Questions to help flesh out for all to benefit who use Firefox on Windows.
1. What are the top dozen (or so) publicly available web-based GPT AI's?
2. Are any aimed at privacy (e.g., no mothership-login requirement)?
Offhand, some free AI chat bots I'm aware of (which isn't all of them)...
<https://askgpt.app/>
<https://chatbotapp.ai/>
<https://chatgpt.com/>
<https://claude.ai/>
<https://copilot.microsoft.com/>
<https://deepai.org/>
<https://gemini.google.com/>
<https://getmerlin.in/>
<https://huggingface.co/chat/>
<https://jasper.ai/>
<https://monica.im/>
<https://my.replika.com/signup>
<https://perplexity.ai/>
etc.
What free web-based GPT chat AI's are we missing in that list?
Which (if any) have no mothership-login tracking requirements?
Questions to help flesh out for all to benefit who use Firefox on Windows.[snip]
1. What are the top dozen (or so) publicly available web-based GPT AI's?
2. Are any aimed at privacy (e.g., no mothership-login requirement)?
On Sun, 22 Dec 2024 09:42:43 -0000 (UTC), Andrew wrote:
Questions to help flesh out for all to benefit who use Firefox on Windows. >>
1. What are the top dozen (or so) publicly available web-based GPT AI's?
2. Are any aimed at privacy (e.g., no mothership-login requirement)?
Services are not even softwares.
What free web-based GPT chat AI's are we missing in that list?
Which (if any) have no mothership-login tracking requirements?
Waste no time on these toys...
On Wed, 25 Dec 2024 11:49:51 +0800, Mr. Man-wai Chang wrote:
What free web-based GPT chat AI's are we missing in that list?
Which (if any) have no mothership-login tracking requirements?
Waste no time on these toys...
There's an advantage to free unlimited talk models which will tell you what your prescriptions do, what the weather is right now, what the world news stories are, what's a balanced news station, how solar panels work, etc.
No typing.
On 12/25/2024 1:34 PM, Isaac Montara wrote:
On Wed, 25 Dec 2024 11:49:51 +0800, Mr. Man-wai Chang wrote:
What free web-based GPT chat AI's are we missing in that list?
Which (if any) have no mothership-login tracking requirements?
Waste no time on these toys...
There's an advantage to free unlimited talk models which will tell you what >> your prescriptions do, what the weather is right now, what the world news
stories are, what's a balanced news station, how solar panels work, etc.
No typing.
Is it really so much trouble to type a few words? So-called
AI not only skips the typing. It also skips the editing and
curating. Do you really want to return to being bottle-fed,
just so that you won't have to chew your food? I'm guessing
there's a drug for that. Maybe an SSRI. You can ask your
chat pal. :)
On Thu, 12/26/2024 7:00 AM, Newyana2 wrote:
On 12/25/2024 1:34 PM, Isaac Montara wrote:
On Wed, 25 Dec 2024 11:49:51 +0800, Mr. Man-wai Chang wrote:
What free web-based GPT chat AI's are we missing in that list?
Which (if any) have no mothership-login tracking requirements?
Waste no time on these toys...
There's an advantage to free unlimited talk models which will tell you what >>> your prescriptions do, what the weather is right now, what the world news >>> stories are, what's a balanced news station, how solar panels work, etc. >>>
No typing.
Is it really so much trouble to type a few words? So-called
AI not only skips the typing. It also skips the editing and
curating. Do you really want to return to being bottle-fed,
just so that you won't have to chew your food? I'm guessing
there's a drug for that. Maybe an SSRI. You can ask your
chat pal. :)
If you are asking the AI...
"Is it snowing where you are?"
then it doesn't really matter what the answer is.
I would recommend building a local query engine in your own home.
An RTX 4090, has lots of grunt (1000 TOPS?), but some in the thread here, think
that some other kind of NPU might be better for queries. The RTX 4090 needs >time to load a model, and other NPU types might take less time to do that.
https://www.reddit.com/r/ArtificialInteligence/comments/1d6g6fg/how_can_i_use_my_geforce_rtx_4090_for_ai/
The absolute best of that thread, is the attempt at an answer at
the very bottom of the thread. "fintech07" copies a trash AI answer
into the thread, helping no one, and the response from another poster ?
"no response is better than ai one"
The participants know exactly the worth of the "gold" in this gold rush.
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)
The way that tokens scale, you can't load the entire tarball
into an AI. Token limits currently are 128KB or 256KB. The
Firefox tarball is huge. AI can't swallow the whole source,
change the architecture, and spit out new source for the
whole thing, as a monolithic "one-gulp" operation. Re-writing
the code a page at a time, helps no one.
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)
I don't think it's reasonable to suggest that the current state of the art for
AI is all we'll ever see. On the contrary, it's probably a near certainty that
what got us to this point will sooner or later be overshadowed by future advancements. In other words, "ever" is a very long time.
I don't think it's reasonable to suggest that the current state of the art for
AI is all we'll ever see. On the contrary, it's probably a near certainty that
what got us to this point will sooner or later be overshadowed by future advancements. In other words, "ever" is a very long time.
******* end: CoPilot AI Test Run *******
Now, isn't that beautiful ?
Paul wrote:
[snip]
******* end: CoPilot AI Test Run *******
Now, isn't that beautiful ?
Don't show that the the Post Office - they will want it to write the replacement for Horizon !!!!
On 12/26/2024 11:06 PM, Char Jackson wrote:
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)I don't think it's reasonable to suggest that the current state of
the art for
AI is all we'll ever see. On the contrary, it's probably a near certainty that
what got us to this point will sooner or later be overshadowed by future
advancements. In other words, "ever" is a very long time.
I'm still surprised that after so many tech advances, over a couple
of hundred years, we're still so naive as to get excited about how
the latest development will change our lives. Of course, there have
been changes. We put clothes into a machine with a button now, to
wash them, and no longer need to hike down to the river with a scrub
board. But those are not changes in quality of life. They're only
details. The tech is only tools.
Meanwhile, people keep getting excited about how the next
breakthrough will transform our lives. I see that as a symptom of
depression. The people who want a push-button world or who thrill
that AI can look something up online without having to think for
themselves are people who find it a hassle to simply relate to their
own experience. "Let's get a device to save us from having to
actually live our lives. That will be so great!"
I remember as a child having a futurism mindset. I used to watch
some science show every Sunday evening, hosted by Union Carbide.
The show told me about what's new in science. The commercials
told me about breakthroughs at Union Carbide. Utopia always seemed
just one discovery away. Now the kids are living through devices,
numbed by disconnection, "traumatized" when they have to deal with
humans, yet the mythology of tech utopia is still in effect. Futurism --
"the miracle of tech" -- is still the naive answer to dealing with life's problems. The glee of "What'll they think of next?!" is still a kind of existential landmark of desperate hope against the abyss of death.
"Life sucks and then you die, but have you seen this new aerial
toothbrush that you don't even have to hold? It's incredible."
On 27/12/2024 2:53 am, Paul wrote:
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)
The way that tokens scale, you can't load the entire tarball
into an AI. Token limits currently are 128KB or 256KB....
A.I. does NOT create, but to read, "analyze", summarize and conclude! :)
Graham J <nobody@nowhere.co.uk> wrote:
Paul wrote:
[snip]
******* end: CoPilot AI Test Run *******
Now, isn't that beautiful ?
Don't show that the the Post Office - they will want it to write the
replacement for Horizon !!!!
You'll probably have to explain that one! :-)
On Fri, 12/27/2024 11:44 AM, Frank Slootweg wrote:
Graham J <nobody@nowhere.co.uk> wrote:
Paul wrote:
[snip]
******* end: CoPilot AI Test Run *******
Now, isn't that beautiful ?
Don't show that the the Post Office - they will want it to write the
replacement for Horizon !!!!
You'll probably have to explain that one! :-)
https://en.wikipedia.org/wiki/Horizon_%28disambiguation%29
Computing
Horizon, the Fujitsu IT and accounting system at the heart of the British Post Office scandal
https://en.wikipedia.org/wiki/British_Post_Office_scandal
"The British Post Office scandal, also called the Horizon IT scandal,
involved the Post Office pursuing thousands of innocent subpostmasters for
apparent financial shortfalls caused by faults in Horizon,
an accounting software system developed by Fujitsu.
"
On 26/12/2024 2:34 am, Isaac Montara wrote:
There's an advantage to free unlimited talk models which will tell you what >> your prescriptions do, what the weather is right now, what the world news
stories are, what's a balanced news station, how solar panels work, etc.
Do you really need to ask A.I. when Google Search can do the job fine?
Mr. Man-wai Chang <toylet.toylet@gmail.com> wrote:
On 27/12/2024 11:00 pm, Paul wrote:
On Fri, 12/27/2024 9:20 AM, Mr. Man-wai Chang wrote:
On 27/12/2024 2:53 am, Paul wrote:They can write code.
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)
The way that tokens scale, you can't load the entire tarball
into an AI. Token limits currently are 128KB or 256KB....
A.I. does NOT create, but to read, "analyze", summarize and conclude! :) >>
Ask the AI for a Hello World program, and that one runs when compiled.
In the following example, it even listens to some of my input.
Yes, they write AFTER reading and deep-LEARNing existing codes written
by humans (possibly dead)! A smart copy-cat! ;)
How is this different to the majority of human creativity? People don't
write code through pure intuition, but often just get "inspired" by stackoverflow posts...
Chris <ithinkiam@gmail.com> wrote:
Mr. Man-wai Chang <toylet.toylet@gmail.com> wrote:
On 27/12/2024 11:00 pm, Paul wrote:
On Fri, 12/27/2024 9:20 AM, Mr. Man-wai Chang wrote:
On 27/12/2024 2:53 am, Paul wrote:They can write code.
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-) >>>>>> The way that tokens scale, you can't load the entire tarball
into an AI. Token limits currently are 128KB or 256KB....
A.I. does NOT create, but to read, "analyze", summarize and conclude! :) >>>>
Ask the AI for a Hello World program, and that one runs when compiled. >>>>
In the following example, it even listens to some of my input.
Yes, they write AFTER reading and deep-LEARNing existing codes written
by humans (possibly dead)! A smart copy-cat! ;)
How is this different to the majority of human creativity? People don't
write code through pure intuition, but often just get "inspired" by
stackoverflow posts...
You're probably correct about the *majority* of human creativity, i.e.
in this case code. But *real* AI should be able to create custom code
from a design spec, etc., not from only adapting other existing code.
That's the kind of code I mainly wrote and I like to think that I'm
somewhat intelligent! :-)
AI has the potential to suggest which searches will be most likely to
lead to the information that you are looking for.
I'm not sure it would be able to tell you about balanced news stations
-- that would depend on the prejudices of the programmers, and of the
sources of the material it's fed and trained on.
Google Search can do that now, using weights and plain old statistics. :)
https://en.wikipedia.org/wiki/Horizon_%28disambiguation%29
Computing
Horizon, the Fujitsu IT and accounting system at the heart of the British Post Office scandal
https://en.wikipedia.org/wiki/British_Post_Office_scandal
"The British Post Office scandal, also called the Horizon IT scandal,
involved the Post Office pursuing thousands of innocent subpostmasters for
apparent financial shortfalls caused by faults in Horizon,
an accounting software system developed by Fujitsu.
"
Yes, I was well aware of that, but probably many (most?) won't, hence
my comment. But, as always, even a simple search on "post office
horizon" directly points to the Wikipedia page, you ony the utterly lazy would remain ignorant! :-)
You're probably correct about the *majority* of human creativity, i.e.
in this case code. But *real* AI should be able to create custom code
from a design spec, etc., not from only adapting other existing code.
That's the kind of code I mainly wrote and I like to think that I'm
somewhat intelligent! :-)
I see no reason this cannot be achieved.
As the human operator in front of the machine, you
break the process down into steps.
1. What are the top dozen (or so) publicly available web-based GPT AI's?
2. Are any aimed at privacy (e.g., no mothership-login requirement)?
AI chstbots and privacy are diametric opposites. They are designed to
collect any and all information.
What free web-based GPT chat AI's are we missing in that list?
For what purpose? They have their own strengths and weaknesses.
Which (if any) have no mothership-login tracking requirements?
You're worried about the wrong thing. As usual.
Paul wrote on Sat, 28 Dec 2024 13:24:02 -0500 :
 You're probably correct about the *majority* of human creativity, i.e. >>> in this case code. But *real* AI should be able to create custom code
from a design spec, etc., not from only adapting other existing code.
That's the kind of code I mainly wrote and I like to think that I'm
somewhat intelligent! :-)
I see no reason this cannot be achieved.
As the human operator in front of the machine, you
break the process down into steps.
Bear in mind the goal is to 'chat' using normal voice inputs.
That means to 'speak' to the chat mechanism using normal voice input.
That also means to hear chat output using normal ears that hear things.
We can kick it off by a single button press.
But constantly pressing buttons is not a voice activated chat.
Neither is typing and reading text.
All of you seem to have resolved this problem - but nobody said how.
How do you manage to get the chat to work purely with voice input.
And purely with speech output?
What am I missing that most of you appear to have already figured out?
Please help me understand how you run Google Search completely by sound.
No typing.
No button presses (or at most, one button press in total to start it).
Just voice.
Speech input.
Speech output.
On 12/26/2024 11:06 PM, Char Jackson wrote:
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)
I don't think it's reasonable to suggest that the current state of the
art for
AI is all we'll ever see. On the contrary, it's probably a near
certainty that
what got us to this point will sooner or later be overshadowed by future
advancements. In other words, "ever" is a very long time.
  I'm still surprised that after so many tech advances, over a couple
of hundred years, we're still so naive as to get excited about how
the latest development will change our lives. Of course, there have
been changes. We put clothes into a machine with a button now, to
wash them, and no longer need to hike down to the river with a scrub
board. But those are not changes in quality of life. They're only
details. The tech is only tools.
  Meanwhile, people keep getting excited about how the next
breakthrough will transform our lives. I see that as a symptom of
depression. The people who want a push-button world or who thrill
that AI can look something up online without having to think for
themselves are people who find it a hassle to simply relate to their
own experience. "Let's get a device to save us from having to
actually live our lives. That will be so great!"
  I remember as a child having a futurism mindset. I used to watch
some science show every Sunday evening, hosted by Union Carbide.
The show told me about what's new in science. The commercials
told me about breakthroughs at Union Carbide. Utopia always seemed
just one discovery away. Now the kids are living through devices,
numbed by disconnection, "traumatized" when they have to deal with
humans, yet the mythology of tech utopia is still in effect. Futurism --
"the miracle of tech" -- is still the naive answer to dealing with life's problems. The glee of "What'll they think of next?!" is still a kind of existential landmark of desperate hope against the abyss of death.
"Life sucks and then you die, but have you seen this new aerial
toothbrush that you don't even have to hold? It's incredible."
On 12/27/24 07:22, Newyana2 wrote:
On 12/26/2024 11:06 PM, Char Jackson wrote:
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)
I don't think it's reasonable to suggest that the current state of
the art for
AI is all we'll ever see. On the contrary, it's probably a near
certainty that
what got us to this point will sooner or later be overshadowed by future >>> advancements. In other words, "ever" is a very long time.
I'm still surprised that after so many tech advances, over a couple
of hundred years, we're still so naive as to get excited about how
the latest development will change our lives. Of course, there have
been changes. We put clothes into a machine with a button now, to
wash them, and no longer need to hike down to the river with a scrub
board. But those are not changes in quality of life. They're only
details. The tech is only tools.
Meanwhile, people keep getting excited about how the next
breakthrough will transform our lives. I see that as a symptom of
depression. The people who want a push-button world or who thrill
that AI can look something up online without having to think for
themselves are people who find it a hassle to simply relate to their
own experience. "Let's get a device to save us from having to
actually live our lives. That will be so great!"
I remember as a child having a futurism mindset. I used to watch
some science show every Sunday evening, hosted by Union Carbide.
The show told me about what's new in science. The commercials
told me about breakthroughs at Union Carbide. Utopia always seemed
just one discovery away. Now the kids are living through devices,
numbed by disconnection, "traumatized" when they have to deal with
humans, yet the mythology of tech utopia is still in effect. Futurism --
"the miracle of tech" -- is still the naive answer to dealing with life's
problems. The glee of "What'll they think of next?!" is still a kind of
existential landmark of desperate hope against the abyss of death.
"Life sucks and then you die, but have you seen this new aerial
toothbrush that you don't even have to hold? It's incredible."
Humanity somehow always dreams up naively utopian scenarios while on the application side the money-changers just ram yet another spruce up their chimney with the same 'innovation'. I ask myself will AI-driven cars end
up dominating and the answer is an unequivocal YES. Will that be a good thing? The answer is an unequivocal NO. But dominate they WILL, not
because they let your family admire the scenery along 1000 miles of
beautiful countryside while enjoying smalltalk and meditating about
noble values between A and B but because all that time you will instead
be CONSUMING the good life sold by online merchants bumber-to-bumper at 200mph. There WILL be fewer although more dramatic accidents, no
problemo, what the insurers want is either no accidents or no survivors. Meanwhile the political 'industry' will have to spend less on roads with
more left in the jar for the deep-state controlling traffic via
satellites. When you arrive on a bridge with the unfortunate file of
someone whose ancestors had made the stupid mistake of having a faecesbook/google account from which it has since been computed that you
will soon become a very expensively sick person the electrical steering (already being imposed on everyone today) throws a fit and your car off
the bridge. "Oops, sorry, computer glitch". Nothing is darker than the
dark side, and it's here.
A few years ago when computer billing was in its infancy I got my dose
of double-bills, mostly minor chickenshit but a few of them were like
this one oil bill for $2000, twice. Busy people swamped in paperwork (or whogivesafuck employees) might not have noticed it at all. "Computer
glitch" they said, and refunded me. Funny thing, random chance would see
as many such 'glitches' giving me free money but of this sacrilegious description there were NONE in all these years. Sez it all, get ready
for 1984 forty years late :-)
Oh yeah, AI will be impressively correct and efficient 99% of the time
so soon there won't be a single human being with a career to protect who
will DARE challenge it; you think political-correctness was bad? You
ain't seen nothing like *AI-Correctness* yet.
Oh yeah, AI will be impressively correct and efficient 99% of the time
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 508 |
| Nodes: | 16 (1 / 15) |
| Uptime: | 238:33:35 |
| Calls: | 9,985 |
| Calls today: | 3 |
| Files: | 13,836 |
| Messages: | 6,358,294 |