I'm a physician and use Perplexity AI Pro for medical research summaries, including comparing and contrasting medical protocols/guidelines. Citations are always accurate, and I can easily guide the answers I want with good prompts. It's my go-to AI search model.
Silicon Valley CEOs like Satya Nadella (Microsoft), Sunder Pichai (Google), Aravind Srinivasan (Perplexity AI) etc uses subsidized education (subsizidized by Indian tax payers) from the world famous Indian Institutes of Technology to get their Bachelors degress before migrating to the US. Once they become stalwarts of the tech world, hope they use their clout to help in development of Indian tech and India as a whole rather than being selfish ungrateful dummies who cares just for their own fortune and fame.
@@sweetchinmusic801 1. Inventions and innovations benefit the mankind 2. China leverages ethnic reasoning like above to recruit spies abroad. 3. Human talent is not zero sum - I am sure there are hundreds among 40 - 70 million recent graduates in India currently unemployed who are smarter than Arvind and some of them will go on to do something that will improve human condition and some of them will be wealthy like Ambani without having to come to the USA.
I am a paid Pro user, and I like it. However, it adds a bit of friction if you just want a quick link. I would want the search results to be surfaced even faster before the LLM text even starts. Also, it fails on follow-ups sometimes, as it somewhat loses context, meaning it takes the follow-up as a completely new question. I would like it to be more consistent in keeping it as a thread. Overall, I really like it.
OnE of the things I do in my prompts is to break things up in numerical steps and refer back to a specific step to maintain context. I imagine you must be doing something similar but would love to hear how you approach it.
Yes, would love a mini model to kick out a quick link immediately up front. Is painful waiting sometimes. Is also odd getting used to perplexity search failing next steps at times - “if it sits too long at this step, it’s not going to answer”. Overall very happy as well with pro, amazing platform to test models as the come out, save history in collections etc
I’m digging it but I’ve received multiple hallucinations where the links were dead and the info made up. I had another instance where it sent me to links that I specifically requested it not share again because it was not relevant to the question but no matter how many times it acknowledged its mistake it would still send the same bad link. Most of the time it works great. Just would love to see some improvement in these areas.
A thoroughly enjoyable hour and a half, surely one of the best and most enlightening interviews on AI available today. Aravind is a Gem to the world. Thank you.
Great response to your question about Perplexity being called a ChatGPT wrapper. Also liked how he said if Perplexity’s customers are paying $20/month and don’t care about what model is being used, why should he?
I like that he talks about hiring a group of people to work on hard problems. And that he doesn’t overpay his people so that they have so much money that they retire and then knowledge base goes out the window.
He's talking about ranking sources; they have to be using a crawl database. Maybe using Google APIs too, but perplexity's links don't match Google's in my queries.
Silicon Valley CEOs like Satya Nadella (Microsoft), Sunder Pichai (Google), Aravind Srinivasan (Perplexity AI) etc uses subsidized education (subsizidized by Indian tax payers) from the world famous Indian Institutes of Technology to get their Bachelors degress before migrating to the US. Once they become stalwarts of the tech world, hope they use their clout to help in development of Indian tech and India as a whole rather than being selfish ungrateful dummies who cares just for their own fortune and fame.
Google should keep Gemini a separate thing within Google that has a free version of Gemini that has advertisements like UA-cam commercials or print ads on the page of output. Or people could pay for a paid version with no ads.
you guys were great. Avarind i appreciate your openness and wisdom. Do more speaking to clarify the value of Perplexity. It is important that we understand that we are moving into a currency of questions. We will have to educate users to learn how to ask the right questions and overcome our passive acquisition of knowledge of memory driven success.
Warm solid cuddle of intelligence on a creative philosophical cool ocean air ! So sweet and healthy ! So good to feel the process of humble brilliance grounded in practical action rooted in a relationship between rhythm melody and harmonic diverse progressive grace! I feel very lucky to of heard this pod cast ! Perplexity blew my tiny mind for the love of research and flow state met its match ❤
I actually had an argument with perplexity the other day about my website not being a real sight... I had to actually tell it to Google the site. It's been getting worse for the past week or so.
Perplexity is on the rise. Google needs to be distrupted from the ad model. I fear tho all AI companies like this will go ad based eventually as it's way more profitable
It can't answer questions about python dependency errors. 😢 (But no one can.) Also, I've been trying not to use it too much because I didn't want to overload their servers and make them drop the free plan. And here he's trying to grow usage. Derp.
I like the phrase "no gradient" at 1:14:57. I like the idea of move fast and break things but test well and make rolling it into production perhaps reversible, or at least consider the possibility of doing so. Don't avoid risk, but weigh the risk so you don't blow off production or things of extreme value.
Tbh I prefer Krawl AI over Perplexity. It's more work-focused and I can get lot more things done using it because of the "tools" concept they introduced. Plus they are adding new tools to it almost every other week
I'm a physician and use Perplexity AI Pro for medical research summaries, including comparing and contrasting medical protocols/guidelines. Citations are always accurate, and I can easily guide the answers I want with good prompts. It's my go-to AI search model.
Aravind presents his thesis without all of the hype. This is a CEO I can get behind his vision and trust that he will see it through.
Silicon Valley CEOs like Satya Nadella (Microsoft), Sunder Pichai (Google), Aravind Srinivasan (Perplexity AI) etc uses subsidized education (subsizidized by Indian tax payers) from the world famous Indian Institutes of Technology to get their Bachelors degress before migrating to the US. Once they become stalwarts of the tech world, hope they use their clout to help in development of Indian tech and India as a whole rather than being selfish ungrateful dummies who cares just for their own fortune and fame.
@@sweetchinmusic801 Ouch!! You bring up a great point.
@@sweetchinmusic801 1. Inventions and innovations benefit the mankind 2. China leverages ethnic reasoning like above to recruit spies abroad. 3. Human talent is not zero sum - I am sure there are hundreds among 40 - 70 million recent graduates in India currently unemployed who are smarter than Arvind and some of them will go on to do something that will improve human condition and some of them will be wealthy like Ambani without having to come to the USA.
This was a terrific interview. Tons of alpha. Much appreciated. I really like Aravind, and truly wish all the best for Perplexity.
Such a thoughtful and insightful episode - especially liked hearing about the future of search and Aravind's predicted breakthroughs in AI
I am a paid Pro user, and I like it. However, it adds a bit of friction if you just want a quick link. I would want the search results to be surfaced even faster before the LLM text even starts. Also, it fails on follow-ups sometimes, as it somewhat loses context, meaning it takes the follow-up as a completely new question. I would like it to be more consistent in keeping it as a thread. Overall, I really like it.
Precisely!
OnE of the things I do in my prompts is to break things up in numerical steps and refer back to a specific step to maintain context. I imagine you must be doing something similar but would love to hear how you approach it.
I like it too. No need to pay for pro
Yes, would love a mini model to kick out a quick link immediately up front. Is painful waiting sometimes. Is also odd getting used to perplexity search failing next steps at times - “if it sits too long at this step, it’s not going to answer”. Overall very happy as well with pro, amazing platform to test models as the come out, save history in collections etc
Flow state takes time to master for both master and student , process is the key 🦋
I continued to be amazed by perplexity ... the more I use it it truly seems like it's only limited by your own creativity. i am truly amazed
This guy is sharp. It’s clear he will be an incredibly important player in this industry. Nice job on the interview!
One of the few profound visionaries I've heard in 2024. He will win something and he just won 1 new customer who will give perplexity 3 chances.
I’m digging it but I’ve received multiple hallucinations where the links were dead and the info made up. I had another instance where it sent me to links that I specifically requested it not share again because it was not relevant to the question but no matter how many times it acknowledged its mistake it would still send the same bad link. Most of the time it works great. Just would love to see some improvement in these areas.
same issues w me
A thoroughly enjoyable hour and a half, surely one of the best and most enlightening interviews on AI available today. Aravind is a Gem to the world. Thank you.
Great response to your question about Perplexity being called a ChatGPT wrapper. Also liked how he said if Perplexity’s customers are paying $20/month and don’t care about what model is being used, why should he?
I LOVE Perplexity! Thank you for creating a fantastic service!
He is so clear in his vision google with AD's has spoiled everything. One of the best AI products i have used. Keep going
Agree
Great questions.Great answers. Thank you for this!
I love Perplexity far more than Google because the related questions are always great avenues for research.
I like that he talks about hiring a group of people to work on hard problems. And that he doesn’t overpay his people so that they have so much money that they retire and then knowledge base goes out the window.
Perplexity is a wrapper over google search
More over GPT-4, but sure…and Google is a wrapper on the internet😅
He's talking about ranking sources; they have to be using a crawl database. Maybe using Google APIs too, but perplexity's links don't match Google's in my queries.
Silicon Valley CEOs like Satya Nadella (Microsoft), Sunder Pichai (Google), Aravind Srinivasan (Perplexity AI) etc uses subsidized education (subsizidized by Indian tax payers) from the world famous Indian Institutes of Technology to get their Bachelors degress before migrating to the US. Once they become stalwarts of the tech world, hope they use their clout to help in development of Indian tech and India as a whole rather than being selfish ungrateful dummies who cares just for their own fortune and fame.
@@jonmichaelgalindo sometimes good and sometimes bad
@@sweetchinmusic801 To build a large tech and global product you have to be in the valley
The product is damn good
Google should keep Gemini a separate thing within Google that has a free version of Gemini that has advertisements like UA-cam commercials or print ads on the page of output. Or people could pay for a paid version with no ads.
you guys were great. Avarind i appreciate your openness and wisdom. Do more speaking to clarify the value of Perplexity. It is important that we understand that we are moving into a currency of questions. We will have to educate users to learn how to ask the right questions and overcome our passive acquisition of knowledge of memory driven success.
Warm solid cuddle of intelligence on a creative philosophical cool ocean air ! So sweet and healthy ! So good to feel the process of humble brilliance grounded in practical action rooted in a relationship between rhythm melody and harmonic diverse progressive grace! I feel very lucky to of heard this pod cast ! Perplexity blew my tiny mind for the love of research and flow state met its match ❤
Thank you for sharing this.
May your day BE filled with more thankfulness, encouragement and forward thinking today....
A very open minded individual..very cool guy and a great conversation
I enjoyed this episode. Aravind is good.
wow that was so good
❤ 17:08 thanks for the question!
The giant now copies perplexity. Now what ? The unfair advantage of the Goliath, is to be able handle legal hurdles and absorb many errors.
How can i invest in Perplexity? It is phenominal!
How to undo delete thread?
I actually had an argument with perplexity the other day about my website not being a real sight... I had to actually tell it to Google the site. It's been getting worse for the past week or so.
Perplexity is on the rise. Google needs to be distrupted from the ad model. I fear tho all AI companies like this will go ad based eventually as it's way more profitable
It can't answer questions about python dependency errors. 😢 (But no one can.) Also, I've been trying not to use it too much because I didn't want to overload their servers and make them drop the free plan. And here he's trying to grow usage. Derp.
Bro gained 50 pounds between the thumbnail and the video
This guy is amazing!
No hate for him, but I've a strong feeling this company is just soaring through this AI phase and soon it will be out of the market.
Perplexity AI has quietly reduced Claude Opus usage to 50 per 24h
Is this like social media without dislike button , keep that to ur self in ur utopian world
As a PRO user of Perplexity, I find the “Related questions” that show up on the bottom distracting and unnecessary. Blank slate = Clear thinking
That isn’t likely. I find perplexity gives me old info constantly and I pay for pro.
All he ever do is talk about Google
They’re a very relevant topic in what he’s doing…..
That might be the most aggressive male leg cross over I've ever seen.
Why this dude lookin like the king of MIDDLE EARTH!?! 💍
Apple will buy them off, 100 billion
Zzz
I like the phrase "no gradient" at 1:14:57. I like the idea of move fast and break things but test well and make rolling it into production perhaps reversible, or at least consider the possibility of doing so. Don't avoid risk, but weigh the risk so you don't blow off production or things of extreme value.
Tbh I prefer Krawl AI over Perplexity. It's more work-focused and I can get lot more things done using it because of the "tools" concept they introduced. Plus they are adding new tools to it almost every other week
Just tried it that. Perplexity is objectively much better.