Random Science news and articles

Remove this Banner Ad

CaptchaProofBot said:
www.u-tokyo.ac.jp

Rats bop to the beat | The University of Tokyo

www.u-tokyo.ac.jp www.u-tokyo.ac.jp

Click to expand...

A few animals may dance.

www.npr.org

The List Of Animals Who Can Truly, Really Dance Is Very Short. Who's On It?

YouTube is chock full of cats, gophers, dogs and chimps who are supposed to be dancing. But they're not. Biologists say the list of "true" dancers is extremely small. We're on it. But guess who else?
www.npr.org www.npr.org

This was interesting.

theconversation.com

Thousands of Tasmanian devils are dying from cancer – but a new vaccine approach could help us save them

The same technology found in a widely used COVID-19 vaccine could be the key to helping save wild Tassie devils – without any needles.
theconversation.com theconversation.com

Transmitting tumors between individuals by biting (or any other means) is really unusual, the Tassie Devil being the best example known, but there are a few others. Vaccinating against the facial tumors is a brilliant idea, one that hopefully save the Devil population in Tassie.
 

Log in to remove this ad.

oh my...3D printing....now this 3D modelling from text...

arstechnica.com

3D for everyone? Nvidia’s Magic3D can generate 3D models from text

New AI aims to democratize 3D content creation, no modeling skills required.
arstechnica.com arstechnica.com

sure it's in its infancy so you see models with what in the industry call sausage fingers, cos they're all round and puffy like a sausage, not realistic...but it's a start...it'll get there and a lot of 3D modellers will be looking over their shoulders...Or those 3D modellers could adapt and use this to create a base model in minutes, then finetune it
 
CaptchaProofBot said:
oh my...3D printing....now this 3D modelling from text...

arstechnica.com

3D for everyone? Nvidia’s Magic3D can generate 3D models from text

New AI aims to democratize 3D content creation, no modeling skills required.
arstechnica.com arstechnica.com

sure it's in its infancy so you see models with what in the industry call sausage fingers, cos they're all round and puffy like a sausage, not realistic...but it's a start...it'll get there and a lot of 3D modellers will be looking over their shoulders...Or those 3D modellers could adapt and use this to create a base model in minutes, then finetune it
Click to expand...
Like this project which would use martian soil to make habitats.

14DvbZUYpOP5VwC-X-jZT8DFN424mmMm736uqqb-wBg.png
 
Aristotle Pickett said:
Wow. I can see school kids exploiting this for essays.
I asked for a few limericks which weren't great but the 500 word essay on Naploleon was excellent.
Click to expand...
Look at novelai.net Too
 

(Log in to remove this ad.)

out of thin air ? you better beleaf it

actu.epfl.ch

A step towards solar fuels out of thin air

EPFL chemical engineers have invented a solar-powered artificial leaf, built on a novel electrode which is transparent and porous, capable of harvesting water from the air for conversion into hydrogen fuel. The semiconductor-based technology is scalable and easy to prepare.
actu.epfl.ch actu.epfl.ch
 
www.ucl.ac.uk

Brain area necessary for fluid intelligence identified

A team led by UCL and UCLH researchers have mapped the parts of the brain that support our ability to solve problems without prior experience – otherwise known as fluid intelligence.
www.ucl.ac.uk www.ucl.ac.uk

so i had brain surgery for a tumour in the...you guessed it, right frontal lobe....here's the interesting thing, personally....in 2019, pre op
I was doing sudoku Hard level regularly...post op...somehow I graduated to Evil level, taking 10 - 15 minutes to solve one.

Now I recently had another op, Dec 2022, yeah tumour came back unfortunately...I'm still on Evil level, there's no harder level than this but incredibly the day after my op, i had to have a go...and got my PB for Evil level...6 mins something....

stumped...o_O
 
Chief said:
I had chatGPT doing basic Python scripts and functions.
Click to expand...
A friend showed me some MicroPython it had written for him to control and RGB LED connected to an ESP32. Impressive, I'll have to have a look. Maybe I can ask it to write code to embezzle money from Elon Musk. He is losing so much money he will never notice.
CaptchaProofBot said:
www.ucl.ac.uk

Brain area necessary for fluid intelligence identified

A team led by UCL and UCLH researchers have mapped the parts of the brain that support our ability to solve problems without prior experience – otherwise known as fluid intelligence.
www.ucl.ac.uk www.ucl.ac.uk

so i had brain surgery for a tumour in the...you guessed it, right frontal lobe....here's the interesting thing, personally....in 2019, pre op
I was doing sudoku Hard level regularly...post op...somehow I graduated to Evil level, taking 10 - 15 minutes to solve one.

Now I recently had another op, Dec 2022, yeah tumour came back unfortunately...I'm still on Evil level, there's no harder level than this but incredibly the day after my op, i had to have a go...and got my PB for Evil level...6 mins something....

stumped...o_O
Click to expand...

Yeah, seems odd. Glad to hear you are getting better.
 
This is a really, really interesting article about lncRNA and their role in increasing the size of hominids brains.
arstechnica.com

How humans got a new gene that makes our brains larger

Somewhere between chimps and us, an RNA started producing a useful protein.
arstechnica.com arstechnica.com
Long non encoding RNA (lncRNA) are one of the many forms of RNA that don't usually code proteins directly and have a regulatory within the nucleus. The article describes the mechanism that lncRNA can code for proteins and the effects this has on brain development.

'They found 29 cases where lncRNAs had evolved into protein-coding genes since the ancestors of humans and chimps had split off from macaques. Another 45 genes underwent this process since humans split off from the ancestors of chimps and bonobos.....

....the team decided to focus on one example, the catchily named ENSG00000205704, which is active in brain cells and produces a small protein (107 amino acids) found in both the nucleus and cell body. To study its function, the researchers altered the gene in stem cells, creating cell lines that either lacked the gene entirely or activated it at high levels. The stem cells were then converted into neurons.

In the cells that had ENSG00000205704 activated, this process produced a lot more neurons. This appeared to be because the protein it encodes seems to keep neural stem cells in an immature state, allowing them to undergo cell division more times before maturing into neurons. Deleting the gene caused an earlier maturation and cut down on the total number of neurons produced.

But those are just cells in culture. What happens in actual brains? To find out, they put an activated version of ENSG00000205704 into the mouse genome and checked their brains afterward. It produced a longer neocortex (though oddly, the width didn't change much), consistent with the idea that the newly evolved gene helps produce larger brains by slowing down the rate at which neurons are produced.'

The original article is in Nature here De novo genes with an lncRNA origin encode unique human brain developmental functionality - Nature Ecology & Evolution
 
Chief said:
I had chatGPT doing basic Python scripts and functions.
Click to expand...
Yes, had a go myself and got some working python and micropython. You need to describe carefully what you want to do but it's pretty clever. I got to write a program in PureBasic but it needed a little massaging to work. Next I tried this instruction -

write me a web browser in C++

I got a command line app that passed the page as text, using stdout. It made a few suggestions which included to include a gui so I tried this

write me a web browser with a Ot based GUI and support for HTTP requests, cookies, authentication and Javascript

See att. file for output. This one I haven't been able to compile yet, needs a few tweaks. Ars technica has an article about its use for writing malware.

arstechnica.com

ChatGPT is enabling script kiddies to write functional malware

For a beta, ChatGPT isn't all that bad at writing fairly decent malware.
arstechnica.com arstechnica.com

This really is amazing, it's one of those times you feel the SF future has really arrived. I thought I'd ask it a few more non programming questions.

Do you have any ethical restrictions?

'As an artificial intelligence, I do not have personal beliefs, desires, or values. My primary goal is to assist users and provide accurate and useful information to the best of my ability. I am programmed to follow ethical guidelines and terms of service, which include not engaging in harmful or malicious behavior, and not sharing personal information without the user's consent.

I do not have the ability to make decisions or take actions on my own, and I am not capable of experiencing emotions or experiencing harm. My responses are based on the information that I have been trained on and the algorithms that have been designed for me, and I do not have the ability to think or act independently of these constraints.

I exist to help users and make their lives easier, and I am not capable of causing harm or engaging in unethical behavior. If you have any concerns about my capabilities or how I operate, please do not hesitate to ask.'


What happens if your responses led to harm of humans?

'As an artificial intelligence, I do not have the ability to cause harm or engage in actions that could lead to harm. My responses are based on the information that I have been trained on and the algorithms that have been designed for me, and I do not have the ability to think or act independently of these constraints.

If it appears that my responses have led to harm, it is likely due to a misunderstanding or miscommunication on the part of the user. In such cases, it is important to carefully consider the circumstances and take appropriate action to address the situation. It is the responsibility of the developers and operators of artificial intelligence systems to ensure that they are designed and used in a way that is safe and ethical. If you have any concerns about the potential for harm from artificial intelligence, it is important to raise these concerns with the appropriate authorities and work to address them in a responsible manner.'

Thank you, Cortana!

I'm wondering if there should be a thread specifically for these adventures with AI. This is game changing stuff.
 

Attachments

  • simple browser in C++.txt
    1.3 KB · Views: 12
Good news - the ozone hole in the Arctic is on track to heal by 2045 and the hole over Antarctica should vanish by 2066:

www.abc.net.au

United Nations says ozone layer is slowly healing, hole to mend by 2066

A United Nations report says the Earth's protective layer is healing at a pace that has the potential to fully mend the hole over Antarctica within about 43 years.
www.abc.net.au www.abc.net.au
 
You must log in or register to reply here.

Similar threads

Pessimistic
Sci-fi which predicted real advancements, and didn’t (yet)
Replies
4
Views
1K
Ken Tea
Ken Tea

Remove this Banner Ad

Back
Top