FSD= fail, smash, dead?

Can't see how you can say that giving how it is performing. As I said from what I am seeing I wouldn't be surprised to see Tesla operating a fully autonomous vehicle next year.

Time will tell.

to an end to end neural net

—

Fancy words for a bit of code.

Anyway, is impressive, but it’s the edge cases that will be its problem (and probably law suits). When some muppet suddenly cuts across it whilst someone else is in the other lane, how it reacts to those things. All five and wellnin normal conditions, it’s the muppetry it will have to contend with that will be problematic

Well if some muppet cuts across anyone, then who knows how they will or should react?

The real edge cases are things like roadworks, traffics light fails, police closures etc.

End to end neural net is far more than code, and is a world away from heuristics. It is a learning Ai, and doing so very quickly. I don't think most epeople understand just how capable Ai is becoming.

comment by Amigawolf no longer MIA (U18508)
posted 38 seconds ago
Well if some muppet cuts across anyone, then who knows how they will or should react?

The real edge cases are things like roadworks, traffics light fails, police closures etc.

End to end neural net is far more than code, and is a world away from heuristics. It is a learning Ai, and doing so very quickly. I don't think most epeople understand just how capable Ai is becoming.
----------------------------------------------------------------------
It’s just a buzzword for code writing code. It’s like calling chat gpt AI- it’s not it’s just clever code.

Sorry Baz, I know you are a decent enough guy, but you don't understand Ai or the profound changes that are going to happen because of it.

Not just in self driving.

comment by Amigawolf no longer MIA (U18508)
posted 4 minutes ago
Sorry Baz, I know you are a decent enough guy, but you don't understand Ai or the profound changes that are going to happen because of it.

Not just in self driving.
----------------------------------------------------------------------
I totally agree that AI will have a huge impact, but right now it’s not what I’d callAI, it’s clever code. Now where there’s code, there’s vulnerabilities , they better be damn sure that the computers on those cars are secure as hell

Ai is starting to code itself. Humans will be out of the loop, as Ai will be smarter than any human. And that future is close very close.

comment by Amigawolf no longer MIA (U18508)
posted 4 minutes ago
Ai is starting to code itself. Humans will be out of the loop, as Ai will be smarter than any human. And that future is close very close.
----------------------------------------------------------------------
Indeed, but not very well it seems https://stackoverflow.blog/2024/03/22/is-ai-making-your-code-worse/?utm_campaign=the-overflow-newsletter&utm_medium=email&utm_source=iterable

What Ai is capable of today will only be better tomorrow. Ai is not going to stop learning, which is why it will be smarter than any human, after that it will be smarter than all humans.

Like I said people do not understand the profound changes Ai will bring. to everything.

comment by Amigawolf no longer MIA (U18508)
posted 4 seconds ago
What Ai is capable of today will only be better tomorrow. Ai is not going to stop learning, which is why it will be smarter than any human, after that it will be smarter than all humans.

Like I said people do not understand the profound changes Ai will bring. to everything.
----------------------------------------------------------------------
Yes but it isn’t AI yet though. It’s not aware or intelligent. It’s silly smart code. It cannot have free thought. It cannot create from nothing. It needs input at all times, therefore won’t be until that is overcome. It’s confined to the borders of its programming currently

We all create from something. An artist will create from his experiences, a music creator will do the same. They might put things together in a creative way, But there is nothing stopping Ai from doing the same.

There are ai programs creating art from simple prompts, creating music, imitating the human voice. There is no limit to what Ai can learn.

comment by Amigawolf no longer MIA (U18508)
posted 1 minute ago
We all create from something. An artist will create from his experiences, a music creator will do the same. They might put things together in a creative way, But there is nothing stopping Ai from doing the same.

There are ai programs creating art from simple prompts, creating music, imitating the human voice. There is no limit to what Ai can learn.


----------------------------------------------------------------------
But they need a prompt. They need someone to give a command. That’s not intelligence. That’s not aware. That’s just clever code . Current ‘ai’ requires a human telling it what to do. Same goes for a car, it won’t just drive, it needs someone to tell it where. I

They do not need a prompt. It can generate it's own prompts if you want.

Again making a car drive anywhere it wants would be a simple task. But an Ai wouldn't want to go anywhere, so of course if a human wants to go somewhere, it will have to tell the car where.

Again making a car drive anywhere it wants would be a simple task. But an Ai wouldn't want to go anywhere,


—/
Exactly. It’s not aware or intelligent. It’s just code reacting. I don’t believe what’s called AI at the moment is true AI. There’s major hurdles to overcome before the true potential of it will be unleashed, til then it’s just smart code doing what it’s asked

Humans are in the loop for sure, but that is because we are putting ourselves there. Ai as I said will create, or drive anywhere if you let it. Not necessarily to human standards yet, but it will soon outgrow humans.

Put it this way, Automation on our roads will be far superior to humans. It is not there yet, but it will be one day. And that could be very soon.

Ai as I said will create

—

It can’t though. It needs some input. Be that’s human telling it what to do, or done human writing code telling it to random or on schedule run something to create something, it still needs input. A human can randomly come up with ideas from just sitting and thinking , this current ai cannot.

I don't know why you think Ai cant come up with random ideas. If you let it like I said it will create music, art, etc.

You are saying it needs a human to tell it to start doing that, yes, but that is because we put ourselves there.

You are saying automation on our roads will fail. FSD is showing it can drive safely, eventually that will be without human intervention, and will be much safer.

Yes the human will tell it where they want to go, but that will be all that is required.

I don't know why you think Ai cant come up with random ideas. If you let it like I said it will create music, art, etc.

—

You’ll have to demonstrate this. Because I don’t know of any dall-e, midjourney etc that can on their own, with no input decide to suddenly create a song in c#, or choose to do a cubist painting? They all need input.

I’m talking about thinking, a true ai that would be capable of the sort of progress you hint at would have to be able to think, they don’t at the moment. They run code based on input, they do not suddenly think, ‘I’d like to create a song in c#..’

Dalle, Midjourney ect will create random prompts if you want. The human input is for more specific images.

And I understand what you are saying. I am not saying Ai is superior to humans level of thinking. I'm saying it will be.

Self driving is a glimpse of that.

And I understand what you are saying. I am not saying Ai is superior to humans level of thinking. I'm saying it will be.

—

I think in terms of doing many tasks, like driving eventually, you are totally correct. But it currently has no free thought, no imagination, it is not self aware- that would be true AI and that’s a huge hurdle to overcome. When that does happen, then we are talking a seismic change and impact on the human race, whether it be skynet or benevolent rulers automating everything, who knows. Right now AI is clever code that’s brilliant at tasks, many of which need Some human input or directive. It’s clever automation at the moment.

Well to be fair that is a long way from your first comment where you said FSD = fail. I am telling you it is happening. You can argue the timescale, but it will happen.

As for the soul so to speak, that is more ethereal in nature, This is just the beginning of Ai remember.

Well to be fair that is a long way from your first comment where you said FSD = fai

—
That was just playing with words and the perception some have of Tesla’s. Don’t forget I also said it was impressive

Yeah but it doesn't equal fail then does it.

As for the soul, I'm not sure giving an Ai free thinking would be a good idea. Although I'm sure someone will do it.