“The computers are now doing self-improvement. They’re learning how to plan and they don’t have to listen to us anymore. We call that super intelligence or ASI artificial super intelligence. And this is the theory that there will be computers that are smarter than the sum of humans.” — Eric Schmidt
“What we are finding is we are not far from a world I think we’ll be there in 3 to 6 months where AI is writing 90% of the code and then in 12 months we may be in a world where AI is writing essentially all of the code.” — Dario Amodai
“I think maybe in the next 10-15 years we can actually have a real crack at solving all disease… Science at digital speed.” — Demis Hassabis
because remember the computers are now doing self-improvement They’re learning how to plan and they don’t have to listen to us anymore That’s why it’s underhyped People do not understand what happens when you have intelligence at this level which is largely free So that was the former Google CEO Dr Eric Schmidt at the special competitive studies project having a conversation with Jan Masserv about the future of AI and biotechnology And honestly it’s a really insightful conversation Let’s actually dive into exactly what he says because the implications are profound because remember the computers are now doing self-improvement They’re learning how to plan and they don’t have to listen to us anymore We call that super intelligence or ASI artificial super intelligence And this is the theory that there will be computers that are smarter than the sum of humans The San Francisco con consensus is this occurs within six years just based on scaling Now in order to pull this off you have to have an enormous amount of power I was here yesterday testifying about this you know and we need like I can talk at some length about how many gigawatts and how many nuclear power plants and all the kind of stuff we can talk about separately This path is not understood in our society There’s no language for what happens with the arrival of this I wrote a book on this with Henry Kissinger called Genesis which you know I recommend obviously um because I wrote it available available available in your usual places um but the important point is this is happening faster than our human that our society our democracy our laws will address and there’s lots of implications that’s why it’s underhyped people do not understand what happens when you have intelligence at this level which is largely free one of the things we can that he talks about here is the fact that this technology is actually underhyped I know that many people currently do think that AI is overhyped and on one spectrum there are people who currently believe the AI is underhyped I think the reasoning here is the fact that if ASI is even remotely true or if the implications of AGI are even remotely true the implications are so profound that you might as well take it seriously I mean if there was like a 50% chance there’s going to be some super intelligent aliens that can pretty much do anything within the next 10 years it might be worthwhile to actually take that risk seriously And I think this is what he’s getting at Like it’s quite underhyped You actually factor in the fact that if this technology does suddenly emerge the societal implications are going to be so profound that the world will literally never be the same And that is the key thing here that most people don’t really you know have an agreement on Some people think okay AI is overhyped And I do agree in some aspects Certainly yes AI is overhyped But the implications of super intelligence most certainly are not This is where we get to the potentially the meat of the video And this is where Eric Smith said a statement that I think is starting to come to fruition just a little bit more every single day He basically says that eventually all of the code is going to be written by AI And right now I do think that you know we’re on a high level of abstraction when it comes to writing code with AI in the sense that you know maybe we are the implementers but AI is the one that is definitely writing the majority of the code right now And it’s super interesting to hear his opinion because he’s not the only one that states that year the vast majority of programmers will be replaced by AI programmers We also believe that within one year you will have graduate level mathematicians that are at the tippy top of graduate math programs There’s lots of reasons to think this is going to happen This is the consensus You go “Okay well that’s pretty interesting Now I can’t do that kind of math Very few people can do that math How can the computer do that math better than anybody else?” To some degree it’s because math has a simpler language than human language So the way these algorithms actually work is they’re doing essentially word prediction So you take you take a a sentence you take a word out and then it learns how to put the correct word back in This is called the loss function and it’s optimized to do that at a scale that’s unimaginable to us as humans So you do the same thing for math but there you use a conjecture and then a proof format through a protocol called lean In programming it’s pretty simple You just keep writing code until you pass the programming test So strangely the first question I always ask programmers is what language do you program in and the correct answer is it doesn’t matter because you’re trying to design for an outcome You don’t care what code is generated by the computer This is a whole new world Okay And so yes a whole new world Now like I said before it’s not just Eric Schmidt Take a look what Dario Amade said recently in another interview on a panel And he basically says that you know 12 months from now which is a year from now AI could potentially be writing all of the code And you have to understand if AI moves as quick as it does 12 months from now there’s going to be a lot of change Powerful tool If I look at coding programming which is one area where AI is making the most progress um what we are finding is we are not far from a world I think we’ll be there in 3 to 6 months where AI is writing 90% of the code and then in 12 months we may be in a world where AI is writing essentially all of the code Now here’s where we get into something even more interesting So basically Eric Smith talks about the fact that all right AI is now writing all of the code but what happens after that in year two take a look So that’s one year Okay What happens in two years well I’ve just told you about reasoning and I’ve told you about programming and I’ve told you about math Programming plus math are the basis of sort of our whole digital world So the evidence and the claims from the research groups in open AI and and anthropic and so forth is that they’re now somewhere around 10 or 20% of the code that they’re developing in their research programs is being generated by the computer That’s called recursive self-improvement is the technical term So what happens when this thing starts to scale well a lot Okay And where do you guys think recursive self-improvement leads us to and this is what gets us to the AGI/ super intelligence talk And this is where things start to get pretty crazy because we all know what happens when AGI is achieved Society is most certainly transformed One way to say this is that within three to five years we’ll have what is called general intelligence AGI which can be defined as a system that is as smart as the smartest mathematician physicist you know artist writer thinker politician maybe not in the same level um but you get the idea Uh just the creative industries and so forth But imagine that in one computer Okay Well that’s pretty interesting I call this by the way the San Francisco consensus because everyone who believes this is in San Francisco It may be the water What happens when every single one of us has the equivalent of the smartest human on every problem in our pocket so it means you have the best architect when you have an architecture problem Okay And another thing that he speaks about here is agents Now agents are growing every single day There are new frameworks new you know protocols that you use OpenAI recently released uh GPT 4.1 which is agent focused in terms of how the model is designed Google recently released A2A which is another agent framework Anthropic recently released Claude MCP I mean there is a lot of stuff going on just making agents a lot better and he basically talks about this as a future and I’m going to show you guys why this is super important Another thing that’s going on is the development of agentic solutions and agents are refer to systems that have input and output in memory and they learn An example here is that I want to uh buy another house Uh I happen to like Virginia I grew up in Virginia I say “Find me a house in the greater MLAN area Look at the That’s one agent Look at all the rules Figure out how big a house I can build.” That’s another agent Do the transaction to buy the land That’s another agent Design the house with a human architect right but sort of ignore them for most of the thing but they have to sign it off and then I approve it and then find the contractor right hire the contractor pay the bills and at the end sue the contractor for lack of performance Okay now I just gave you the stupidest possible explanation I just described every business process every government process and every and every sort of academic process in our nation So it isn’t just the programmers that are going to be out of work We’re all going to be out of work No that’s not a consequence I’ll come to that But but the reason I want to I want to make the point here is that in the next year or two this foundation is being locked in and it’s not we’re not going to stop it Now of course with the rise of agents many people are wondering whether or not automation is going to replace everyone And I got to be honest I’m 50/50 on this because I do think that this time is different Whilst Eric Schmidt basically says that you know every single time has been pretty much the same we just find new things to do I do kind of struggle to see how humans find their role in the world But of course I’d love to know your opinions and Eric Smith basically says that you know it’s basically going to be fine as it has been in the past And by the way on the jobs thing everyone assumes that automation will replace will eliminate jobs If you look at the history of automation ever since the the looms and uh in uh 300 years ago the jobs are changed but more jobs are created than destroyed In this case you’d have to convince me that this time is different If you look in Asia where they for whatever reason are choosing not to have children the Asian reproduction rate is in the order of 1.0 or lower So they’re rapidly disappearing So the Asian countries are very very quickly automating The tools that I’m describing will allow the few humans that will be working very hard in 30 or 40 years If these trends continue the rest of us will be dependent on those hardworking humans it’ll make their productivity more much greater And of course another thing to mention is the fact that the deepseek moment occurred and he talks about you know to to quite some extent here about the fact that China are playing no game when it comes to artificial intelligence They are taking this so seriously more seriously than they’ve taken anything before So this is something that of course the United States needs to take seriously as well In China the deepseek moment is equivalent to our chat moment I was there with Henry Um and this is what happens when you’re talking to to the Chinese about AI with Henry And this means we are alive and we’re listening to you Thank you very much Right that’s not what they’re doing anymore When the when DeepS seek showed up and our stock market lost a trillion dollars in one day all of a sudden they began to understand the scale of what it was So now there is a massive program in China to accelerate these things I had thought I and some of the other people in this room worked really hard on these um chip controls and the chip controls have been um in my view largely effective How did China get around them well some of it was straightforward theft and evasion of the tariffs but they also they’re sufficiently smart They created new algorithms that use different kinds of computing to move forward because they because China operates in open source that is they they release the software to everyone There are two things that happen We we Americans immediately saw their idea and incorporated in our own So thank you very much China You invented something new immediately incorporated it But second because it’s free the proliferation issues around Chinese models have now become a very big deal And our government is trying to figure out uh without success so far how to handle this question It’s a very tricky question And one of the ways that the US government has actually tried to play this is that they’ve actually tried to restrict American access to deepseek They are thinking about you know an app store ban They are thinking about limits on how cloud providers could offer deepseek models There are just a vast amount of ways that the US government is going to handle this and I don’t think it involves them integrating deepseek into US culture at all I mean American competitiveness is something that they really do want for the country and I don’t think they want Chinese models proliferating the ecosystem Now remember how I said that China are taking this super seriously What happens when China gets ahead of us or we are so far ahead of China that they start to be concerned that is when we have to start looking at things like data center attacks data centers being bombed and of course mutually assured destruction So basically this is where Eric Schmidt talks about the fact that if one country is you know so far ahead of another they might try to undermine that country by you know stealing the weights doing a lot of different things and this has been spoken about extensively by those who try to predict the future on AI having this whole debate in our nation about what to do about Iran’s nuclear program and I’m not an expert in that but these are the kind of conversations that happen here in in DC So when we get to the point where China is n months ahead are we willing to bomb their data centers my favorite example here is I was in I’ve been working on this I was talking to somebody said the answer is obvious I said what the good lady and the bad guy We agree to a treaty where each of us puts dynamite on each other’s uh electricity supply You get to blow up my electricity if you get mad and I get to blow up your electricity if I get you get the idea Now some would say we’ve already done that’s already happened Well kinetic attack on people’s data centers is probably an act of war Yeah Eric Schmidt actually talks about automatic drug discovery and potentially synthesis which is a remarkable implication because it means that you know in the future finding cures to many different diseases and all these ailments is going to be simply the push of a button and just waiting for the AI to figure it out And he talks about you know how there’s literally a startup that he’s funded and how they’re actually working on this the primary funer of a particular group that has built a model It first learned how to do chemistry and uh it was trained as a foundation model for chemistry and it’s attached to a robotic lab and what this model does is it generates hypothesis for drugs of one kind or another and it just generates them God knows if they’re right and then overnight the robotic lab tests them and gives the report overnight and then it starts again and the reason I’m mentioning this is this is the future model of the fusion of AI and bio right the AI system generates all sorts of candidates to reduce the um essentially the um search space if you think about it algorithmically it’s an exponential with too many degrees of exponential So you have to come up with some way of reducing the space So this particular group is using AI to reduce the space run the things and so forth Their objective we’ll see if they pull it off This is research project is to identify all human druggable targets within the next two years If that occurs then that information goes straight into the drug industry Now it’s a different way of thinking and it’s profound in that it gives them the targets they need to go build drugs against That’s interesting to me It’s the combi combination of AI and a robotic lab that does something in a wet lab essentially So one model that you should think about is wet labs will be roboticized And the wet labs will have AR they’re essentially they’re not humanoid robots They’re arm robots and they go boom boom boom They and they do the pipeetting and so forth and so on and they do it 24 hours a day That’s a major change in the way bio bio the biotech industry works And I mean I don’t have to you know really even get into the implications of that I mean if you could simply push a button and the AI could immediately begin working on some of our most common problems I mean the life expectancy continuing to increase is something that has been predicted for quite some time now especially with the advancement of AI technologies and of course AI if this works if it really does work we could actually be on the cusp of longevity escape velocity so that is basically where you know we could potentially live forever and I know that sounds completely crazy but in theory it could work and then here’s where he actually responds to someone in the audience that asks if that you know this is a real possibility a recent PhD in biomeical engineering Very excited I’ve been following you Just a very quick question Do you think there’s implications for ASI via drug discovery for like curing cancer and or personalized medicine just something um yes because under the under the assumptions of super intelligence these are systems that see things that we don’t see And so the assumption is that ASI for example could understand biological and cellular mechanisms that you are an expert in and I’m not at a level that humans will not So that’s why this is such a big deal We’ve always assumed that humans would know there would be at least one human right we call these people polymaths that would understand these things We’re going to end up in a world maybe 10 years from now where we won’t actually understand why But you as our scientist will say I use it every day when I when I was at college I was studying quantum physics and my friend who was a graduate student who is much better than I and I said is this stuff actually true you know it’s like too weird to be true and he said yes we use it every day and I imagine in 10 years some young student will come up to you and say is this stuff true and you’ll say frankly I use it every day no human understands What an interesting situation for you as now a senior researcher 10 years from now to have to deal with Wasn’t the only person who actually shares this opinion Demesis Haris recently in an interview around 2 days ago literally said the exact same thing and they’ve been working on this for quite some time now So it isn’t something that’s just come about because you know AI and LLMs are now popular But this has been something that’s been in the works for a really long time and we’re only starting to see small signs that this could really be something in the future now Maybe not the immediate future but definitely sometime soon I think maybe in the next 10-15 years we can actually have a real crack at solving all disease That’s the mission of isomorphic And I think with Alphafold we showed what the potential was um to sort of do what I like to call science at digital speed And why couldn’t that also be applied to finding medicines um and so my hope is 1015 years time we’ll we’ll look back on the medicine we have today a bit like how we look back on medieval times and how we used to do medicine then you know and and that would be I think the most incredible uh benefit we could imagine from AI