Skip to main content

What ozone-depleting substances can tell us about governance of AGI

 



There are not too many YouTubers that get it. That balance of fascination and constrained horror of what we are witnessing as AI developments occur, that seek out the latest papers, that seek to explain their salient points, and know which ones to choose from the multitude. Thankfully there are channels, only, a very few, like AI Explained, and thankfully too readers of this blog like Just Matthew, who help inspire the content. 

In this latest video, that he published just three hours before writing this, the person (or persons) behind the AI explained channel explored a number of different papers, some of which I've covered in this blog, some of which I've partially read. There's also some tasty surprises. Whilst I was researching through some less than original work, in order to write today's offerings, I missed the launch of the paper, 'Governance of SuperIntelligence' by OpenAI. (Do note that Altman finished his Ted Talk with his stated aim of creating AGI, now the statement concerns itself with ASI. That's a significant change of intent). 

So thanks AI explained. Please watch the video, it's under 20 minutes long and covers far more than I will in this blog.

 Back to the document from OpenAI. It starts: 

'Given the picture as we see it now, it’s conceivable that within the next ten years, AI systems will exceed expert skill level in most domains, and carry out as much productive activity as one of today’s largest corporations.'

It ends, all to briefly, with:

 we believe it would be unintuitively risky and difficult to stop the creation of superintelligence. Because the upsides are so tremendous...

If you aren't concerned that a corporate entity is stating this, then I don't know what to say.

After the first few weeks, whilst researching for these blog posts, I would have questioned the feasibility of OpenAI's opening statements. Now, I hold slight doubts. My optimism for the control of more advanced AI systems are mostly held in these slight doubts, that what is being promised/threatened, is not possible. My doubts, the more I learn, are diminishing. Do I think human institutions are up to the task of combatting my fears of AI? Well, let's have a little journey back into recent history. 

When chlorine and bromine atoms come into contact with ozone in the stratosphere, they destroy ozone molecules. One chlorine atom can destroy over 100,000 ozone molecules before it is removed from the stratosphere. Ozone can be destroyed more quickly than it is naturally created. In 2000, the ozone hole reached its maximum extent since 1979 and has stopped increasing in size in subsequent years, which is attributable to the phasing out of ozone-depleting substances under the Montreal Protocol (for more information, see the EEA indicator 'Consumption of ozone-depleting substances' 

This sounds great, legislative action by the world's nations can affect positive change, what I was concerned about in the early 80's is in recession... but there is a graph that I'd like to present to you, from Copernicus


Look at the last two years on the infographic. What's the new trend? Stopping ozone-depleting substance is a relatively simple task. Ask yourself though, how does that compare with stopping our current form of capitalism through legislation and by the institutions that protect capitalism. Because without such a drastic approach, how else would you propose stopping the biggest companies, and military contractors, from the further development of AI, 'Because the upsides are so tremendous'?


 

Labels:

Comments

Post a Comment

Popular posts from this blog

The AI Dilemma and "Gollem-Class" AIs

From the Center for Humane Technology Tristan Harris and Aza Raskin discuss how existing A.I. capabilities already pose catastrophic risks to a functional society, how A.I. companies are caught in a race to deploy as quickly as possible without adequate safety measures, and what it would mean to upgrade our institutions to a post-A.I. world. This presentation is from a private gathering in San Francisco on March 9th with leading technologists and decision-makers with the ability to influence the future of large-language model A.I.s. This presentation was given before the launch of GPT-4. One of the more astute critics of the tech industry, Tristan Harris, who has recently given stark evidence to Congress. It is worth watching both of these videos, as the Congress address gives a context of PR industry and it's regular abuses. "If we understand the mechanisms and motives of the group mind, it is now possible to control and regiment the masses according to our will without their
Post a Comment
Read more

Beware the Orca, the challenge to ChatGPT and Palm2 is here

  So Google's 'we have no moat' paper was correct. If you train an LLM wisely then it's cost effective and cheap to produce a small LLM that is able to compete or even beat established, costly LLMs, as Microsoft has just found. It's another excellent video from AI Explained, who goes through some of the training procedures, which I won't get into here. Orca, is a model that learns from large foundation models (LFMs) like GPT-4 and ChatGPT by imitating their reasoning process. Orca uses rich signals such as explanations and complex instructions to improve its performance on various tasks. Orca outperforms other instruction-tuned models and achieves similar results to ChatGPT on zero-shot reasoning benchmarks and professional and academic exams. The paper suggests that learning from explanations is a promising way to enhance model skills. Smaller models are often overestimated in their abilities compared to LFMs, and need more rigorous evaluation methods. Explana
Post a Comment
Read more

What is happening inside of the black box?

  Neel Nanda is involved in Mechanistic Interpretability research at DeepMind, formerly of AnthropicAI, what's fascinating about the research conducted by Nanda is he gets to peer into the Black Box to figure out how different types of AI models work. Anyone concerned with AI should understand how important this is. In this video Nanda discusses some of his findings, including 'induction heads', which turn out to have some vital properties.  Induction heads are a type of attention head that allows a language model to learn long-range dependencies in text. They do this by using a simple algorithm to complete token sequences like [A][B] ... [A] -> [B]. For example, if a model is given the sequence "The cat sat on the mat," it can use induction heads to predict that the word "mat" will be followed by the word "the". Induction heads were first discovered in 2022 by a team of researchers at OpenAI. They found that induction heads were present in
Post a Comment
Read more