{"id":67921,"date":"2024-05-21T17:02:14","date_gmt":"2024-05-21T17:02:14","guid":{"rendered":"https:\/\/www.noemamag.com"},"modified":"2024-05-30T15:12:06","modified_gmt":"2024-05-30T15:12:06","slug":"mapping-ais-rapid-advance","status":"publish","type":"wpm-article","link":"https:\/\/www.noemamag.com\/mapping-ais-rapid-advance","title":{"rendered":"Mapping AI\u2019s Rapid Advance"},"content":{"rendered":"

Nathan Gardels:<\/strong> Generative AI is exponentially climbing the capability ladder<\/a>. Where are we now? Where is it going? How fast is it going? When do you stop it, and how? <\/p>

Eric Schmidt:<\/strong> <\/strong>The key thing that’s going on now is we’re moving very quickly through the capability ladder steps. There are roughly three things going on now that are going to profoundly change the world very quickly. And when I say very quickly, the cycle is roughly a new model every year to 18 months. So, let’s say in three or four years. <\/p>

The first pertains to the question of the \u201ccontext window<\/a>.\u201d For non-technical people, the context window is the prompt that you ask. That context window can have a million<\/a> words in it. And this year, people are inventing a context window that is infinitely<\/a> long. This is very important because it means that you can take the answer from the system and feed it back in and ask it another question.<\/p>

Say I want a recipe to make a drug. I ask, \u201cWhat’s the first step?\u201d and it says, \u201cBuy these materials.\u201d So, then you say, \u201cOK, I bought these materials. Now, what’s my next step?\u201d And then it says, \u201cBuy a mixing pan.\u201d And then the next step is \u201cHow long do I mix it for?\u201d<\/p>

That’s called chain of thought reasoning. And it generalizes<\/a> really well. In five years, we should be able to produce 1,000-step recipes to solve really important problems in medicine and material science or climate change. <\/p>

The second thing going on presently is enhanced agency<\/a>. An agent can be understood as a large language model that can learn something new. An example would be that an agent can read all of chemistry, learn something about it, have a bunch of hypotheses about the chemistry, run some tests in a lab and then add that knowledge to what it knows. <\/p>

These agents are going to be really powerful, and it’s reasonable to expect that there will be millions of them out there.\u00a0So, there will be lots and lots of agents running around and available to you.\u00a0<\/p>

The third development already beginning to happen, which to me is the most profound, is called \u201ctext to action.\u201d You might say to AI, \u201cWrite me a piece of software to do X\u201d and it does. You just say it and it transpires.\u00a0Can you imagine having programmers that actually do what you say you want? And they do it 24 hours a day? These systems are good at writing code, such as languages like Python.\u00a0<\/p>

Put all that together, and you’ve got, (a) an infinite context window<\/a>, (b) chain of thought reasoning in agents<\/a> and then (c) the text-to-action capacity<\/a> for programming. <\/p>

What happens then poses a lot of issues. Here we get into the questions raised by science fiction. What I\u2019ve described is what is happening already. But at some point, these systems will get powerful enough that the agents will start to work together.\u00a0So your agent, my agent, her agent and his agent will all combine to solve a new problem.\u00a0<\/p>

Some believe that these agents will develop their own language to communicate<\/a> with each other.\u00a0And that’s the point when we won\u2019t understand what the models are doing. What should we do? Pull the plug? Literally unplug the computer? It will really be a problem when agents start to communicate and do things in ways that we as humans do not understand. That’s the limit, in my view.<\/p>

\n