Public web-based AI GPT chat URLs

On Sun, 22 Dec 2024 09:42:43 -0000 (UTC)

Post by Andrew
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?

You are wasting huge resources, clear off bird-brain.

--
Bah, and indeed Humbug.

2024-12-23 13:03:34 UTC

[snip]

Services are not even softwares.

Peter

2024-12-24 03:24:25 UTC

Post by JJ

Services are not even softwares.

I tested the links provided in the OP and two of them work beautifully in
Firefox on Windows WITHOUT creating an account on the Internet.

Mr. Man-wai Chang

2024-12-25 03:49:51 UTC

Post by Andrew
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)?
....
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... go take a walk! :)

Isaac Montara

2024-12-25 18:34:59 UTC

Post by Andrew
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.

Newyana2

2024-12-26 12:00:46 UTC

Post by Isaac Montara

Post by Andrew
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...

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. :)

Paul

2024-12-26 18:53:22 UTC

Post by Isaac Montara

Post by Andrew
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...

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.

Paul

Char Jackson

2024-12-27 04:06:58 UTC

Post by Isaac Montara

Post by Andrew
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...

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 :-)

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.

Post by Paul
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.

Newyana2

2024-12-27 12:22:10 UTC

Post by Paul
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)

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."

Salvador Mirzo

2024-12-27 15:06:35 UTC

Post by Newyana2

Post by Paul
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)

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."

You said it all. ``Depression'' describes it.

bad sector

2024-12-30 12:14:31 UTC

Post by Paul
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)

   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.

Newyana2

2024-12-30 13:43:15 UTC

Post by bad sector

Post by Paul
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)

    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.

That's a vivid vision. I hadn't imagine such evil intentions. Though
I have experienced choicelessness, when people on the other end of
the phone are required to act like simple computers, or there's
no one on the other end of the phone. Our ISP company charges us
through the nose, but we're always officially getting a discount because
their official prices are so high, and no regulation limits it.

A newish trend is that they jack up prices because most people
are on automatic payment. Then if anyone calls to complain, those
people will get a "discount". But as with your oil bill, most people
are not paying attention. Paying bills is a hassle. They want it on
automatic withdrawal. We neglected to consider that by having
computers run our lives, we'd have to live like computers.

We bought our own modem to stop the $18 fee. I got a VOIP
phone to stop the periodic trick of them suddenly doubling my bill.
Yet we're paying more than ever. And these are the good guys.
Our other choices are Comcast or Verizon. Or I guess I could pay
$600 + $90/month to Vice President Elon.

On the bright side, I read on Slashdot the other day that Hertz
bought 100,000 Teslas. It turned out that most people didn't want
to rent an electric car. It's a lifestyle, after all, and not just a case
of visiting a different gas station. And the repairs are apparently
very expensive. Now Hertz is having trouble selling the cars, even
at a 50% discount.

Graham J

2024-12-30 16:32:51 UTC

bad sector wrote:

[snip]

Post by bad sector
Oh yeah, AI will be impressively correct and efficient 99% of the time

Like the infamous Post Office Horizon computer system in the UK !!!!

--
Graham J

Paul

2024-12-27 14:29:48 UTC

Post by Char Jackson
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've done as much testing as I need to do.

S t r a w b e r r y Answer: "2"
^ ^ ^

Only fifty years to go.

Maybe the first full-scale Fusion Reactor,
will have an AI doorman.

Paul

Mr. Man-wai Chang

2024-12-27 14:20:02 UTC

Post by Paul
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

2024-12-27 15:14:00 UTC

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 !!!!

--
Graham J

Frank Slootweg

2024-12-27 16:44:49 UTC

Post by JJ
[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! :-)

Paul

2024-12-27 17:11:20 UTC

Post by JJ
[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.
"
Paul

Frank Slootweg

2024-12-27 18:52:00 UTC

Post by JJ
[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.
"

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! :-)

Andrew

2024-12-28 20:31:32 UTC

Post by Paul
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.
"

All of you seem to have already "solved" the problem, so may I ask any of
you to explain how you solved it using purely speech input & output please?

No typing.
No button presses (or at most, one button press in total to start it).

Just voice.

Speech input.
Speech output.

How did you accomplish that?

Paul

2024-12-27 15:00:19 UTC

A.I. does NOT create, but to read, "analyze", summarize and conclude! :)

They can write code.

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.

******* CoPilot AI Test Run *******

Write a Linear Least Squares calculation program in the C language:

1) Input ordered pairs as floating point X and Y 23.17 14.75
2) Loop on input, accepting new lines of X and Y
until a blank line is detected on the input instead.
3) Work out the linear least squares best fit y = (a * x) + b
Print to stdout "The slope is <value of a> and the intercept is <value of b>"
as to floating point numbers.
4) At the point, the main routine terminates.

#include <stdio.h>
#include <stdlib.h>

int main() {
float x, y;
float sumX = 0, sumY = 0, sumXY = 0, sumX2 = 0;
int count = 0;
char line[256];

while (1) {
printf("Enter X and Y (or blank to finish): ");

// Read a line of input
if (fgets(line, sizeof(line), stdin) == NULL || line[0] == '\n') {
break; // Exit the loop if the input is blank
}

// Parse the input as two floating point numbers
if (sscanf(line, "%f %f", &x, &y) != 2) {
printf("Invalid input, please enter two floating point numbers.\n");
continue;
}

// Accumulate sums for least squares calculation
sumX += x;
sumY += y;
sumXY += x * y;
sumX2 += x * x;
count++;
}

if (count < 2) {
printf("Not enough data points.\n");
return 1;
}

// Calculate the slope (a) and intercept (b)
float a = (count * sumXY - sumX * sumY) / (count * sumX2 - sumX * sumX);
float b = (sumY - a * sumX) / count;

// Print the result
printf("The slope is %.2f and the intercept is %.2f\n", a, b);

return 0;
}

******* end: CoPilot AI Test Run *******

Now, isn't that beautiful ?

Paul

Mr. Man-wai Chang

2024-12-28 08:55:31 UTC

A.I. does NOT create, but to read, "analyze", summarize and conclude! :)

They can write code.
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! ;)

Chris

2024-12-28 10:41:14 UTC

A.I. does NOT create, but to read, "analyze", summarize and conclude! :)

They can write code.
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...

Mr. Man-wai Chang

2024-12-28 13:17:37 UTC

Post by Mr. Man-wai Chang
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...

Back then there was no internet, but only books and magazines, and
programming work hidden in various organiztaions. All codes were typed
into the computers by hands and fingers. :)

Anyway, I agree that it's hard to prove that someone (a human)
invented/created something, especially when it comes to writing.

Chris

2024-12-28 21:41:14 UTC

Post by Mr. Man-wai Chang
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...

Back then there was no internet,

Did you mean "back when"?

Post by Mr. Man-wai Chang
but only books and magazines, and
programming work hidden in various organiztaions. All codes were typed
into the computers by hands and fingers. :)

Ah, the dark ages when RAM was measured in kilobytes and a whole OS could
fit on a floppy disk.

Post by Mr. Man-wai Chang
Anyway, I agree that it's hard to prove that someone (a human)
invented/created something, especially when it comes to writing.

The Horny Goat

2025-01-03 07:58:00 UTC

Post by Chris
Ah, the dark ages when RAM was measured in kilobytes and a whole OS could
fit on a floppy disk.

I only got rid of my unused pirated copy of Windows 1 which was on a
single 5.25" floppy about 5 or 6 years ago when I finally retired my
last machine that had a 5.25" drive.

The Horny Goat

2025-01-03 07:55:42 UTC

On Sat, 28 Dec 2024 21:17:37 +0800, "Mr. Man-wai Chang"

Post by Mr. Man-wai Chang
Back then there was no internet, but only books and magazines, and
programming work hidden in various organiztaions. All codes were typed
into the computers by hands and fingers. :)
Anyway, I agree that it's hard to prove that someone (a human)
invented/created something, especially when it comes to writing.

Absolutely - I typed several thousand of those IBM 80 column punch
cards in my undergraduate days back in the mid to late 1970s.....

On your second point about the only way is a manuscript copy and even
that's not proof - I remember at one point transcribing the first
chapter of Genesis which you can reasonably assume was written before
I arrived on this planet...

Frank Slootweg

2024-12-28 16:19:37 UTC

A.I. does NOT create, but to read, "analyze", summarize and conclude! :)

They can write code.
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! :-)

Paul

2024-12-28 18:24:02 UTC

A.I. does NOT create, but to read, "analyze", summarize and conclude! :)

They can write code.
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...

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) "Prepare an architectural block diagram of the Firefox browser source code"

2) "Using the architectural block diagram of the Firefox browser source code,
remove the redundant or non-functional portions of the browser architecture."
"Optimize the architecture, for modern web browser usage, without losing
any functional blocks while doing so."

3) "Using the answer from the previous step, write source code modules implementing
the functions in the amended architectural block diagram."

While it is theoretically possible to give "hint prompts"

"Work slowly and methodically, thinking about each step and checking your work..."

the test experience is an AI can just drop complete sentences and ignore them.
And the AI does not tell you what it did. You the user, figure out later
that "shit, it dropped my prompt!". It drops prompts that are too hard to satisfy,
and you can tell from the temporal behavior of the screen, that you are about
to experience one of those. There is a pregnant pause, then suddenly the output
picks up speed as if the AI is "having a good day", when in fact it has shed
the constraints you had applied to it. It can drop one or more keywords
from a query, or it can drop a whole sentence that does not lead to a quick
solution.

One of the runs above would cost you thousands of dollars of data center
time, and you would have no assurance the prompt is exactly
right for the job. The job could have cost overruns.

The thing is, when an AI makes a mistake, the mistake
could be trivial, or the mistake could be catastrophic
to the mission ("really really stupid fundamental mistake").

S t r a w b e r r y Answer:"2" (can't count worth shit...)
^ ^ ^

If an AI was passing the time with an old person,
the conversation would go like this...

"Is it snowing where you are ?"

"Did I tell you about my highest bowling score?"

Both parties share the same defect.

If we were back in the year 1980, and running the port
of Eliza, the conversation would go like this...

"Did I tell you about my highest bowling score?"

"What is it about bowling that you do not like?"

Notice that the conversation is slightly more focused,
and the keyword "bowling" was not dropped by Eliza.
It no longer appears that both parties have dementia.

Paul

Andrew

2024-12-28 20:35:04 UTC

Post by Frank Slootweg
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?

Paul

2024-12-28 21:24:18 UTC

Post by Andrew

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?

*******

I have no experience with this, so the following is a purely imaginary list.

1) Check in the permission section, that the "App" which will conduct
the section, has the Microphone permission in the ON state. If
you were doing a Zoom session, you would turn on Microphone Access
and Webcam Video Access. These permissions must be enabled, for
the program or application to get input.

Settings : Privacy and Security : Microphone <=== an example from W11
(No, there is no CoPilot entry)

2) With the application running, look for the Microphone icon in the GUI.
Click the Microphone, to remove the diagonal line through it.
The microphone icon is your mute. If you don't want the Ai to
hear you fart, click the Microphone button and suppress input temporarily.

3) Once the Microphone is enabled, and in the Recording audio level
panel, the gain on the microphone is set to non-zero, say to the AI.
(In the Recording panel, LineIn could be enabled instead of Microphone.)
In the past, a VU meter may have been included at this point, so the
user could calibrate that input was working. Voice recognitions in the
past, had a calibration phase just so the user would know it was primed
and ready to go.

Hey, Cortana
Hey, CoPilot

Use one of the popular prompts, to start voice recognition.

The process is, of necessity, not fully automated as such.
If a session does not work, you have to debug it.

Sometimes, there is a logical failure. If the network cable
is unplugged, you could be babbling away to the Microphone,
to no effect. Do not expect the software to be smart enough,
to alert you in clear English text, to plug in the network cable.
It does not do that for anything else and the AI session will
be no different.

This should not be much different from getting Zoom to work :-)

Check the Accessibility section, and see if there is anything
specific about enabling microphones or speech by default.
There may be such a feature, if the OS supports accessibility
via microphone.

It's a research topic, not a slam dunk. This is not a smart
phone, it's a desktop, and desktops hate automation.

How many times have I tested audio and got a punch in the face ?
Most all of my electrets, don't even work on the computer.

Paul

Chris

2024-12-28 21:43:44 UTC

A.I. does NOT create, but to read, "analyze", summarize and conclude! :)

They can write code.
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...

It can do that. Not perfectly, but pretty well.

Mr. Man-wai Chang

2024-12-27 14:21:11 UTC

Post by Paul
Will an AI ever re-write Firefox in one gulp ? The answer is NO :-)

I prefer the answer "not yet today :-)"

Wrong! Once Firefox's source codes become publicly available, A.I. can
... you know... :)

Mr. Man-wai Chang

2024-12-27 14:18:10 UTC

Post by Isaac Montara
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?

In the future, Google Search can possibly embrace A.I. seamlessly! :)

Steve Hayes

2024-12-28 02:36:08 UTC

On Fri, 27 Dec 2024 22:18:10 +0800, "Mr. Man-wai Chang"

Do you really need to ask A.I. when Google Search can do the job fine?

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.

--
Steve Hayes from Tshwane, South Africa
Web: http://www.khanya.org.za/stevesig.htm
Blog: http://khanya.wordpress.com
E-mail - see web page, or parse: shayes at dunelm full stop org full stop uk

Mr. Man-wai Chang

2024-12-28 08:56:30 UTC

Post by Steve Hayes
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. :)

Andrew

2024-12-28 20:31:57 UTC

Google Search can do that now, using weights and plain old statistics. :)

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.

I don't know how to do it but you've apparently figured it out.
How did you all get Google Search input and output to be purely by voice?

Andy Burns

2024-12-29 13:01:51 UTC

Post by Andrew
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 a Pixel phone with with an associated google account, it can be done.

To be totally handsfree the phone must be unlocked (e.g. it's kept
unlocked by geolocation, or detecting a nominated bluetooth device) just
enable the "hey google" assistant wake-word in settings.

It will wake-up, launch gemini and answer a voice question with a voice
answer (plus a transcript on screen), if it recognises it as a command
rather than a question, it will pass it to google assistant instead of
gemini.

I believe this applies to [some?] Samsung phones as well as Google phones.

But I don't normally have "hey google" enabled on my phone.

The Horny Goat

2025-01-03 08:02:02 UTC

On Sat, 28 Dec 2024 16:56:30 +0800, "Mr. Man-wai Chang"

Google Search can do that now, using weights and plain old statistics. :)

It would be an interesting exercise to feed Jonathan Swift's "A Modest
Proposal" into ChatGPT and see if it understood that it was a
magnificent satire.

Anybody want to try? Here's the text: (which you can copy and paste)

https://www.gutenberg.org/files/1080/1080-h/1080-h.htm

Chris

2024-12-27 11:55:07 UTC

AI chstbots and privacy are diametric opposites. They are designed to
collect any and all information.

Post by Andrew
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?

For what purpose? They have their own strengths and weaknesses.

Post by Andrew
Which (if any) have no mothership-login tracking requirements?

You're worried about the wrong thing. As usual.

Andrew

2024-12-28 20:41:38 UTC

Post by Andrew
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.

No argument here. Privacy is a complex web. That's why the need for the
lack of login requirements is critical. You understood that requirement.

Most people do not understand there can never be privacy with a login.
You do.

That's good.

But that doesn't change the question since only TWO of the list below allow
the privacy of NOT having to need to log into a mothership Internet server.

Post by Andrew
What free web-based GPT chat AI's are we missing in that list?

For what purpose? They have their own strengths and weaknesses.

I volunteer to help old people in town and most of them love the idea of
chat but they can't interact well with the mechanisms, so another important
goal is to have speech input and output (which I forgot to add to the OP).

What speech input/output chat mechanisms do you know of that are free to
old people and which do not require them to create a mothership account?

I only know of one, but it has a fifteen-minutes-per-month limitation.