Perplexity AI

[lnofeisone]

10 hours ago, mattee2264 said:

Isn't it a little worrying for Mag7 investors that a start-up founded only a few years ago can release LLMs that compare favourably to the ones put out by OpenAI/Microsoft and Google? Reinforces the point that in AI no one really has a moat at this point. Mag7's ability to stockpile chips gives them an edge and they can outspend everyone on R&D. But in the internet age it was Google who ended up with the dominant search engine not AOL/Netscape/Microsoft. 

Any company with a good idea/product will have little problem attracting funding and users aren't locked in to a specific LLM at this point so will switch if something better comes along. 

 

And so far it is LLMs that are generating all the buzz and drawing in all the users and seeming to have the most practical use as a lot of people are using LLMs to write emails, assignments, research papers, marketing copy etc. 

From what I see, there is now a trend to tap into AWS/GCP to accelerate the journey to LLM and other analytics. Lots of orgs have proprietary data but have no clue how to get it to LLM. 

 

As far as the moat for Google/Msft/Aws, I'd argue the moat now got bigger. There are only few start ups that are really thriving here. Anthropic is one. Perplexity is another. Many fold shop before they get to market. The cost to train these models is north of $1M per run. There are a lot of great ideas but few companies with funding to try them so AWS/GCP/Azure have plenty to pick from and fund if they so desire.

[bargainman]

On 3/4/2024 at 11:10 AM, rogermunibond said:

Isn't prompt engineering like Google-fu?

It's more than that.  I mean there's user level stuff but then there are things like https://www.langchain.com/langchain that create components that get chained together for different flows.  It's usually not just a single statement.  When you message Chat GPT for example it doesn't just take what you wrote and give it to the LLM.  It goes through layers and tries to prevent attacks and uses what are called "system" level prompts to wrap your user level prompt.  Plus there's no memory so it actually sends your entire conversation through the LLM every time.  So there's a lot too it.

[bargainman]

On 3/6/2024 at 1:54 AM, mattee2264 said:

Isn't it a little worrying for Mag7 investors that a start-up founded only a few years ago can release LLMs that compare favourably to the ones put out by OpenAI/Microsoft and Google? Reinforces the point that in AI no one really has a moat at this point. Mag7's ability to stockpile chips gives them an edge and they can outspend everyone on R&D. But in the internet age it was Google who ended up with the dominant search engine not AOL/Netscape/Microsoft. 

Any company with a good idea/product will have little problem attracting funding and users aren't locked in to a specific LLM at this point so will switch if something better comes along. 

 

And so far it is LLMs that are generating all the buzz and drawing in all the users and seeming to have the most practical use as a lot of people are using LLMs to write emails, assignments, research papers, marketing copy etc. 

 

Made me LOL.  Sorry but this is the nature of tech, always has been.  Gates always said he was more worried about a couple of folks in a garage cooking something up that would turn the entire industry upside down.  These days a LOT of stuff is out in the open.  Take a look at huggingface which hosts community and models of many many open source LLMs.  This is academia driven sort of.  they always were driven to publish or perish and release their stuff to get reviewed.

 

These days though the limits are starting to show with training.  It's crazy how expensive it is to train an LLM, how many GPUs and how much energy it takes.  it's really only very large companies with huge budgets, GPUs and data that can afford it.  Plus the data annotation process is difficult and error prone, nevermind ethically challenging.

 

Still, once models are out, a lot of them are open source, so you can finetune for less money, and just try prompt engineering to get what you want/need.

 

It's still day one.. early early days which is shocking but true.

[backtothebeach]

Anyone feel like Perplexity sucks recently? I find myself searching for some information, not getting a satisfactory answer, and finally going to Grok, which spits out a super comprehensive answer 5 times faster.

[UK]

I do not know much about Perplexity, but am a bit shocked, in a good way, how good and useful Gemini now is. Maybe it fits into what I am looking for, but wow, no delays, very coprehensive and useful answers, basically solving problems for which I just recently was literally paying money (and still was not able to get satisfactory answers:))

[Luke]

1 hour ago, UK said:

I do not know much about Perplexity, but am a bit shocked, in a good way, how good and useful Gemini now is. Maybe it fits into what I am looking for, but wow, no delays, very coprehensive and useful answers, basically solving problems for which I just recently was literally paying money (and still was not able to get satisfactory answers:))

Gemini is great...also switched from ChatGPT to them. 

[Spekulatius]

I have not used Perplexity for many month. I use a Gemini as a default (it’s good and it’s quick) and ChatGPT (good but sluggish). I have very little need for anything beyond the two.

 

I used Perplexity in 2023 and 2024 when it seemed to be ahead of the pack  but I don’t think I have used it this year at all. I think it has become the Altavista of Chatbots.

[Dalal.Holdings]

https://www.perplexity.ai/finance

 

FYI there is a Finance Perplexity that does earnings transcripts, financials, etc

 

I use Perplexity as a Google/Wikipedia search hybrid. I use Gemini for deep research only really

 

...The thing about all of these is---is there really much differentiation for day to day use ?

 

Edited December 10, 2025 by Dalal.Holdings

[WayWardCloud]

The differentiation will come from integrating your personal AI chatbot (and soon agent) with everything else you already do. I use Gemini Pro and it's connecting to my gmail, google calendar, google photos and google docs and let's me retrieve and send information back and forth between all those separate apps. It's still early and not super useful all the time but I can see the unified vision taking place. Something that Perplexity can never do. Just like The Gemini team describes itself as the engine room at Alphabet, the Gemini model is planning to be the engine room or control center from which you pilot the ecosystem of all of your Google services. This is when decades of offering things for free pays off, in particular the trove of unique data assembled inside Gmail and Google docs. They know everything about you and now they're going to make it centralized, useful and accessible to you. The vision has been layed out a long time ago : "organize the world's information and make it universally accessible and useful". I would say the twist is that it's actually going to become much more "personal" than "universal".

 

Microsoft has the opportunity to do the equivalent but for enterprise. So far they're bungling the execution though.

Edited December 10, 2025 by WayWardCloud

[lnofeisone]

You can use Gemini in Perplexity (I do this). I see results much better than straight Gemini without any modifications. I can course correct Gemini to give similar results. 

[treasurehunt]

12 hours ago, Luke said:

Gemini is great...also switched from ChatGPT to them. 

 

Me too. I'll give ChatGPT another shot when 5.2 becomes available though.

 

Jensen Huang said in September after Nvidia announced a $100B investment in OpenAI that "OpenAI is going to be the next multi-trillion dollar hyperscale company" and "this is some of the smartest investments that we can possibly imagine". I wonder if he still feels that way.

 

 

[Longnose]

4 minutes ago, treasurehunt said:

 

Me too. I'll give ChatGPT another shot when 5.2 becomes available though.

 

Jensen Huang said in September after Nvidia announced a $100B investment in OpenAI that "OpenAI is going to be the next multi-trillion dollar hyperscale company" and "this is some of the smartest investments that we can possibly imagine". I wonder if he still feels that way.

 

 

Thats because chatgpt is his best customer. 

[treasurehunt]

4 minutes ago, Longnose said:

Thats because chatgpt is his best customer. 

 

Good point. Maybe there is some wishful thinking going on here. But it's still early days, and I'm very curious to see how it all plays out.

[WayWardCloud]

1 hour ago, lnofeisone said:

You can use Gemini in Perplexity (I do this). I see results much better than straight Gemini without any modifications. I can course correct Gemini to give similar results. 

You can absolutely use the Gemini model on its own inside Perplexity but since its siloed off it won't provide you with the Google ecosystem integration you get from using Gemini directly and giving it access to your private data from all the other Google apps you've been using for years. So you will get great general responses but not personalized ones tailored to your current projects, your personality or your history. Which means for multi steps projects you will have to keep copy pasting things from one app to another in multiple tabs who don't speak to each other and also that your AI assistant will never learn how you like things done and how to serve you better specifically. I think for now this cross-functionality is barely starting but as agentic behaviors and personal date integration increase, the models who have permission to navigate into the already existing plumbing of your digital everyday life (maps, emails, calendar, docs) will prove much more useful and seamless.

Edited December 10, 2025 by WayWardCloud

[lnofeisone]

4 hours ago, WayWardCloud said:

You can absolutely use the Gemini model on its own inside Perplexity but since its siloed off it won't provide you with the Google ecosystem integration you get from using Gemini directly and giving it access to your private data from all the other Google apps you've been using for years. So you will get great general responses but not personalized ones tailored to your current projects, your personality or your history. Which means for multi steps projects you will have to keep copy pasting things from one app to another in multiple tabs who don't speak to each other and also that your AI assistant will never learn how you like things done and how to serve you better specifically. I think for now this cross-functionality is barely starting but as agentic behaviors and personal date integration increase, the models who have permission to navigate into the already existing plumbing of your digital everyday life (maps, emails, calendar, docs) will prove much more useful and seamless.

You can connect perplexity to your google docs (or any google service like calendar) through what they call File Connections or use n8n or Zapier. You can extend perplexity to also connect to other products concurrently - e.g., connect to msft OneDrive. Gemini can't do that out of the box and requires an external thing like n8n. 

[WayWardCloud]

Thanks for explaining!

[bargainman]

On 12/10/2025 at 8:31 PM, lnofeisone said:

You can connect perplexity to your google docs (or any google service like calendar) through what they call File Connections or use n8n or Zapier. You can extend perplexity to also connect to other products concurrently - e.g., connect to msft OneDrive. Gemini can't do that out of the box and requires an external thing like n8n. 

 

While this is true, there is a lot to be said for the default behavior.  gemini pro is connected by default good parts of the google ecosystem, copilot to the Microsoft ecosystem etc.