvious. Also obvious is that Grandpa Joe forgot his meds this morning.
Google is also good
Why is the host so angry? All of these AI tools explicitly say that it can be inaccurate.
They all hold a lot of Google stock and are scared 🤣
The issue is that he works for a media company and Perplexity has not been as kind as google when it comes to addressing the established patterns related to crawling copyrighted content. Media companies are not happy Perplexity will draw content from their platforms without paying or crediting them to the degree they would like.
@@ashwinkumar675 CNBC employees can only hold broad ETFs
@@MrMountain707 but not their family members and friends
@@floroos I did not know that lol
why is this guy on TV?
My thoughts exactly.
Oldies should stick to blueberry and Siemens phones.. pure trash talk
@@Did_you_know448 I think he is still using typewriters and fax machines. He is the type who can not set the timer on a VCR.
Garbage in, as always. 🤭
I stopped after 2:31 into the video. To quote him "I was looking for sentience" and found he had none.
Perplexity is my favorite. It gives me the best results. Someone funded a hit piece
WIRED THEMSELVES ARE GETTING SCREWED..........WHY WUD I GO ON THERE WEBSITES IF I USE PERPEXITY...........THEY DNT NEED ANYONE TO TELL THEM TO GO AFTER THEM............PERPEXITY IS COMING AFTER THERE JOBS N THEY ARE SCARED
I agree it does feel like a hit job. I very rarely have any complaint with my perplexity search results.
Probably because you're a sheep
The anchor seems to not know what he is talking about
He and Biden are about the same age and have similar mental functions.
It seems he should be better reviewing adult diapers. 😆😅😂
If it’s “garbage in garbage out”, then if you go directly to the information source, it’s still garbage. Only now you’ve spent precious brain cycles having to look through ads to get to it.
I don’t use Perplexity to find the right answer immediately, I use Perplexity to help me get closer to the right answer more efficiently without having to do manual searches and looking at irrelevant content.
Exactly! I think that both Wired and CNBC are scared of becoming obsolete due to AI. Why do I need a journalist to gather information and give me their biased take on it? I can use AI and ask as many follow up questions as I like to understand what is happening.
I you're a smart person, then you can ask AI the right questions and sift through the garbage. If you ask a 4 or 5 word question and expect a perfect answer without any further research or questions, then you're just an idiot. This CNBC guy is an idiot.
This feels like a hit job. Perplexity is great! It is a thoughtfully designed knowledge discovery engine. It's for the curious!
Why is it confusing? This seems like a coordinated effort to bring down perplexity
They must figure google will bless their SEO rankings for kissing the ring
Joe has given us a demo of how a discussion between Luddites would look like. There is an ongoing bet in the AI community on whether Joe will be replaced by a bot before his natural retirement.
I hope thats soon, he's such a moron.
Might not work. AI chat bots aren't that angry (or clueless).
@@hugooc 😂
Joe is a terrible interviewer. He has no idea what a.i does, he himself spew misinformation all the time😂😂😂😂
I had to check the comments because I use perplexity all the time and I love it. So to call it trash makes no sense whatsoever 😂.
Exactly, rather than embracing it and making it work for news journalist to amplify their reporting, they are totally threatened and lying about it - missing the point completely. He just said everything that Perplexity is not, and gave every user reason to continue utilizing it. AI fears and negative messaging will be their own demise.
It is odd that attack on Perplexity by CNBS. The anchor admits he doesn't know much, then jumps to the valuation that has nothing to do with the criticism on quality of the product. I use many of the LLMs, and I like Perplexity because it reads searched web pages and present a summary of the sources. If you don't trust or like a source, you can tell Perplexity to ignore it. The additional followup questions at the bottom is quite helpful too.
These tech experts can't even explain what is wrong with the product. He probably asked what color is the shirt that i'm wearing right now? and it gave him 10 different colors and told him where to buy it.
He said the truth at the end. He is a conspiracy theorist and gets angry when AI doesn't support his conspiracies. That is what the whole segment was about.
The fellow's description doesn't compute. Perplexity was recommended to me this morning by someone with a very wide grasp of the digital world and with my very brief experience today I'm finding Perplexity very different from the fellow's description in this YT video above. Unrecognizable, actually.
Exactly! It doesn't compute because it's a smear campaign - a hit piece. Everything these news media journalist lie about will only be verified and proven wrong, anyways, they're making themselves look like idiots!
What an odd interview.
The presenter just convinced me to purchase a perplexity subscription where I wasn’t going to before. 😂
Wow, thats harsh even for wall street.
Was the old man was whining?
He wants us off his lawn.
@@hugooc
😄
😂😂😂
If this is what corporate media has to say about it, it must be awesome.
If you ask it opinion questions you are going to get opinion answers. Joe definitely asked it " do people like me "... For technical or fact based stuff Perplexity is amazing..
Note to CNBC: Your viewers expect relevant , accurate information that keeps them up to date on new Technology not misinformation spewed by an out of date out of his element host who seems to fit the mold of a FOX NEWS cast member. I was expecting to hear him say "What in tarnation is this ?" at any moment .
What ever thr topic is on the planet. Joe manages to make it political. 😂
While reminding us of the clown we know him to be 🤡
the fact he also says that whats the point of AI if you have to check it... well Joe, if something used to take analysts an hour to do but takes the AI seconds and analyst now spends 5 minutes checking the work so that they can use 55 minutes to do more higher value stuff, that's pretty valuable to a lot of businesses
Haha.. sounded like Joe asked AI about himself and reddit answers didn’t disappoint.
Truth
On point!
I've never heard of Wired but Perplexity is great.
🤣 True. Grandpa loves Wired. Now get off my lawn!
Why is it okay for Google to use their journalism in its search results, but not okay for Perplexity?
Makes me wonder about their motives. What I got from this is stay away from CNBC and WIRED. Perplexity is great.
They just gave us the nails to hammer them in the coffin. Hahah!
So: wired lady agree that perplexity is "garbage in garbage out" because perplexity scanned wired magazine? Does it mean that the wired thing is garbage?
Dear CNBC, this interview was a catastrophe. The host interrupted the guest on numerous times to make irrelevant points that do not reflect the Perplexity product. He should not speak on technology.
Even at its base use of “summarizing ‘Google’ results” that is a huge time saver and the host is missing the value. Would you rather run 20 Google searches and read 50 articles or have an AI do it for you?
As alluded to by the guest, Wired (Conde Nast) has a bone to pick against perplexity as they do not want their content being used to train an AI. With an interview like this, you should interview someone from Perplexity as well to defend against the accusations. It could be that the wired site was misconfigured or used as a honeypot to wield against Perplexity. Please work on the quality of your journalism.
Perplexity has a free version and it has a pro version. When you use the pro version, you get to choose what premium AI model will be used to answer your question. The fact that this report doesn’t say which version of even which AI model they were using makes it absolute junk journalism
This interviewer barely let the guest talk. It was mostly his own angry ramblings.
He wants people off his lawn.
This CNBC host needs to be replaced with an intelligent AI host that actually understands AI. "I expect sentience" - What?!
I wouldn't trust CNBC at this point. If they couldn't get it right the first time, considering their legacy branded reputation, which proves why mainstream media is dying. I paused my workflow to watch the CNBC just because of Perplexity reporting and now I will never watch another reporting. Although I do like Andrew Sorkin. But the news reporter and Wired magazine, which seems to be racing to the bottom, have each lost my viewership. I wouldn't trust any mainstream media reporting about AI because they're full of negativity, lies and hatred. AI is supposed to enhance, support and increase productivity, our lives, etc. Perplexity is wonderful - CNBC birthed out of the internet, one would think it would transition with AI. Reporters with wisdom and are needed but ones with negative attitudes and anger like this guy are the reason why mainstream media journalist like this man are not worth listening to. He's 100% wrong and disgruntle.
Did he just say, "I'm looking for sentience?"
My gosh.
He wouldn't let her speak , just forced his opinion down the audience's throat.
He seems mad that he didn’t get the answers he was looking for.
She has a valid point.
Yep
To everyone watching this video I would absolutely love your opinion on if this is a Hit Job on Perplexity or not.
Hit job.
hit job
Perplexity is great. I use it every day
Please pin this page when 'AI' crash.
This guy is so frank because he has not invested any money like the rest.
Perplexity is amazing and almost always gives accurate information in a much better way than any other product. This analysis of Perplexity is garbage in garbage out.
Gotta love these hit pieces. He said all you need to know at 2:35 when he gushed over Googles' brilliance. Perplexity absolutley is a useful AI tool that is threatening Google because it answers questions with cited sources vs Google just providing search reaults.
Use a search engine if you want to find websites. Use Perplexity if you want to answer questions. Perplexity is incredibly good at sorting through information and bringing you all the relevant data. It's a time saver. You can try to sort through the information yourself, but it will take you ages in comparison.
Perplexity is not an oracle or a super intelligence. It's limited by the information that's available on the internet just like you are. I don't think this news anchor understands that.
Joe Kernan has turned into a right-wing mouthpiece over the last 8 years, since Trump first ran. Thank god Bloomberg TV exists as an unbiased source of financial news. Whenever Kernan is on CNBC, I change the channel to Bloomberg. Kernan seems to forget that Democrats buy stocks too. I just wish that the producers at CNBC who keep him on the air would understand that too.
yeap, can't stand him. He hardly does homework and his questions are poor.
And Bloomberg poached Joumanna from CNBC.
I am a perplexity user it is amazing not sure what they are talking about. Seems like they have an agenda …
100% - another journalist making up a narrative and creating 'misinformation'... sad.
Perplexity is advertised as an AI search engine so of course all it does is search the internet. It does all the searching for you and then compiles the information it finds. I use it often for basic things I'm curious about or for gaming information.
I agree it does make mistakes sometimes, but if you're not sure about something it says, or if you want to double check something it provides all of its sources.
Perplexity Bots and persons of interest have taken over the comment section.
It's becoming more obvious that bots have taken over the Internet every day
there is no muslim called ana anywhere.
I just tried Perplexity. I tried asking it very basic engineering questions and it consistently gave me wrong answers and presented those wrong answers very convincingly.
This can't be true!
How is it garbage or BS? They give you answers and cite the source. Kernan is confused again.
Meta is essentially doing the same thing . Ask it any question and it will scrape various websites and formulate a response quoting their source at the bottom. Yesterday Meta AI actually gave me two wrong answers.
What is this man doing on this topic?
Lol this guy has a personal vendetta against perplexity or something? He was very harsh in his word.
really cringe how he keeps cutting her off
Funny conversation of a non-tech guy trying to have a conversation with a highly skilled tech lady. Her big words just flew right pass his head. LOL
the host is definitely talking like a boomer
Would Joe feels to understand is that AI is incredibly powerful but far from where it should be. It's fantastic at writing recipes, even writing code for my Reaper DAW. I threw in some HTML today and asked it to parse it out into tab delimited text and it just ate it up and gave out the data correctly. Writing code is quite difficult but it can still have good results after many many many failed attempts. It is able to work on relatively complex things. It also lies so you have to know exactly what kinds of questions you ask it because if the questions are consequential at all you may be in big trouble if it lies to you which it can do easily and has. It's easier just to call AI Google 2.0 which is what it really works as for me
Oh man, there can't be a TV host who is so far behind the knowledge curve as this one.
Never saw perplexity site reddit.
*cite.
The judicial system in the USA has ruled that web scraping is legal. They don’t have to comply with a request to not scrape a website.
What case law are you citing?
@@KK-pm7ud hiQ v. LinkedIn, The Ninth Circuit court of appeals. It was brought before the Supreme Court but they ordered the appeals court to hear the case again.
@@KK-pm7ud Bright Data vs. Meta though the case does not end the many legal cases that will revolve around this issue.
Perplexity is an excellent tool for addressing the problem of hallucinations, contrary to what this hit piece suggests. How? It keeps the "human in the loop" by allowing users to instantly check the results because it provides citations. It’s a great service, and you can try it yourself. Just be sure to craft the prompt properly by directing it to look at authoritative sources. I use it as part of my prompting sessions to illustrate how CPAs and other business professionals can get preliminary research done fast.
I’ll accept Perplexity isn’t completely transparent about how they train models but it’s good to use.
“Garbage” is clearly an exaggeration.
This BS Machine is worth billions today you fools 😂
CNBC…seriously? The anchor’s prep for this story was right up there with, “the dog ate my homework.”
I like how solid and accurately Katie responds, even while the CNBC guy is extremely biased and also seems to be the “I’m shouting the loudest, so I’m important” kinda guy. He’s not knowledgeable enough to lead.
Sounds like Perplexity is taking market share from their own investments.
Train of thought:
- What if i searched covid?
- What if it says covid jumped from one animal?
- It could do that
- Are any of them fair at this point
Epic logical reasoning from this genius.
Perplexity should probably get rid of the free tier or make it more limited so they can use more powerful models. The subscription version works well
It is obvious that the moderator does not truly understand how Perplexity works. If he did, he would recognize that with Perplexity, you can search using many different models to get a more nuanced response. If he is just using the default model, he is really not using Perplexity in the best way possible. Of all the AI projects out there, you are more likely to get a balanced response from Perplexity because you have so many different models to pick from. Legal issues will be hashed out in court, but as for me, Perplexity gets a big thumbs up.
AI needs to be empircally validated. It hasn't achieved independent measurement validation, He is just responding the way any skeptic would reasonably respond.
Perplexity is great. It does great research and hardly ever gets things wrong. I suspect these people did not even subscribe and were using the free version.
I find Perplexity to be an excellent tool for research, for grasping deep profound philosophical concept, making connections. Just everything you want in an AI toll, the collection feature to keep your past thread organized for easy reference is a great addition!
Can't believe CNBC allowed this. This guy should never be allowed to report on tech news! He has absolutely no idea what he is saying!
Did Google pay CNBC to downplay Perplexity? I think so.
It was the first thing I thought. And also to Wired, they wrote the original article.
There could not be a more effective promotional message to get me to try Perplexity AI.
He was looking for sentience and perplexed he didn't find any! CNBC and their reporters needs some, no doubt.
Perplexity is awesome. It's not confusing and it's not difficult to figure out how to work. These people are idiots.
I have used perplexity and now is my default to find factual information. I have stopped using Google that is plagued by sponsored links.
Let me do a step by step rebuttal.
1. First, Google has been in the search business for far longer than perplexity but their AI-enabled previews are qualitatively worse than Perplexity. There is hallucination in Perplexity but it hallucinates to a lesser extent than Google's feature which I must say was so that they can compete with Perplexity.
2. Isn't what Joe is saying and the Wired piece actually opposite. Joe says Perplexity uses garbage data to train on stuff like reddit posts. If that is true, then it is not a threat to websites owned by Wired's owner Condé Nast as they produce legitimate websites I assume or at least what the Wired reporter would like to say. So, if Perplexity is using genuine websites, then Joe's argument is wrong. If Perplexity is only using "junk websites" as according to Joe, then the Wired lady does not have anything to worry about. They should first discuss what they are agreeing on before ranting. Is the topic: Perplexity is bad and hallucinates, or is it Perplexity broke the robots.txt norm of scraping/crawling.
3. For Joe, it seems like he is feeling personally threatened that Perplexity is good. If Perplexity is bad, then like all startups which give unsatisfactory results, it will perish irrespective of who funds them.
4. So Wired has a bone to pick with Perplexity so their piece is not an objective piece then. If this was coming out of another company (TechCrunch or Verge), one can listen to their argument. But if it comes from Ars Technica or Wired both owned by Condé Nast, then there is conflict of interest in how they will present the case for Perplexity.
5. Last but not the least, robots.txt is still seem as proper etiquette and in the US, the 9th circuit court had considered following robots.txt as terms of service/agreement of web use but had not sided with the aggrieved website in a past case. So, there is no clear legal precedence in US law which would guarantee that Perplexity is 100% legally wrong even if it is proven that they did do what Wired is alleging. Their investigation at this point has no meaning to any neutral party.
Conde Nast does have a reasonable point. Joe is just off the rails.
Facts? Trust me, bro..
4:23 "I've looked at things. That I disagreed with consensus on"
My experience has been that Perplexity always supplies source links and also almost always warns where warnings seem to be needed - either the commentator is out of his depth big time (most likely) or less likely it's hit job..
It's a hit job - 100% Another journalist that's a murder.
The Wired’s grief is that Perplexity scraped the data and not about quality and trying to go after them with the article.
Don’t they have standardized tests to compare different models in unbiased way? How does Perplexity compare?
couple of months ago, CNBC (this guy included) was praising how much its journalists depend on Perplexity (which I never used)
Joe is a sillyhead, he will read whatever shows up on the prompt and give it an angry spin
Fascinating... I am a boomer software engineer since the 80's, I successfully use Perplexity with prompts on technical topics (coding, science, health). Am knowledgeable in a wide variety of subjects so can discern when Perplexity is wrong, no TLDR here. The host might have a basic non-technical understanding but his presentation is a bit unhinged. Granted, if your prompt is about non-technical info eg politics, social issues it will reflect the wide chasm of data in these areas. Market valuations for "AI" reflect outsized investment in infrastructure not based on revenue generated by "AI" products and services.
"LLMs violate the copyright of all our journalism."
"Everything they produce is illogical garbage."
Oh? Do tell
Garbage in garbage out, very true for all AI models. But the input is what the user is giving in the search prompt of perplexity. So is he inputting garbage? Arvind and Co are undoubtedly building the best answer engine.
Are they planning to sue? They should… great story Katie - agree with everything you’re reporting
CNBC: The Original BS Machine Reviews Its Competition' 😂
Watson seriously made a mistake that big? I was possibly looking into Watson for my company but after seeing how much they are charging I am shocked and even more shocked to learn they are charging that kind of money for errors! These models need to have a logic engine at the last mile to the user to verify and confirm the information. Hell it could write its own logic code. I think the problem is our RAG methods we are using, cutting things into chunks causes context loss.
I’m perplexed at their take on Perplexity Ai. So far, I find it to be the best of all of them out there. It hallucinates way less than any of the ChatGPT models.
This fails to note Perplexity cites all sources used to generate answers , it’s not garbage when the sources are cited from reliable sources including medical journals for example
Joe is great example of just how bad you can be at a job, yet still not quite bad enough to get fired.
What a wasted opportunity to interview the guest so that the viewer is provided with further useful information (regarding use of secret IP addresses and scraping websites even when they say no, and exploring that issue further etc.)
I used free version of perplexity and it was amazing. I asked it many questions that only some one with deep knowledge of machine learning could answer. Perplexity not only answered it but gave me links to the material for further detail study. This looks like a jealous competitor doing a hit piece and this will paradoxically make perplexity a bigger hit.
I understand that perplexity might not be a breakthrough product but it’s a good product nevertheless. I have used it a few times but the value of the product lies in it being a type of copilot.
Google hasn’t been working lately. To many sponsored ads and garbage cluttering up real information. It’s been harder to find things