Jeff Bezos Bets on a Google Challenger Using AI To Try To Upend Internet Search

Perplexity, a startup going after Google's dominant position in web search, has won backing from Jeff Bezos and venture capitalists betting that artificial intelligence will upend the way people find information online. From a report: Started less than two years ago, Perplexity has fewer than 40 employees and is based out of a San Francisco co-working space. The company's product, which it calls an answer engine, is used by about 10 million people monthly. Those ingredients were enough to persuade Institutional Venture Partners, Bezos and other tech executives to invest $74 million in the company, the largest sum raised by an internet search startup in recent years. The investment valued Perplexity at $520 million, including the new money, said Chief Executive Officer Aravind Srinivas.

Amazon.com, the e-commerce company chaired by Bezos, has committed to investing billions in Anthropic, the AI startup behind the chatbot Claude. The path to competing with Google is littered with carcasses, and Microsoft itself has struggled for years to dislodge the company's roughly 90% market share in online search. Others, including tech giants, are incorporating AI into their search engines. Perplexity's founders said their advantage is using advances in AI to provide direct answers, instead of website links, in response to search queries, without some of the limitations felt by larger companies.

Note to self

[Rosco P. Coltrane]

Never use Perplexity.

Re:

[Keick]

You know what, I like new shiny things, so why not. How bad could it be? I mean ChatGPT impresses me on the daily.

So your just another ChatGPT clone?

No, I am not just another ChatGPT clone. I am an AI language model developed by Perplexity AI. While I share the same underlying technology as ChatGPT, I am a unique instance of this language model, trained to provide accurate and helpful information on a wide range of topics. If you have any questions or need assistance, feel free to ask, and I will be happy to help.

How are you better than ChatGPT?

I don't have access to the content of the links you provided. However, I am an AI language model developed by Perplexity AI, and I am trained on a diverse range of data sources to provide accurate and helpful information on a wide variety of topics. My training data includes up-to-date information and knowledge from reputable sources. If you have any specific questions or need assistance, feel free to ask, and I will do my best to provide you with a reliable and informative response.

I'm pretty sure Eliza was smarter than Perplexity is.

Re:

[Rosco P. Coltrane]

How bad could it be?

- AI here, AI there. I'm sick to the backteeth of the AI circle-jerk.
- It doesn't give links, meaning I've forced to learn things through its interpretation of whatever I asked for.
- It helps Bezos become richer.

Tthree excellent reasons to stay the hell away from it.

Re:

[Shag]

I asked it "who is" questions about my wife and myself. It gave correct information about my wife, who's become much more visible online in the last several years.

I, on the other hand, have become less visible, and although my name isn't particularly common, there are a few other people around the world with the same name. So it decided that I was a freakishly multi-talented person who worked as a Wall Street trader, actor, astronomer, rugby player and writer. While amusing, at least 40% of this was wrong

Re:

[Shag]

...and I see I need to update my signature here, since Perplexity did not say anything (this time) about my doing brain surgery while racing motorbikes.

Re:

[Ferocitus]

You know what, I like new shiny things, so why not. How bad could it be? I mean ChatGPT impresses me on the daily.

So your just another ChatGPT clone?

I'm pretty sure Eliza was smarter than Perplexity is.

But much nastier.
She would have replied: It's you're, you dumb cunt.

wsj.com

[Anonymous Coward]

Paywalled article? I would expect better from /. but I guess the good 'ol days are over...

Direct answers?

[JaredOfEuropa]

their advantage is using advances in AI to provide direct answers, instead of website links

I think the lack of website links (or references) is one of the disadvantages of AI.

Re:Direct answers?

[Calydor]

Very much so. The 'answer engine' was supposed to be Wikipedia, which is very adamant about citing your sources. Hell, that's required for any kind of schoolwork. Don't cite Wikipedia, cite whatever Wikipedia cites.

But I guess this is just another step towards making it so only the elite actually know things; the rest of the rabble just need to have faith in what they are told. Just like the good ol' dark ages.

Re:

[EvilSS]

They actually do give weblinks to references. I think what they mean is that don't JUST give you web links like a Google search will (although Google will sometimes also give a direct answer, so.... yea).

Honestly, I'd be more interested in a search engine like Google, but using AI to give better results than Google currently does. From what I can tell, Perplexity is just another ChatGPT clone. As a matter of fact it seems to use ChatGPT and "Copilot" among some other models. It's just a different interfac

Re:

[jenningsthecat]

Honestly, I'd be more interested in a search engine like Google...

I'd be interested in a search engine like the Google of fifteen years ago. Heck, I'd even settle for the Google of five years ago - it wasn't nearly as good as the earlier days, but it was at least a bit better than today's pile of dreck.

Re:

[EvilSS]

Wouldn't we all. Unfortunately "SEO"s ruined that version of Google. Then Google injecting more and more ads and sponsored results finished it off.

Re: Direct answers?

[LindleyF]

Have you tried the Google Generative Search Experience? It's opt-in but public.

Re:

[EvilSS]

I have not. Looking at it though, it looks like it only works in Chrome on desktop right now. Bummer, I use Firefox as my main browser.

Kagi

[Xenolith0]

I've found Kagi[1] search results to be substantially better for the type of queries I do. So much better in fact, that I actually pay for it.

I choose to believe that, supposedly, they're honestly not keeping search-history logs[2]

[1] https://kagi.com/ [kagi.com]

[2] https://kagi.com/privacy [kagi.com]

Remember A9.com?

[xack]

Amazon's last attempt, they even had street view.

Does Google even Do Search anymore?

[databasecowgirl]

Given how the goog has devolved into an ad service, I would imagine it wouldn't be two difficult to do better.

Re:

[jythie]

This is the problem with anyone trying to 'dethrone' Google. Google's current search behavior is not a technical issue, it is a market one. Any company that enters the space will be under the same economic pressures with the same rewards as Google, and thus will likely fall into the same patterns.

Given the total mess that is Amazon's search

[david.emery]

It's rich to think of Jeff Bezos trying to beat Google from a usability/quality perspective. BUT this makes great sense if Bezos' metric is not 'quality of search results' but 'ad revenue and click-through fees'.

Re:

[ArchieBunker]

Amazon and google are similar in terms accuracy. If you search for anything that isn't popular you will get one or two hits and then a million adjacent hits. eBay was the lone holdout on accuracy but now I see the same bullshit. You look for a part number and get many similar part numbers sprinkled in. Similar part numbers don't cut it.

Re:

[david.emery]

Actually, I usually 'search' Amazon through DuckDuckGo when I'm looking for something specific. The chance top item actually matching my full search criteria is much better in DuckDuckGo than Amazon. Amazon gives preference to their 'sponsors' which often do not meet the search criteria, which DuckDuckGo does not (instead providing the fully conforming item in the first link.) I rarely use Google these days, so I can't comment on how well/poorly Google search works in the same situation. But I suspect t

Re:

[dargaud]

About ebay, yes. I've had a longstanding running search for a rare bike part reference. It would find one every 2 years. And for the past month or 2 I get several emails per day with things that are totally irrelevant.

Help us, Beff Jezos, you're our only hope

[elcor]

He will do it, fix his lame duck Antropic with something not as terrible as Claude and compete with our overlord of ad service. His window of opportunity is very large thanks to bard being a pile of steaming pooh. Will he fall prey to the temptation of heavy lobotomy to serve his political agenda?

One way to compete is to

[Tablizer]

...allow garbage and embarrassing shit. Big brands don't want to alienate sponsors and investor egos. But a small startup can move fast and break things by accepting a percentage of riff-raff. Whether Bezos will accept such a strategy is unknown. Musk would in a heartbeat. But, you'll get weird sponsors.

Things Bezos invested in that failed

[toxonix]

This is an AI generated result from Google:
In 2000: Bezos invested $60 million in Kozmo.com, a shampoo delivery service.
In 2012: Bezos lost $175 million when he purchased LivingSocial, a competitor to Groupon.
In 2014: Bezos was involved in the design process for the Amazon Fire smartphone, which was a major flop.
In 2019: Amazon wrote off $170 million for its failed Fire smartphone.
In 2020: Amazon launched the free multiplayer game "Crucible" after five years of development, but it performed poorly and never

It wouldn't be hard to do better than Google

[rlwinm]

These days Google Search is useless. hexdump -C /dev/urandom provides more relevant results.

Perplexity is fast but very shallow

[m00sh]

I know they are doing their funding (huge ad in their search bar for paid customers) and this is probably a hype piece sponsored by them. Anyways, I've been using them for a bit.

Perplexity is improved web-search but compared to ChatGPT and Claude it is quite lacking in depth.

It can't do as good an answer as ChatGPT and it can't do as good a retrieval as Claude. Yet, they claim to use ChatGPT and Claude in the backend.

I believe to save API costs and token costs, they are using tricks to summarize and make th