point you made very early is that energy transformed into work has substituted for labor and so we trade energy for labor and now we're trading energy for intelligence and intelligence itself is a labor multiplier so yeah exactly so the question is well is trading energy for intelligence a good trade it's like well that's that's what we do that's what human being is buy buying intelligence really cheaply a good thing right right good thing for every individual who buys it so in the aggregate it's probably going to be pretty good yes yes well and it's what
you want to do when you hire someone to do a complex job you're gonna yeah and people are just like it's so just is one final digression people are just so they just don't yet realize how much how efficient it's going to be to use these things for not every application by any means but for many applications they have all these things like oh it's not always it doesn't always get directions proper oh really all your human employees always get the directions correctly like at this point Chach BG Pro is better than almost any human
you will ever employ in terms of in terms of following directions like it is really really good in terms of just if you write out what you think if you say it in a secutus way it is pretty damn good and in terms of like the output this is another thing I think people don't get is they like for instance I'll give you an example I was running this CO2 thing like this natural gas CO2 thing and I ran it with the non-pro version of chat GPT and it gave me a false thing that was
something like I wanted to know how much volume of natural how much uh of how does the volume of natural gas compared to the volume of CO2 generated because one way to think about this idea of let's capture let's use fossil fuels and capture the CO2 is you need to build a whole new industry for the captured CO2 so if it's a one: one ratio right if it's a one:1 ratio of uh natural gas and CO2 then you need to basically double the size of the indust right in terms of piping it and putting it
ground and I thought it was a one to one and then I was in the airport and talking to chat gbt plus and it told me it was like 59 times I'm like that sounds wrong but if that's true that's that's crazy and I was sort of excited like oh wow that's a blackbuster but then I was like I ran it the other day with Chachi B pro and I said no no it's it's one to one which made more sense to me but it's like okay yes sometimes it'll make a mistake like that but
I don't go to print with that thing without confirming it by an expert but often but even if it's right 95% of the time or 98% of the time you can often get I don't have the perfect words for this but you can often get like the shape of what the solution will look like even if not every value is correct so you can get the idea of yeah you do need to build a whole new parallel industry and it would cost a lot of money even if it's your estimate is 50 times to high
you can help think through it so just the fact that it can make errors aumans can make errors too but it can really help you explore territory and develop ideas very very quickly and then what you might find is that you make one or two errors at the end but usually those are not decisive errors that throw out the whole project usually because every kind of thought process involves lots and lots of little uh assumptions and and interrelationships and if you get 98% of that right probably the overall thing is going to be right probably
a very small percentage of the time it's going to be oh there was one false assumption and the whole thing is wrong I have yet to find that it could happen it could happen but so yeah I think people are just underestimating how good this stuff is going to be although not the people who are demanding more energy so that they can make faster artificial intelligence systems they're not underestimating it and that's why they driven to do and they may they may be even overestimating certain things and it could be we're back in five years
and yeah there might even be a bubble in certain ways but at this point it's obvious that for a lot of people it's going to help them and I would just say as an entrepreneur who runs a small business that's very intelligence-based it obviously helps me think much much better it helped me it helped me a lot when I was writing my last book I used I used the AI systems a lot yeah yeah yeah to sketch out research domains you have to check the references you have to make sure it's not lying to you
but you can do that you have to be careful with it and you have to you know interact with it intelligently but yeah and we've built specialized AI systems too that um some of them based on my work that I consult because well that's an extension of my thinking and that's been extremely helpful and so oh yeah I should mention Alex Epson is now free so people should check that out oh yeah okay so that's that's now the latest version of energy talking points is we're already having some elected officials use this for like floor
speeches is you can have Alex AI write you a floor speech and and it's so we've spent a lot of money customizing an AI and particular sein yeah a and and what it's really done is it's engineered with prompting that that is very based on this even-handed and pro-human uh right you built that ethos into it yes and and really one thing it does really well is question assumptions so it scans everything for is does this questioner statement have an assumption that Alex Epstein would disagree with so for example if you say hey Alex AI
how do you how do we get to Net Zero by 2050 as quickly as possible it doesn't just try to manufacture answer it says would Alex agree that that's the right goal well actually I disagree with I don't think this is the right goal I don't question the premise yes exactly funny yeah but that's one thing we had to that means it accurately reflects you yes yes exactly it's just as annoying in some circum all right let's go back to these Five Points so in terms of liberating so let's make sure a couple at least
one of the big things for this um so I would say with the liberating domestic development one of the key things we need to do is address what's called NEPA I don't know if you've heard this term you might have heard it's National Environmental Policy Act so this is one of the early environmental laws and NEPA is the thing and and I forget what the Canadian equivalent is I think you just passed a new version of this that's nuts but it's basically it's a duplicative review process that's what NEPA is so it basically says like
any agency that does anything it has to also go through an additional review for its quote unquote environmental impacts or impacts on the human environment I mean it's it's worded something like any significant impact on the human environment and significant is not defined human environment isn't defined so what it means is basically and then it has to do with Federal actions so it's like a major federal action but what is a federal action is it anything where federal law applies is it so originally it was supposed to be okay the federal government is building like
a giant Bridge you know that's a mile long or something like that is it going to cause any kind of major damage or something like that and you write like a 10-page report now it's become every project imaginable is covered and it can take 10 years and one of the major mechanisms is judges can sue so NEPA has no official authority to stop anything it's just a review thing but you people can sue activist groups can and they can they can say you didn't you left out this bird on your review so you throw it
back and then the judge says yeah you have to this so in practice this is the thing which why we can't build any roads why everything takes forever why a m you know takes something like 15 years to per if you put your Five Points here yeah into one of these AI systems yeah could you ask it for example if we were moving in this conceptual Direction what policy changes should be implemented prioritized by their benefit to cost I'll have to look now when I tried it a year ago as horrible it was but but
it didn't have search back then it didn't have a sophisticated processing so with the the search it would be it's it's a good thing I'll do it right after this and see because it's a predo distribution issue right there's going to be a couple of things you could change they're going to have a disproportionately positive effect and everything else is going to sort now now with the interestingly with this kind of thing it's often very efficient to just talk to a very smart lawyer so like lawyers that we pay like at the top end it's
like like over $1,800 an hour like we'll pay lawyers and and it's worth it cuz you can just ask them like an expert in Nea expert in electricity like what really needs to change because you often find that the thing people talk about isn't the thing really matters pointa some of the things are like one of the big issues is um is like one of the big kinds of solutions is you can what's called limit NEPA to agency discretion so basically the agency can do the NEPA but it basically decides okay we've done the review
and it can't be challenged like something like that and and it's fine because it's it's the agency's responsibility to review the thing you don't need to put everything in Double Jeopardy and just take forever and have outside people allowed to question your review and basically if it's that just makes it impossible yeah that's that's what it is so that's this kind of thing where when politicians will talk about NEPA they'll often say something like let's limit the length of the process right let's L limit an environmental impact statement to two years or one year or
what's called an environmental assessment to a smaller amount of time but they don't fully get that if you you can still have infinite litigation on that so even if you set a shot clock if you have infinite ability to challenge it so that's the kind of example where and there's like specific that's a fix that won't work yeah [Music]
