Posted on 11/21/2023 9:10:02 AM PST by unlearner
There are some solid uses of the technology for sure—but I have zero confidence in homo sapiens ability to keep AI benign—even for the medium term.
Perplexity:
The comparison between ChatGPT and Perplexity is not straightforward, as both tools offer unique experiences and cater to different needs. Here are some key points from the search results:
Perplexity is described as a real-time fact-checked information machine that excels in providing real-time data and trustworthy information for web search-related tasks
1
4
.
ChatGPT, especially the GPT-3.5 and GPT-4 versions, is highlighted for its stable and quick performance, making it suitable for various tasks such as content creation, software development, and streamlining work processes
1
5
.
Perplexity is praised for its user interface, which is considered slicker compared to ChatGPT
1
.
ChatGPT Plus is noted for providing natural responses and detailed answers, while Perplexity is more suitable for providing brief answers and real-time information with sources
2
.
Perplexity Pro is mentioned to be better at information lookup by default, and it can also enable GPT-4 to be always on, offering a writing focus
3
.
Perplexity AI is described as less of a chatbot and more of a search bot, resembling the Google search engine in its layout and providing real-time information
4
.
Both ChatGPT and Perplexity are reported to have their strengths in providing fast responses, with not much difference between them in this aspect
4
.
In conclusion, the choice between ChatGPT and Perplexity depends on the specific needs of the user. ChatGPT is praised for its stable and quick performance, while Perplexity is highlighted for providing real-time information and a slick user interface. Both tools have their respective strengths and can be valuable depending on the user’s requirement
“First to go, I hope, will be the ‘dumb phone menu systems’. They’ll actually be able to handle natural speech, without ‘key words’, and intelligently be able to get you in the right direction. That alone is a massive number of opportunities.”
I think this use case should be obvious, but the movers and shakers have not yet seized upon it.
I’m particularly interested in implementing something using Twilio (programmable voice) because I think we’re nearing a point where human callers will not be able to tell they are talking to an AI rather than a human agent, except maybe because the AI is more effective.
Twilio has a big market cap right now but is underutilized, selling its services more like commodities. I believe that there is massive potential to leverage this as a value-added service to individuals and small businesses.
As AI what it thinks.
I don’t believe that AI thinks yet. It appears to mimic human language and thought. But it will be interesting to see if science or philosophy can come up with ways to test whether other entities can think or feel.
I use Perplexity sometimes because it is connected to the Internet to access more current information than ChatGPT 3.5. And this feature comes without a subscription fee. Perhaps I’ll subscribe to a paid version when I can find a way to make money from it. Or maybe there will be free offerings for students and researchers.
90B valuation at peak, wonder what it goes to now. Also Binance CEO stepping down. Crazy week.
Yes
“Most naive post I have read today.”
I use AI for generating structural geometry for fabrication. There are designs that are so light weight and efficient that it will radically change every day products and machines. BUT... As with all new tech. It will end up being used to promote pornography and the marxist democrat agenda.
https://www.autodesk.com/solutions/generative-design
:) Very much so...
I firmly believe that...if “IT” CAN be done, where people have the MOTIVATION TO DO “IT”, and are given the RESOURCES TO DO “IT”...then “IT”, whatever it is (e.g. hydrogen bomb), will be done.
...very much applies here. Good luck putting up firewalls and safety guards. This technology will go at lightspeed, if we don’t then our enemies (i.e. China) WILL do it....just like in WWII with Germany and the atom bomb.
Your knowledge is almost certainly as shallow as your answer.
Lots of people post online thinking they’ve made a point or proved something when it is a delusion in their own minds.
You’ve added nothing definitive to the conversation. It is merely an assertion from someone who neither provided evidence to support the shared opinion nor any basis to believe his or her credibility.
Here’s a good idea: Have a point. It makes it so much more interesting for the listener.
https://youtu.be/FG81ptfPz7I?t=101
“I use AI for generating structural geometry for fabrication.”
You’re wasting your time using things like facts and real-world examples. Some people prefer to operate in a deluded realm of self-aggrandization.
I haven’t looked into that one (Twilio)...will do.
I used the ChatGPT python libraries and integrated a demo using voice to text (speech rec) and text to speech to implement an “Iron Man” style Jarvis application.
It’s *really* surprising the level of conversation you can have with it. It’s very natural and you can ask details on very complex topics, all while it keeps track of the conversation context history. I’ve leveraged its “call back APIs” (e.g. defined by my application) to do more than chat, upon prompting, it can respond with a callback API (instead of a chat message) and my local machine will execute the functions exposed - enabling any local capabilities available.
I’m initially targeting an in-vehicle assistant. All the vehicle functions (media, navigation, comms, HVAC, etc.) controllable via callback functions. It’s quite amazing to have a chat with it and then say something casual like, “...oh, btw, the car is really hot, make it cold, and play some Metallica, and navigate home”. Resulting in each of my callbacks being invoked in its response (temp set to 69, Metallica playing, navi routed to home). Then the context is remembered, so if I say “it’s still too hot, make it colder”, it calls my HAVC callback function again but to (e.g.) 67 degrees - it knows that is cooler than the last call.
It’s just a demo - but I want it in my car now :) I’ve since learned (of course) some other OEM’s are doing similar. It’s just one application...but this stuff is just beginning too.
I think Tom Anderson, the co-founder of Myspace, may be available to pinch-hit. Groupon’s founder and CEO, Andrew Mason, has some availability as well.
Meg Whitman might be available. I don't think Quibi is very busy these days.
:-)
https://en.wikipedia.org/wiki/Colossus:_The_Forbin_Project
Even before “War Games” and “The Terminator” and Skynet.
I never saw that. Looks interesting.
It wasn’t a ‘big’ movie at the time, and never got much TV replay either, but probably more provocative now than it ever was before.
The issues with this are (1) Microsoft owns a huge chunk of OpenAI and has exclusive agreements with them and (2) Poaching the CEO might be possible but there was no way to know all the staff would bail too and (3) but even if they did guess that there was no way to know he (and they) would come to Microsoft, they could have gone over to Musk’s new AI venture, for example and that would devastate the value of Microsoft’s investment in Open AI and (4) Microsoft, through it’s ownership share, already has access to all the tech in OpenAI anyway.
In spite of all this it seemed to work out incridibly well for Microsoft.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.