Thank you @TypeHrishi ,
That seems to be a sensible use of AI.
I shall try it.
Thank you @Rosika .
Well yes, I can, because I have worked with model fitting.
It would be fun to try and make a small version of one of these systems, but it would take some years of effort.
I appreciate the efforts prople make to try and help me keep up with these modern computing developments. It would be difficult on my own.
Perplexity uses DeepSeek
This AI world is evolving rapidly
Maybe I’ll try it locally in a VM detached from any network.
Just to see, how clever it is forced to be locally and offline? 
I went searching for Firefox extensions to enable use of Perplexity.
It found 37 addons
For all of them it said
This add-on is not actively monitored for security by Mozilla. Make sure you trust it before installing.
That makes me nervous.
Hi Neville, 
To be able to use it less nervously you´d just have to run peplexity in a firejailed browser, either adding --private or --private=PATH-TO-WORK-DIRECTORY as parameter.
No extensions needed this way. 
E.g.:
firejail --read-only=(pwd) --private=(pwd) firefox -no-remote "https://www.perplexity.ai/"
Cheers from Rosika 
Hi Rosika,
Thank you, that is much simpler and safer.
I thought of trying the perplexity app, but my tablet has too old a version of Android.
Regards
Neville
Hi Neville,
I see. Well, before trying to install an app (which wouldn´t work on your tablet anyway, as you said) I think it´s always a good idea to get acquanted with perplexity the safe way (i.e. firejailing the browser).
BTW:
Mind the -no-remote option from my above example. It´s useful.
By default, a single Firefox process instance handles multiple browser windows.
If you already have Firefox running, you would need to use-no-remotecommand line option, otherwise you end up with a new tab or a new window attached to the existing Firefox process.
( from here, bold by me)
Cheers from Rosika
I just emailed that link (to myself)
Thank you
@nevj and all,
for anyone interested:
As far as perplexity is concerned I may add that you can create (but you don´t have to) an account for using it.
So I got a bit curious and asked perplexity itself about it.
Q:
I have a question regarding perplexity itself.
I use it in a web-browser. There is the option to create an account (“sign up” button). I can use perplexity without an account. So what´s the benefit of creating an account?A:
Creating a Perplexity account offers several benefits over using it without an account:
Save search history: With an account, you can save and organize your search history for future reference39.
Customize interface: An account allows you to personalize your Perplexity experience, including setting preferences for language and interface appearance (light or dark mode)5.
Access to Collections: You can organize your queries into collections for better management48.
File uploads: Account holders can attach files like documents to provide more context for their questions48.
Copilot features: Signing up unlocks additional AI-assisted features to enhance your search experience6.
Five Pro searches per day: Free account users get access to 5 daily Pro searches, which use more advanced AI models for complex queries48.
Follow-up questions: You can ask up to 5 follow-up questions every 4 hours to explore topics in more depth4.
While you can use Perplexity for quick searches without an account, creating one enhances your overall experience and provides access to these additional features at no cost36.
Many greetings from Rosika 
Strange, I was just now asking myself the same question.
Thank you.
I am not into A/i and don’t use it. I am going to assume that like all A/i platforms, it has been invented as a stealth privacy invader, data collector and indoctrinator. It is glorified spyware and I am not ready to trust the Chinese but it’s nice to see them come up with something that may be better than the likes of chatgpt and perplexity.
The Australian Government has banned DeepSeek from all government offices
Nice! I think it should be banned from all countries…
Quod licet Iovi, non licet bovi.
English translation: What is permissible for Jupiter is not permissible for an ox.
First time I have come across this latin phrase, they did not teach that at my school
My probably wrong opinion:
Doing that without further thinking and analizing would just kill competition between AI developers/providers.
So I think either ban them all, or none of them.
I could say the statement:
ChatGPT is safe, it does not collect any data, it respects privacy.
Would you beleive it is 100% true? If not, what’s the difference to the chinese AI model?
Maybe nothing in this regard, but I emphasize: I don’t know!
What is interesting, that (according to my readings on the topic) Deepseek can work using way much less resources than ChatGPT.
That should be examined, analized, evaluated instead of simply banning it.
I still did not have the time to make that chinese model work on a VM, but I’m going to do it maybe next week. I’ll tell you how clever it is totally offline, so apparantly no data transmitted to China 
Not convinced a ban is the correct action either, everyone should have the choice, like social media I choose not to use it, but I know my family do, just do not join in.
Government offices are a different thing to individual use.
All for government bans on some sites, tools, etc. Internally same for bigger organisations.
The energy usage of these AI systems is huge