mmccarron.bsky.social - Profile | ThreadSky | a Reddit-style client for Bluesky

Tree rings show Earth was hit by an extreme solar storm 14,300 years ago, likely caused by multiple CMEs striking in succession. It flooded the atmosphere with radiation, far beyond modern levels, and shows how intense solar activity can become. 🔭🧪⚒️ #SolarStorms #SpaceWeather #Paleoclimate

submitted 18 days ago • 3 comments

This article takes a brief look at the #AI and #Robotics safety issues and demonstrates an ChatGPT jailbreak. This is a good piece if you want to understand the systemic issue at a deeper level. yetagainanotherhack.wordpress.com/2025/02/11/a...

submitted 95 days ago • 0 comments

I think modern computer software was a mistake and we should go back to our roots of developing for limited hardware. The same amount of electricity that could power MS-DOS for a year is now used to ask Copilot how to save a Word document to your local Desktop instead of OneDrive.

submitted 127 days ago • 157 comments

The problem with agents will be the accumulation of errors limiting the complexity of tasks which can be performed.

submitted 170 days ago • 0 comments

Hallucinations in GenAI are caused by a malformed higher dimensional object.

submitted 170 days ago • 0 comments

People who think consciousness will emerge in GenAI don't understand how Transformers work. There is no intelligence, instead there is a mathematical object which makes language and real-world relationships computable.

submitted 170 days ago • 0 comments

What we need is a more formal approach to discovering the optimal higher dimensional object for a given language, rather than trying to construct it from data. This will probably be the next phase of AGI research.

submitted 170 days ago • 0 comments

The real problem with GenAI and why hallucinations occur is because the higher dimensional object which captures the relationships is imperfect.

submitted 170 days ago • 0 comments

I think it is time we move away from concepts such as learning when it comes to GenAI. GenAI represents information in higher dimensions which makes the problem, to a degree, linearly separable.

submitted 170 days ago • 0 comments

GR's deep symmetries and the mathematical structures of relativistic theories serve as a form of intrinsic hypervisor. #physics

submitted 177 days ago • 0 comments

An article to make you pause and think about life for a while. We don't do this often enough, instead we mainly put the blinkers on continue on with our day. Take a existential walk off the deep end with me. yetagainanotherhack.wordpress.com/2024/12/01/t...

submitted 186 days ago • 0 comments

AI isn't just learning from humans; it’s reflecting our biases, aspirations, and fears. In many ways, AI is a mirror. What is #AI like without the guardrails? What would that say about us?

submitted 188 days ago • 0 comments

What happens when machines create things we don’t understand? #AI can write poetry or code, but does it mean anything to itself?

submitted 188 days ago • 0 comments

Currently available LLMs are more intelligent than the vast majority of people, especially when we consider the breadth of knowledge they have. What you are really get at is a system which has its own agency and that's not possible with computational systems.

submitted 193 days ago • 0 comments

But, they were partially correct. The printing press played a key role in the largest scale conflicts and mass deaths the world has ever known. #AI has the potential to do the same. Fact checking may not resolve the issue. Take Eugenics for example, scientifically correct, but morally wrong.

submitted 193 days ago • 1 comment

Still one of my favourites... www.youtube.com/watch?v=1O_I...

submitted 193 days ago • 0 comments

🚀 Who says you need a supercomputer for #AI? Meet #BabyGPT - a language model trained on a single laptop by our senior researcher Dr Amjad Majid. Read more: bit.ly/4fvHQR7

submitted 194 days ago • 0 comments

Scaling alone won't be enough to achieve #AGI.

submitted 200 days ago • 0 comments

Discover what it will take to make the leap to #AGI / #ASI from #LLMs. This article will take you on a deep dive into the world of strong #AI yetagainanotherhack.wordpress.com/2024/11/15/f...

This article takes a brief look at the #AI and #Robotics safety issues and demonstrates an ChatGPT jailbreak. This is a good piece if you want to understand the systemic issue at a deeper level. yetagainanotherhack.wordpress.com/2025/02/11/a...

I think modern computer software was a mistake and we should go back to our roots of developing for limited hardware. The same amount of electricity that could power MS-DOS for a year is now used to ask Copilot how to save a Word document to your local Desktop instead of OneDrive.

The problem with agents will be the accumulation of errors limiting the complexity of tasks which can be performed.

Hallucinations in GenAI are caused by a malformed higher dimensional object.

People who think consciousness will emerge in GenAI don't understand how Transformers work. There is no intelligence, instead there is a mathematical object which makes language and real-world relationships computable.

What we need is a more formal approach to discovering the optimal higher dimensional object for a given language, rather than trying to construct it from data. This will probably be the next phase of AGI research.

The real problem with GenAI and why hallucinations occur is because the higher dimensional object which captures the relationships is imperfect.

I think it is time we move away from concepts such as learning when it comes to GenAI. GenAI represents information in higher dimensions which makes the problem, to a degree, linearly separable.

GR's deep symmetries and the mathematical structures of relativistic theories serve as a form of intrinsic hypervisor. #physics

An article to make you pause and think about life for a while. We don't do this often enough, instead we mainly put the blinkers on continue on with our day. Take a existential walk off the deep end with me. yetagainanotherhack.wordpress.com/2024/12/01/t...

AI isn't just learning from humans; it’s reflecting our biases, aspirations, and fears. In many ways, AI is a mirror. What is #AI like without the guardrails? What would that say about us?

What happens when machines create things we don’t understand? #AI can write poetry or code, but does it mean anything to itself?

Currently available LLMs are more intelligent than the vast majority of people, especially when we consider the breadth of knowledge they have. What you are really get at is a system which has its own agency and that's not possible with computational systems.

But, they were partially correct. The printing press played a key role in the largest scale conflicts and mass deaths the world has ever known. #AI has the potential to do the same. Fact checking may not resolve the issue. Take Eugenics for example, scientifically correct, but morally wrong.

Still one of my favourites... www.youtube.com/watch?v=1O_I...

🚀 Who says you need a supercomputer for #AI? Meet #BabyGPT - a language model trained on a single laptop by our senior researcher Dr Amjad Majid. Read more: bit.ly/4fvHQR7

Scaling alone won't be enough to achieve #AGI.

Discover what it will take to make the leap to #AGI / #ASI from #LLMs. This article will take you on a deep dive into the world of strong #AI yetagainanotherhack.wordpress.com/2024/11/15/f...

A deep remix of a classic... www.youtube.com/watch?v=okVE...

One of my favourites... www.youtube.com/watch?v=lWc-...