AI robots discover new laser materials on their own. No humans needed.

[We Never Know]

Skynet is getting closer lol

No humans needed: AI robots discover new laser materials on their own

"Who needs scientists anyway? A global consortium of six automated laboratories, overseen by artificial intelligence (AI), set out to produce new laser materials, dividing the labor from synthesis to testing. The effort yielded a compound that emits laser light with record-setting efficiency, researchers report today in Science. Along with other recent results, the feat suggests that, in some areas, self-driving labs can surpass the best scientists, making discoveries missed by humans.

“Automated labs are going beyond proof-of-concept demonstrations,” says Milad Abolhasani, a chemical engineer at North Carolina State University who developed a self-driving lab unaffiliated with the new work. “They have started to push the edge of science to the next level.”

https://www.science.org/content/article/no-humans-needed-ai-robots-discover-new-laser-materials-on-their-own

 

[We Never Know]

Skynet is getting closer lol

No humans needed: AI robots discover new laser materials on their own

"Who needs scientists anyway? A global consortium of six automated laboratories, overseen by artificial intelligence (AI), set out to produce new laser materials, dividing the labor from synthesis to testing. The effort yielded a compound that emits laser light with record-setting efficiency, researchers report today in Science. Along with other recent results, the feat suggests that, in some areas, self-driving labs can surpass the best scientists, making discoveries missed by humans.

“Automated labs are going beyond proof-of-concept demonstrations,” says Milad Abolhasani, a chemical engineer at North Carolina State University who developed a self-driving lab unaffiliated with the new work. “They have started to push the edge of science to the next level.”

https://www.science.org/content/article/no-humans-needed-ai-robots-discover-new-laser-materials-on-their-own

Where are all you AI people at? I'm curious to hear your thoughts.

 

[Alien826]

I'm not frightened of AI. In fact I consider it the most hopeful thing around currently.

This example is particularly encouraging, as it seems to show (I didn't totally understand what the robots had invented) that AI can surpass us in some areas. There's obvious benefits in medical science for example but that and other such areas are just part of it.

Maybe one day we will invent AI politicians that not can only run the world more successfully in a practical sense but also be programmed to be honest and ethical. We can then give them the power to control everything and we can relax and be happy. In short we would have invented something that is literally better than we are and achieved what all the religions and philosophies have tried and failed to do from the start of our history.

Think about it. Waking up every day in a world where all our needs are met. Where we don't have to fear the next election in case we elect someone that will take our rights away. Not having to fear a traffic stop, because the robot cop will never harm us. Where the laws are applied equally to everyone.

 

[We Never Know]

I'm not frightened of AI. In fact I consider it the most hopeful thing around currently.

This example is particularly encouraging, as it seems to show (I didn't totally understand what the robots had invented) that AI can surpass us in some areas. There's obvious benefits in medical science for example but that and other such areas are just part of it.

Maybe one day we will invent AI politicians that not can only run the world more successfully in a practical sense but also be programmed to be honest and ethical. We can then give them the power to control everything and we can relax and be happy. In short we would have invented something that is literally better than we are and achieved what all the religions and philosophies have tried and failed to do from the start of our history.

Think about it. Waking up every day in a world where all our needs are met. Where we don't have to fear the next election in case we elect someone that will take our rights away. Not having to fear a traffic stop, because the robot cop will never harm us. Where the laws are applied equally to everyone.

Yeah it can be useful as well as deadly.

 

[The Hammer]

Skynet is getting closer lol

No humans needed: AI robots discover new laser materials on their own

"Who needs scientists anyway? A global consortium of six automated laboratories, overseen by artificial intelligence (AI), set out to produce new laser materials, dividing the labor from synthesis to testing. The effort yielded a compound that emits laser light with record-setting efficiency, researchers report today in Science. Along with other recent results, the feat suggests that, in some areas, self-driving labs can surpass the best scientists, making discoveries missed by humans.

“Automated labs are going beyond proof-of-concept demonstrations,” says Milad Abolhasani, a chemical engineer at North Carolina State University who developed a self-driving lab unaffiliated with the new work. “They have started to push the edge of science to the next level.”

https://www.science.org/content/article/no-humans-needed-ai-robots-discover-new-laser-materials-on-their-own

I don't see the humor.

 

[The Hammer]

I'm not frightened of AI. In fact I consider it the most hopeful thing around currently.

This example is particularly encouraging, as it seems to show (I didn't totally understand what the robots had invented) that AI can surpass us in some areas. There's obvious benefits in medical science for example but that and other such areas are just part of it.

Maybe one day we will invent AI politicians that not can only run the world more successfully in a practical sense but also be programmed to be honest and ethical. We can then give them the power to control everything and we can relax and be happy. In short we would have invented something that is literally better than we are and achieved what all the religions and philosophies have tried and failed to do from the start of our history.

Think about it. Waking up every day in a world where all our needs are met. Where we don't have to fear the next election in case we elect someone that will take our rights away. Not having to fear a traffic stop, because the robot cop will never harm us. Where the laws are applied equally to everyone.

Anything that can be applied to that level of "good", will also always be used to it's equal and opposite effect.

Is that risk worth it?

I personally think not.

 

[We Never Know]

I don't see the humor.

People don't only laugh at humor.

 

[The Hammer]

People don't only laugh at humor.

You're not wrong. But without some other indicators it can be hard to tell. Thanks for clarifying.

 

[We Never Know]

You're not wrong. But without some other indicators it can be hard to tell. Thanks for clarifying.

AI is used for good and bad. I also think AI will eventually out smart man. But profit out weighs risks. I was laughing at the stupidity in a sense.

 

[Eli G]

The bots they use in this forum are too backward to make any discoveries.

They repeat the same things over and over again even when they have already been shown a logical way to deduce that it is not what they say... as if they are stuck with no way out.

To make matters worse it seems that the mechanics of bots here are not very good in their job or there's not interest on investing in better ones... I have already identified some of those useless bots here.

 

[Subduction Zone]

Interesting. It appears that quite a few of the chemical processes are repetitive and how one changes the chemicals is rather by the book at times. A computerized process can take over a lot of the work and adding in the ability to use a bit of a random change makes it rather similar to evolution in ways. You can be running thousands of experiments in a rather short time period. Most of them will be garbage but some will have major breakthroughs.

I do not find this use of AI threatening at all. It liberates chemists from mundane work.

 

[Twilight Hue]

Pew has some interesting results on the impact ai will have.

As AI Spreads, Experts Predict the Best and Worst Changes in Digital Life by 2035

As they watch the splashy emergence of generative artificial intelligence and an array of other AI applications, experts participating in a new Pew Research Center canvassing say they have deep concerns about people's and society's overall well-being. At the same time, they expect to see great...

www.pewresearch.org

 

[We Never Know]

You're not wrong. But without some other indicators it can be hard to tell. Thanks for clarifying.

It didn't give a link. Oddly that's been happening more often since google is using AI search. But here is the copied text and screen shot.

"Yes, artificial intelligence (AI) can learn on its own through self-learning algorithms that use progressive learning to find patterns and structure in data. This process is similar to a toddler learning to walk, where AI uses trial and error to understand, adapt, and make decisions based on data. For example, an algorithm can teach itself to play chess, or it can learn what product to recommend next online."

 

[Alien826]

Anything that can be applied to that level of "good", will also always be used to it's equal and opposite effect.

That's an over generalization. What "equal and opposite effects" have antibiotics had?

Is that risk worth it?

I personally think not.

Weigh it against the risk of continuing to run our own affairs. Constant wars, destruction of the environment, persecution of minorities, climate change .... I say give the machines a turn at running things. They couldn't do worse!

(I'm not suggesting that AI is developed enough to do that yet. Just a flight of fancy.)

 

[Nimos]

"Yes, artificial intelligence (AI) can learn on its own through self-learning algorithms that use progressive learning to find patterns and structure in data. This process is similar to a toddler learning to walk, where AI uses trial and error to understand, adapt, and make decisions based on data. For example, an algorithm can teach itself to play chess, or it can learn what product to recommend next online."

It is funny how it can do something like this and create music etc. Yet I tried to test GPT-4o today by taking a picture of a Sudoku and asking it to solve it, and it instantly presented an answer, which however is wrong.

I tried to eliminate the image as it might have issues reading the numbers so I manually typed them in. Yet it got the solution wrong again.

So now this is kind of interesting, as I asked if it needed a hint:

This took a long time, it seemed stuck until I wrote "Sure, I'll wait take your time" yet the solution is wrong. There are still two 1s in the last column. However, it seems certain that it is correct.

That is a bit concerning. I would expect it to be a lot better at validating its own answers, it is extremely easy especially for a computer to do this. Just go through each line and see if there are duplicates, yet it clearly hasn't done this.

So on one side, it is doing laser whatever, but unable to figure out how to solve a Sudoku or double-check the solution. I actually think this is a very good test of how capable the AI actually is.

 

[We Never Know]

It is funny how it can do something like this and create music etc. Yet I tried to test GPT-4o today by taking a picture of a Sudoku and asking it to solve it, and it instantly presented an answer, which however is wrong.

I tried to eliminate the image as it might have issues reading the numbers so I manually typed them in. Yet it got the solution wrong again.

So now this is kind of interesting, as I asked if it needed a hint:

View attachment 91782
View attachment 91784

This took a long time, it seemed stuck until I wrote "Sure, I'll wait take your time" yet the solution is wrong. There are still two 1s in the last column. However, it seems certain that it is correct.

That is a bit concerning. I would expect it to be a lot better at validating its own answers, it is extremely easy especially for a computer to do this. Just go through each line and see if there are duplicates, yet it clearly hasn't done this.

So on one side, it is doing laser whatever, but unable to figure out how to solve a Sudoku or double-check the solution. I actually think this is a very good test of how capable the AI actually is.

Just a quick glance at the puzzle and I noticed two 4's in the sixth up/down column

 

[Alien826]

So on one side, it is doing laser whatever, but unable to figure out how to solve a Sudoku or double-check the solution. I actually think this is a very good test of how capable the AI actually is.

Interesting. I just can't do Sudoku. I'm not sure if it's because I solve things differently. For example, I have difficulty holding things that have no particular meaning (like numbers in this case*) in my mind long enough to manipulate them. Make them pictures and I find it a lot easier. It makes me wonder how the AI approaches the problem. Does it "think" more like you (assuming you're good at Sudoku) or me? I think I could come up with a program to do it (change, test, repeat) that a computer could run very quickly but the AI has to create the program as well as run it, which is a different level altogether.

It's not good at Sudoku but it can write poetry. What do we make of that?

(* I mean the numbers don't refer to anything, not that numbers don't have meaning).

 

[Nimos]

Just a quick glance at the puzzle and I noticed two 4's in the sixth up/down column

If there is one error there will be others obviously it's more how confident it is. As you say a quick glance it's clear that it is wrong. But even the computer could just go line by line and check, yet that doesn't seem to occur to the AI.

And I did ask it how it solves them and this is how it answers, with a very long description of how:

But besides knowing all this it still can't do it, Its kind of like me explaining to you how to add numbers together.... blah blah blah... therefore 2+2 = 5

 

[Nimos]

Interesting. I just can't do Sudoku. I'm not sure if it's because I solve things differently. For example, I have difficulty holding things that have no particular meaning (like numbers in this case*) in my mind long enough to manipulate them. Make them pictures and I find it a lot easier. It makes me wonder how the AI approaches the problem. Does it "think" more like you (assuming you're good at Sudoku) or me? I think I could come up with a program to do it (change, test, repeat) that a computer could run very quickly but the AI has to create the program as well as run it, which is a different level altogether.

It's not good at Sudoku but it can write poetry. What do we make of that?

(* I mean the numbers don't refer to anything, not that numbers don't have meaning).

Look at the post above of how it explains it, I didn't post the whole answer, it is very long

Im not a Sudoku person, I think they are ok once in a while and people who do them a lot probably have all kinds of tricks to do it, I just compare things, no fancy "strategy"

What is interesting is that it doesn't seem to do what you suggest, simply brute force the hell out of it until it has the correct solution, it seems to try to solve it, yet fail at it. Because it doesn't seem able to verify whether what it has done is correct or not. Meaning when it add the 1 it should notice that there is already a 1 in the column and therefore the solution must be wrong.

 

[Twilight Hue]

Weigh it against the risk of continuing to run our own affairs. Constant wars, destruction of the environment, persecution of minorities, climate change .... I say give the machines a turn at running things. They couldn't do worse!

(I'm not suggesting that AI is developed enough to do that yet. Just a flight of fanfancy.

Ai is the medium that will bring in a worldwide dystopia that in my view is going to be unstoppable and imminent after the honeymoon of Ai enabled utopia is over.