I want to pivot a little bit away from the scenarios specifically and talk about If one of the more optimistic scenarios occurs in which AI is automating lots and lots of activities that humans previously did your paper, and I think maybe even a sort of consensus, is that that would create abundance broadly speaking, right?
The first part of that schedule, I think, is AI automating software, right? AI automating software engineering. And we saw back in November, I think, when Claude Opus 4.5 was released, Anthropic took this model, and whenever they're trying to hire a software engineer,
Okay, here's the growth rates over time. So automating 75% of tasks can raise wages.
This is an important point. We've been automating production for 200 years. The Industrial Revolution, right?
So it was clear that there was problems with the existing financial system. I was excited about automating different kinds of things using software. I had my own cryptocurrency portfolio.
and automation can amplify it and culture can adapt it? If you're automating, let's say, a chatbot for a bank, you know that a lot of your interactions are going to be about how to change your password,
And at the lower level, we're tending to create jobs in a lot of societies that almost don't need to exist. We could be automating a lot of work a lot faster, but the societal limits to it, our fear of creating-- the Oxford Martin Institute says 47% of all jobs are automatable within 30 years-- our fear of doing it means we don't do it, and we instead create lots of person-to-person jobs
For example, molecular cooking like this, where people experiment and broaden what cooking is, and make it more complicated, expands what we can do with cooking, as opposed to automating. So we're not saying there's something wrong with technology but we're saying, sometimes it's a good thing to understand what your core idea about the human behavior
It essentially gets, although the way--. He's working on automating it a little bit, so the lens, for example, is laser cut. So, it goes through and cuts all the lenses.
So I think it's a yes, and, a both, and scenario. We do need to think about automating the menial, meaningless functions to free ourselves up to think about higher order things. But we also need to think about, what's working?
Says she's going to do it. into our real lives and basically automating everything that we can.
For example, Boeing is asking us to add new SMV, which is one of the fastest and most efficient model checker out there, to this list. It could be that you're automating a function.
we can imagine that Mickey does this by writing a computer program and then having the BM execute that computer program uh and is thereby automating his chores very very clever uh so what might this program look like well Mickey might start by writing down some sort of scoring function or objective function uh where in this case the broom gets one point if the cauldron's full and zero
I doubt very many of us would have ever heard of the field today. The field would have just motored along automating various tasks while we marvelled at the cleverness not of what the creations were, but of the engineers.
humans tend to be complements for the automated power systems. It's not so clear that automating cognitive tasks and augmenting cognitive tasks is going to be complementary, versus a substitute, for human labor. And as we work through the implications of the second machine age, we need to think about how we can make the kinds of adjustments
Williams College, up the road from Lake Geneva in a town called Williams Bay. So just something as simple as automating the dice-rolling, having a pre-programmed button that rolls 15 D6, really keeps the narrative going.
Okay? And so this sort of makes you realize automating one thing really well- Is not enough. What you need to do is you need to keep automating the weak links. Right, we need a dynamic model, not just this static model, And that's what the model does.
Textile looms, tractors and railroads, right? Computers. We've been automating for a long time, and yet 2% growth Right? But if you continue doing that, something may change.
to allow the user to innovate easily on the thing that the user knows a great deal about and wants to innovate on and automating the rest so there are other things lots of other things I mean product innovator gender is strongly affected by the suitability of toolkits here's a amazing example uh
You want to find out, OK, some alarm went off, what distinguishes the things that are firing alerts versus things that are not? Here, we're just automating that process. On the ML side of the house, we're applying unsupervised density estimation here in a univariate setting.
And we start to get new technologies for automating jobs.
We have to decide that which is capitalist and that which is not capitalist. I see it as a one-off subsidy for rapidly automating the world.
and later, technologies like the internal combustion engine. The second machine age is much more about automating cognitive tasks, mental work, and we think it's going to have equally profound implications. While, when you automate make muscle tasks, it actually increases the demand for control systems, or cognitive tasks, and in that sense,
Okay, so it just continues. You know, we've been automating the economy. This is an important point.
Okay, here's the growth rates over time. The Waymo, the self-driving cars are really automating everything that an Uber driver does.
and automation can amplify it and culture can adapt it? So we want to make sure that we're finding meaningful patterns and automating those.
intelligence as a natural continuation of the long-standing historical process of automating tasks that require human
What I can say is that most of these tools help us in automating some of the low-level infrastructure provisioning and resource allocation.
You can argue that the invention of the wheel was an example of automating something, and I don't think any of us regrets that.
Because all the other weak links are bottlenecking us. Okay? And so this sort of makes you realize automating one thing really well- Is not enough. What you need to do is you need to keep automating the weak links.
On the other hand, based on the business as usual scenario, we've got weak links, right? Weak links tell you, okay, automating some stuff but not the rest, the chain is still weak because of the weak links. Okay, so what if you put both these ingredients in a model, calibrate it to what we've observed in the past, and run it forward?
So, right now, they're focusing on automating coding. They're taking their AIs, they're making them bigger, they're
about that more. And they also design their tools to be labor automating so that when the technologies are deployed,
You look at accountancy tools and apps, just reducing the need for accountants and automating the whole process.
One example in B2B industry is robotics process automation, where AI can automating large amount of repetitive works.
Or I can talk about living a rich life, automating your money so you can get on with your life because no one wants to be a money expert or a negotiation
That these things are coming, and we're kind of marching along the schedule that they laid out for us 10 years ago. The first part of that schedule, I think, is AI automating software, right? AI automating software engineering.
mentioning, you know, uh recursive self-improvement, AIs automating the whole research process, superintelligence.
are there significant behaviors going on in the stream? So to give an example of this type of functionality we're supporting in Macrobase, what we've enabled is basically automating a process that a lot of people do today when something goes wrong.
We were talking to Vinod Khosla yesterday, and he brought up that one company that he's aware of is trying to just think about automating driving trucks
It's also the fastest growing slice of the pie, so its-- you know, as we hear about computing work automating other jobs, this is the one part of the job pie
And they have this phrase, each time you repeat a task, take one step toward automating it.
But the solution, and what these people were doing, these successful people, is they were automating those decisions.
When you talk about income inequality, there's a lot of discussion about, will digital technology really end up automating so many jobs,
Male #10: Well, just in general. Like, I realize your books are about the risks of automating technology, but what do you think about the risks of trusting humans?
Daniel: I'm much more comfortable with the idea of algorithms informing humans. 'Cause here's the thing. Automating things, again, things change. Conditions change. And anybody
And I, I call that, in a certain sense, automating communication in a way that it has never been automated before. And I, I assert in the book and, and people in here know a lot more than
is HDR qualitatively different from that or, eh, is it just a way of automating it, eh