Interview with Matt Hsu - How Cerebras is Taking on Nvidia in the ML Arena

all right hello everybody uh and welcome today this is a very special episode uh because it's not just you know you listen to me and shashan rattle on back and forth about nonsense uh today uh we have a guest um so uh we have uh actually I would even call him like a friend as well right at this point I think yeah yeah I think so you've done that for for a while yeah for for uh quite a while so uh Matt is a regular attendee at the South a gen Meetup and always has is always one of the most interesting people I would say that I get to talk to on uh like whenever I can it's always a pleasure uh when I'm able to talk to like Matt is just like a wealth of knowledge about a bunch of stuff you know whenever I talk about something he tends to know a lot about it he's the type of guy who knows a lot about a lot um so I think that you guys are gonna enjoy the conversation that we're gonna have today so Matt he currently works at uh cerebrus um which is a company which is in a certain sense competing with Nvidia at like the highest levels so um they make the largest Computing I don't know what would you call that like the largest platform yeah it's like the largest yeah the it's like probably one of the most powerful Hardware computer uh Platforms in the world and um I'm just like super honored that like you were able to take the time and uh chat with us today uh so anyways Matt why is uh cerebrus better than Nvidia Mark thanks for those kind words um yeah I could ra ramble on about a long time but but the gist of it is that um Nvidia is a GPU and it was designed to move textures from memory from a centralized memory through the graphics processor and process triangles and pixels and whatnot and um because of that it it it has a it doesn't have much knowledge of the content of the data and so it does have however have a lot of processors in it and back in the day if you wanted you know tens of thousands of processors for 400 bucks you could buy a video card and once once Nvidia opened up Cuda um uh and and Google got a hold of it and made tensor flow and the ml guys realized that they could get access to all the micr code and all those cores they were they were a at the races and um the only game in town actually AMD was really late in releasing any kind of API for their for their GP um and so Andia was the only game in town and that but it it it was primarily designed to move textures and because of that when you want to do a workload you you know ml is often uh moving a n bym rectangle of data from memory into the processors and oftentimes back to memory um and what what you do what Nvidia doesn't know how to do is if in ml workloads after the very first layer because of the nonlinear activation function you often have a lot of zeros and so there's this thing called sparsity where a good portion of your data um you know from 20 30 40 even up to 50 60 70% is going to be zeros and so Nvidia doesn't really know how to move not move zeros from memory into the processor so they've got to move that n by rectangle of which a good portion may be zeros so that takes both time and energy to do and and when you do that and you do you spread it a bunch of all your processors um you you've also got to do multiplies and ads and they don't know how to not multiply and add zeros and everybody from grade school math knows that you know adding and multiplying zeros is a no op and so they they use they use bandwidth and processing power to to to move those nonzeros now to move those zeros now that's that's that's that's one thing right and that's one one advantage um the other one is uh we we pioneered um this technology called wafer scale and it's been tried in the past unsuccessfully and what wafer scale is is rather than cut the retical off the wafer to make individual chips and put them in packages we leave them all on the wafer and cut the routed parts off to make a giant square of a chip so so it becomes the biggest chip in the world it's roughly 12 in on a diagonal um by a long shot it's got you know many many dozens of rectangles in it or reticles in it it's got uh um trillions of transistors it's got close to a million processing cores in it just in the same just in the same chip and by doing that when you move data from one chip to the next it's just a tiny little Trace that is extraordinarily fast way faster than paying the uh Time Savings of charging up capacitance to get off of a single die through a package and then a board and getting through the inductive and capacitance of the chip on the that it's talking to which is which is slower and takes more power and so we got a couple of things going for us there's the the the scale of the of the chip itself that saves power and time there's the groundup design for ML workloads um we don't have any shared memory so memory is is local so there's no moving zeros in and out of memory um the groundup design is data flow so whenever there's a calculation that results in a zero it just gets dropped and and the next nonzero data gets moved in its place um one of the problems with this though is that all the benchmarks are designed for dense data workloads uh and so we don't do benchmarks because they're all you know sparcity is is where we get hum significant gains in our in our performance and so we we tend to talk more about um about how how fast our customers can get an answer from the product and so those are just two examples of why we think we have a a better solution than than than our competitor okay so uh I think there's a lot of things that you said there so I'll just try to um you know DB it down for at least myself uh sure and like you know for the the listeners so I guess what you're trying to say is you know cerebrus is better because it's specialized for the particular work of training llms would you see that's like Fair I'd say training any you know we're not specific to llms we we're a general purpose ml training LMS just happen to be the the rage right now but before that it was uh computer vision and and and image recognition and things like that that's where the first real progress got made and then and then people realized and I think our our our management realized that hey well that's not what what's really going to happen is LS are going to be there and so a couple years ago we started investing really heavily and making sure we could run L more clothes and that that proved to be a pretty impression thought okay I guess uh do you have a sense for what kinds of models have more sparse um weights as opposed to dense it sort of doesn't matter actually the ml the activation function like there's an activation function called relu and what it means is any value that result that is uh less than zero becomes zero and any value that's greater than zero just stays itself so it's kind of like and the and the nonlinear part is right at zero you got this sort of steep angle there which makes it a nonlinear function and so um that means that you know after the first after the first layer of processing you could wind up with a significant number of of zero data I guess um you mentioned two things that cerebrus as much better than Nvidia one is the way for scale Computing you guys just have bigger more massive chips and the other one is dealing with sparity so isn't uh isn't that like a software or design problem that Nvidia could also handle like they could just ignore sparse values somehow with some fancy math or computer science they can and I don't want to I don't want to speak for their their marketing or anything but um my understanding is they can do it on a stride basis and if the if if they can kind of align the data uh the zeros to that stride they can skip it but they can't do it um completely asynchronous to the to the the way the data comes in at the moment so you would you would need to I mean you would need to drop the zeros you would need to have special instructions and things to know that it can drop the zeros because it you know oftentimes The Matrix operation is you know here's n by m and here's a thing you're multiplying and by and it just does all the values and fills them in and that's that's what they're really good at but um it it's not a data flow architecture it's more of a uh mimy kind of architure where you're you're processing multiple data and multiple instructions but they all have to process at the same time in order to get the data through and then they has to go back to memory and so you need special Hardware to kind of figure out where the zeros and non-zeros are but if you have data flow it just flows through this so you mentioned uh data flow versus like MD architecture can you maybe Define like what that is so in in a data flow architecture um the data just goes from one stage to the next and doesn't isn't necessarily that aware of anything else um any adjacent data or um or things like memory size or or whatever just it just moves as it moves through the pipeline whereas if you have a a a a matrix and you're trying to do an operation like a true parallel processor would it it it doesn't it just it knows that it's got n byn pieces of data move it moves it and then does its operation and then does the next operation and does the next operation and that's more of a I don't know what You' call it a matrix flow or the chunk of data is much bigger okay so in a certain sense uh to like you know kind of summarize is like uh cerebrus is better uh because or like specifically for ML workloads U because it's op the chips are optimized for it um right and like Nvidia is more of like a a general purpose chip that since they're like doing like Mass Market appeal that like those chips need to work for like gaming and like ethereum Mining and whatever um therefore like they wouldn't be able to make the optimizations that like uh cerebrus would be able to do and because of that uh cerebrus is potentially like a far better chip for training ml workloads yeah exactly and early on in the company our CEO you know he's he's a good marketing guy Andrew Filman he he he basically said you know CPUs are like the hyena is a Savannah right they'll eat they'll scavenge for Foods they'll kill their own food they'll basically can just do anything um and they you know they can they can move around they're a little mobile and then you get to gpus which are a little less like hyenas but they do General P purp workload and what we make is we make a cheetah it it just does the one thing really really fast and everybody you know has that immediate visualization of that cheetah running through the saada from whatever nature show you saw and it was a it was a beautiful kind of marketing thing because not only does it illustrate the point that we're not a general purpose thing but you know when you visualize a cheetah versus a hyena you have this positive negative I thought well that's really clever let's just cast our competitors in this this SL is big hyas and we this beautiful Majestic Savannah animal it was really hilarious so um you mentioned you guys are cheetah at training training us and uh I I guess I that made me realize that cerebrus doesn't really focus on inference as much and I I was curious can your chips even do inference well oh absolutely so so there's there's this thing coming up and and and and we knew that we needed to be in the inference game at some point but if you consider that a lot of the end result of training models you want to work on cell phones you want to work on personal computers you want to work on much much smaller things so it makes sense to train a model on a on a big Hefty compute you know resource heavy thing and then use sparsity to make that model really small and then run it on some things with maybe some a few percentage points loss and accuracy but there's this whole Market coming out now which like if you think about llms and and some of the um uh image generation networks right now they're not what I would call real time they're you know you give it some data and it comes out with an answer and you give it some data and it comes and it can take milliseconds it can take you know you can make a video that you know may take you know a minute or two or music but it's not really real time in the sense that nothing is depending on that coming out but now we have this thing which I don't know I I call it like Edge inference where you're you're doing inference that the data is required to come out really fast and and get get used and um once you fill the pipeline um um your your your inference then becomes this sort of real time model which is a little different than than what uh a lot of models have done in the past um so we we we've we actually partnered with uh I forget the name of the company recently well I I saw you partnered with Qualcomm Qualcomm right to to to help us with the inference part of it and the funny thing is you know some of these Edge inference tasks are so compute intensive that we think we're actually faster at inference for those bigger jobs and so there's there's a you you might train a model on Nvidia chips and then do inference on our box because it's throughput is faster right because it's just a big chip and everything's moving through very fast an inference I've heard takes roughly a quarter of the compute um required than the training part which has this whole backward path and weights and activations and things so would cerebrus chips be better for inference for some of the larger models sure absolutely but remember our box takes 20 kilowatts and it's you know it's a it sits in a you know it's 16 u high and sits in a 19inch rack and um you're not going to put it in a car and you're not going to yeah not going to carry it around with your anything like so it's a very big compute heavy requires water cooling for instance so you can do inference on it but you have to have a workload that you're going to want to commit that kind of resource too it'll be some of the big dogs running some of the larger models or maybe some of the Poss uh some startup that's taking maybe llama 3's bigger parameter model fine-tuning it and maybe serving that quite possibly yeah I mean they're use cases it's you know it's a generic box right you you have a model you've got weights you got activations you know how big it is you know how many layers you are you can you describe it with you know tensor flow or pie torch or something and once you've described it and you have your weights and activations they're just a matrix of data you just load it onto whatever system you want to load on to and and and you're off the races but uh maybe going back to your original Point um you don't Market the cus chip for inference you Market it for training it's the cheetah of training it's not the cheetah of inference so assuming assuming no one wants to use it for inference how big do you think the market is for just people wanting to continuously train models indefinitely because like let's say you train llama 3 for example um and then you're done and then you release it to people uh what are these chips doing in between successive model releases is there any point to having them or is it just like um a dead weight well at at the moment well I can't talk about our marketing and how they're infer so so we'll leave that one aside but as far as the the the training part you know you train a llama model and you're going to train it on data from say today 2024 right and then two years from now there's going to be you know another large increment of data out there in the universe that you're going to want to retrain the model on so it's not dated so I I have a feeling they're going to be retraining models over and over again all the time so that's one point the other point is there is a ton of untapped training markets like we our partner uh g42 um that that that's that spot the line sh our systems right now um they they just released a 30 billion parameter um Arabic model right which you know nobody else has and so um I think as far as workloads are concerned for the next decade I could imagine there there being a demand for training over and over again bigger and bigger models to get more accurate and and it's you know the the training say of an llm is um um itial with the number of parameters you have right so if you have if you want to train a 3 billion parameter model like I think gbt 4 is on that order of magnitude uh trillion trillion three trillion perimeter model um you might use x amount of Hardware well if you want to go to a Six Trillion parameter model it won't be 2x amount of Hardware it might be 9x amount of Hardware taet because it's just it's just that much more data that much more parameters um the network is it has two Dimensions rather than one so you might need significantly more more compute to do so there's this sort of like arms race going on and who who can build the biggest compute for for um for training training models yeah that's true I think that uh Sam Alman is not trying to raise you know whatever it is like9 trillion dollar or whatever that crazy number is uh just because he doesn't want to do more training right I I think that right now all these big companies like are trying to do as much training as possible actually I remember in a interview with her heard with Mark Zuckerberg he said that the biggest bottleneck right now for training these models is not actually Hardware but just like straight up energy uh so if if we're run into energy constraints uh I I think that we'll want to get as many optimizations that we can out of the hardware as possible and with Nvidia gpus or any gpus in general if uh there is issues with sparsity uh if there's issues with the chips being small um I could see uh a lot of these companies wanting to use uh cerebrus Hardware just because it's you know optimized for training ml workloads because these companies are spending billions and billions of dollars uh training so to me like I don't know why you if you're going to spend that much money I don't understand why you'd want to get like some sort of general purpose chip uh but I think that you want to get like a chip that is uh specialized for that now I guess like the the counterargument of that is like well if you spent a bunch of money on Specialized hardware for training then you could only use it for training uh and then like you couldn't use it for other things um so I guess maybe uh some companies are thinking that okay like we have this hardware and we can use it any way we want in like a general purpose fashion so it's like we can use it for training and then like I don't know but then we could use it to help run our web servers or um we could do some like Bitcoin mining with it you know whatever like you could do that I guess but still I think that there should be like a high enough critical mass of people with is like hey look if I'm going to spend like1 billion dollar on this training like you might as well get something that's specialized for the training I'm pretty sure that there will there won't be that the optimal use of the hardware will be training for the next decade yeah they'll just be adding more Hardware over time now I don't know when when successively new generations of Chip out maybe some of the old Hardware gets a little a little obsolete but a lot of models actually can work multimodal on on multiple versions of Hardware so I think that I think the training Will Will just will continue to be there and and a lot of companies have like for instance we know we we work with some Pharmaceuticals and some medical industry and they have very special privacy requirements in terms of their servers and how they need to be isolated and so they're not going to want to buy they're not going to want to put ml Hardware in the cloud they're going to want to have it isolated somewhere so that um they can use their own data and their own you know protect their their their patients data and whatnot in their own in their own premises so I suppose uh to solve some of these issues uh you guys have your own cerebrus cloud to maintain Peak utilization across all your uh chips so that someone doesn't need to buy their own chip and have it sit idle after they're finished trading uh yes and no uh the S we worked with the Comico serus for cloud service kind of like Behavior okay I don't I don't know that it's something you can just kind of like spin up an instance like AWS because you need a lot of compute for it most people that are serious about it will and want to use our system will come to us and we have different levels of service from White Glove where we we have people that handle your your model and all that kind of stuff all the way to I know what I'm doing I just want to I just want to run on a cluster of your systems that points out another Advantage I think we have that's that's like ease of use early on when you trained um and then on a bunch of uh Nvidia things you had a you could tell it how many gpus you'd want to use in would do it but it it actually takes some amount of coding to scale out to a number of gpus and um uh I think our our our CEO pointed out that you know it might take many tens of thousands of lines of code to scale out to to more gpus where for ours you you really truly just put another number in and the user doesn't know how many um how many boxes he's putting his thing on as far as sharing workloads our our biggest customer g42 is the is is kind of the largest cloud provider in the Middle East we we've got two six to4 known clusters spun up from that so it's 64 of our chips uh in in each box they they while they they use it most of the time there are periods of time where we'll let we'll have other customers that'll that'll rent time on it so I I grew up in the Middle East in Abu Dhabi and I have a rough sense for the cult culture and like the interests there um they have a ton of money they don't have as good of a talent pull as here in Silicon Valley but I mean I think most of the researchers are trying to go where they have the most compute so it seems like a great play very Forward Thinking Play and uh uh these Kings and the shakes uh and the rulers there are very aware that oil may not last them forever it's a good move on their part um I I think they were thinking about moving into clean energy at some point but um they've been trying to get into the tech game for a while yeah and I think now it really makes sense for the Middle East to try to do something big because compute is a very valuable resource so sure I just hope the US is listening and we decide that you want to we want to have our own compute here as well and not let it not let it go completely overseas well I I did watched the cerebrus AI day keynote and uh they mentioned a couple different nodes here Santa Clara San Diego and one other location I think so yeah the initial announcement with g42 included um three 64 node clusters at how how big is that is that a supercomputer is that uh maybe like just a big server how do I answer that uh it's it's the equivalent of many thousands of gpus right so so you could call that a superc computer though it it's a super computer yeah 64 so so what they said is we want well we we went into we went into g42 and they had been like you said they're not an expert in ml and so they'd been trying to get um one of their models to work on a on a on a a bunch of gpus and I'm not going to say how many because I don't remember the number exactly and they've been trying for a bunch of weeks you know several weeks to get it to run and it hadn't finished yet and so that we saw this opportunity we came in with our box or a bunch of our boxes and we we finished the the the Ron in a week and it finished and it gave them the accuracy they wanted and they said okay we want this and I don't think they're the they're the experts at ml models so someone at one of the hyperscalers like open AI or somebody could probably make a model better and more efficient and have finished but most customers are not like that most customers I've got a model I've got data can you help me run it I don't want to spend you know hire hundreds of Engineers to optimize that down to the point where I get every last ounce of they just say I just want to run you know can you can we run it somewhere and so g42 announced they're going to buy three 64 node clusters now the initial 64 node cluster the Press was quoting in about 100 million I obvious can't talk about how much they really cost but the first one was 100 million and then two more after that each hit say 100 million uh per per cluster and then right after that they said we want nine more 64 node clust or six more 64 node clusters we total nine in in the U and wow at the moment it's easier to have them here in the US because you don't have to do export stuff so we originally were going to ship all our systems over there and then there's got to be so much import export paperwork that we said well you know if we just put it in the data center here there's zero of that and access to it is the same as is immediate anyway so why not just host it here so we're hosting it here but to the point that Mark mentioned energy is a big cost for training these models and they're in the Middle East they have all the energy so I'm surprised that they didn't want to go through the paperwork to ship what was it close to a billion dollars of chips you know I don't I don't know for sure I can imagine that somewhere along the line there will be systems shipped to them and they will have a um on their premises they will have they will have noes it's a lot E I mean it's a very complex system so it's a lot easier to support when it's here if there's a um we have we have we we ship like 64 nodes but we actually ship a number of nodes as spares to see it cycle in and out if one of them shows a problem we'll we'll cycling it out and as you can imagine because we're a wafer scale system sometimes you find latent defects and we have significant amounts of redundancy and so we'll swap out that system figure out how to repair that Noe out and then it becomes the next spare there so that's all fairly seamless for them so um that was going to be one of my other questions so when when you're making um traditional chips whether CPUs gpus on the scale of like you know for a laptop mobile phones and even servers they're they're pretty tiny and you I guess like etch out different uh you take you take a die and stamp a bunch of different chips onto a large wafer and see which one of those stamps looks good and if any one of those stamped G CPU GPU die uh doesn't look good you kind of just toss it out that's right right but with a wafer scale Computing you have a massive um chip I think it's maybe 10 15 20 times the size of a regular GPU um you don't want to throw that away you can't because you would have a yield of zero we know that you know the defect density is there going to be defects it's you know back one of the technologies that I worked with was like 16 defects per so many square centimeters or something like that so there are defects silicon doesn't come out perfect so you know they're going to be there now whether they fall a transistor or trace or something that's going to kill you that's a thing but they're going to be there and they're going to wipe out certain chips so you have to have redundancy and and and other companies have been doing this forever right like memory companies in order to get high yields they have whole extra Columns of bits that they can just redundancy out right if they find that there's too many bid errors on that column they'll just they'll just take it out so they'll get a really high yield and when you apply the power to it they'll apply the redundancy uh programming and no body anybody using it is none the wiser so we have a we have a standard size that is smaller than the actual total size um that uh we use for redundancy to map out to map out uh places where defects are so I guess the chip is bigger than it needs to be so that if there are any redundancies in the middle you would block out that row and column um I can't talk completely about the technology but suffice to say that there's plenty of overhead that we can we can map out um um we can map out what what where the defect is and pro still provide a standard size for the for the comput know rectangle I mean software C is a you know if you start with a an X by y rectangle of physical processing cores um software sees an M bym rectangle that's slightly smaller than that and we have enough overhead to guarantee that um the software sees that n by rectangle um and and and can can run on it and uh so you you take all these designs for this massive wafer scale chip and you send it off to tsmc right tsmc is our manufacturer yeah yeah the only guys that can do it is there any uh constraint there cuz I assume everyone in the world wants tsmc's new 5 nanometer or even smaller they keep going uh to 3 nanometer I think Apple's at um is there any challenges you guys face when getting the latest and greatest from tsmc um or making enough chips uh our our quantity is so small compared to everybody else I mean a lot for us is minuscule compared to their normal daily throughput so they don't they don't really care about that so the the manufacturing throughput is not a problem for us um I think you know in the early days of the company we we got access you know t since he wanted to do this near as I can tell and and that was before I joined I I was I was employee number 30 so I I joined about a year after the company started but they but they had engaged with tsmc and got access to researchers there which is pretty unheard of for a tiny startup who wanted this technology to happen and um and and they made it work and they continue you know obviously it's a slightly different process because you have to make the connection between the die as a as a one or more steps in the process afterward um and that's different from what everybody else does they just cut it up and put it in packages so it's a it's a slightly different process um and so we need to engage with them on on that um but but for the most part it's you know the the wafer scale part I think was actually the easiest part but the much harder part is if you imagine like your common video card and youve got a GPU on it and you've got this you know fairly large surface area to power it you can put big fans on it you can put Power Supplies all over the board you can put decoupling caps and whatever you need to to power that when you have all the DI eye on a single wafer you've only got the vertical space above the reticle and below the reticle to power and cool it so you got to imagine you got to you've got to power and cool a gpu's worth of worth of chip but only in this tiny little little volume of of above and below what that reticle is on the on the wafer wait you don't send the power from the sides no it comes in vertically oh and you can look on our website we we show just blow up diagrams of power planes and cooling planes and like I don't think yeah it's crazy how even from the nvidia's uh announcement like a month or two ago most of the challenges weren't in increasing the uh chip performance itself but everything that goes on top of the chip like the EnV link connects and uh memory uh bandwidth and actually even memory allocation to store all these weights and move them around to these different gpus it's it's it's a hard problem and the Mechanicals alone our our um mechanical Chief mechanical guy won an award from from the French uh I forget what instit it is there's only like a couple hundred people that have won this reward for the mechanical design on rassis because it's so unique and and and uh um you know it's a good design that's cool yeah cool so so one thing I'm kind of curious um you mentioned that the there was a lot of paperwork to to ship the chips to the Middle East I don't know I don't know for certain I'm not involed I'm an engineer the L so I have absolutely no idea I just know that the if you look at the the government you have to get waivers and things for anything you want to ship over to the Middle East it can be done and there are countries you can't ship to but the Middle East is one where you have to get wers and yeah because what I what I like to kind of touch on is I'm just sort of curious if you've heard anything about I think like the US has like their chips act where they have been trying to bring like a lot of uh Computing into the US I think like tsmc built a plant here um I know they've been like giving a lot of money uh to a bunch of different companies have you noticed anything uh since like you know I'm like a little bit farther away from Hardware than you are like I don't know if you like well we hear we hear stuff I mean it none of it actually relates to our business okay at the moment and it won't for a really long time uh but it because it's a really hard infrastructure to build out this speci not only is there special like you can build a Fab but you know one of the things that tsmc has is they've got packaging companies and part supplier companies and and testing companies and all this collateral business around their Fabs that you don't just build a Fab because then you have to ship stuff here and there to get it packaged to get it tested to get it and and it and that takes time and money to do that and everybody wants things really fast so so tsnc is truly optimized in terms of they have a whole infrastructure rure much like Silicon Valley is for startups and things you got a whole infrastructure in place that's really hard to replicate and so they're going to have to build Fabs and they're going to have to grow their own infrastructure around it to all the other hard Parts which are you know packaging and testing so imagine you were like um you know like a government uh official in the United States uh and you were trying to you know bring TSM like a tsmc like company here or like that infrastructure like what do you think you would do like uh cuz I think that a lot of people maybe listening to this I mean hopefully some uh government officials will listening to this podcast and they're they're thinking like you know what should be done like uh like what would you do like if you had like a magic wand and were uh you had like let's say like unlimited funds and you were trying to you know bring some uh chip manufacturing into the United States I think like in the United States for sure we do like all the design a lot of the design is done here but a lot of the manufacturing is done right like you know what would you do like bring that here I'm the last person to be an expert on this so anything I say is going to be just totally off the cuff I am very little expertise in this um I mean if if you had infinite money you just build it and and and and let people use it for free right that's that so the problem is exactly that you don't have infinite money you somebody somewhere has to see some return on investment for building building this out so right but I think that in a certain sense right like uh the money part aside is like you know you so you could give me a billion dollars and I would have no idea like even where to start right like I would just start asking around like you know guys like you like hey like where do I start like building this cuz I like um it's like I just like would have no my skill right find expert exactly so I think that like in Taiwan they have like a lot of people who like know what they're doing when it comes to building the chips and I I don't know if like we have that um skill set here as heavily in the United States like I I think you'd have to throw an enormous amount of money at it that's that's all there is to it and um you know there there it's like any any kind of Technology you want to succeed like take solar panels for instance you have to you have to you have to put an amount of money in it so the product reaches an economy of scale to where it sustains itself right and that's a really hard thing to do with chip manufacturing because it's the the investment is really large and then you have to have the throughput to to sustain it so in a certain sense like in order to do that if going back to like the thing if you were going you like You' want the Fab but then you would also want to have all of the things around the Fab so like what are the types of things that like you would maybe need in a Fab to um that like you know for like a steady supply of things like I know one thing it's like asml they have like the LI machine like I assume that's one thing like the only company that makes that yeah I there like a bunch of other stuff that you'd need too you know I think I think they're going about it the right way um they're encouraging companies like tsmc um to build Fabs here and giving them breaks tax braks and whatever in order to um sort of like seed the seed that industry but I mean you gotta you've got you know uh engineers and technicians that you have to train you've got other infrastructure that has to build like I said and you just have to I guess you would have to pay attention to all those things I would say make sure the soil is fertile you know don't just don't just don't just throw a Fab in the middle of the desert because there's sand there right you got to make sure that there's going to be you know um uh housing and Engineering uh train uh um uh schooling you know we got to make sure they're graduate students that can do the research you got you got to provide all this stuff all this fertile ground for the seed to take root otherwise it just it's just you know cogress throwing money into a hole um without without taking care of all the all the all the stuff around it yeah and I think that's like the a really important part because I I think it's very easy uh when all you have is money um it looks like the money can solve everything but in a certain sense uh like the money is useful but like all these things that you mentioned like you know the the right people um the right infrastructure having the uh the right location um like you know uh for example if there's uh if there was a a fab here uh let's say in the Bay Area I think a lot of people would maybe want to go work at it right but if you're going to build in let's say I don't know like rural Alaska where um it's it's really cold so you don't have to worry about cooling right but um you know it's dark like four months out of the year and then in the summer uh it just like the Sun never set so you might uh have to really do some convincing to get people to move there so it's it's even more than that there's a certain Serendipity to it as well I mean people have tried to replicate Silicon Valley in other places and and my thinking is one of the reasons it's really hard is you know there's so much there's a critical mass of talent here that if somebody takes a job at a startup and the startup fails um they can go to another startup and another startup and another startup or a big company for a little while then to a startup whereas in a lot of places if there aren't enough startups for them to risk career on if they go to a startup and the startup fails they don't get to go anywhere else right there is another startup around the corner right and so to build up that critical mass just happened organically here and it's really hard to to to make something like that just happen there's there's I don't know um kind of random things that have to happen to make that work that you might not be able to predict yeah that's very true I mean I think like at the bay area there's so many smart people um like highly concentrated in this area and I think that's due like a ton of reasons right like I mean uh Stanford's right here uh Berkeley uh I mean now all of the big tech companies uh you have all the money uh here you have like all the money yeah all the tech money right and um you're all within like a pretty small area like I mean you could probably throw a rock and hit a software developer um in this area which sure sure uh which you know is uh bad for the software developer that just got hit by the rock but uh you know it's uh it's it's good if you're looking for like a co-founder you're looking for money and I I think it's like it makes it a really like so M Place sure sure absolutely and and and starting with like you know uh um Fairchild and National sum way back in those days they had a ton of really smart Engineers that came here and settled here and um and it just grew from that and and I I know tons of stories of people that been here for a long time talk about you know hi yeah my dad invented that you know and as a kid I got to play with that um kind of thing and that's not true everywhere are you originally from this area I'm from Rochester New York oh okay and then college at Northwestern in Chicago and then I did a stent several years at Ron company on the East Coast Ron is a military they are they are they're well they're kind of funny vertically oriented thing they had a um like both Beach aircraft they own they they owned the Amanda radar range came out of that from it's called the radar range because some Engineers put their lunch in their radar transmission cone and when they took it out they realized it was warm and they were like why is that well it turns out that the radio waves will stimulate water you know molecules and that causes them to heat up and lo and behold a Mana was was formed in the radar range was was born and so there's a bunch of and they had own publishing companies with but primarily um military military stuff um components for planes and things and radar which is where I kind of cut my teeth doing radar I guess uh what made you want to leave that and come to Silicon Valley I wish that I had come to Silicon Valley right out of right out of college I really do because um there were so many startups like at a company like raon and it was great working there because they're they're big enough that if you if you're if you're kind of any little bit motivated you can do you can take the time to learn technology and things and that's really that's really awesome so I learned about Radars I worked in Virginia Beach on a radar for a long time I rode operating systems uh um all kinds of you know board design all kinds of really fun stuff um but Silicon Valley um had this startup mentality and this is purely a financial bid because people would come here and startups would be you know you'd get some giant chunk of shares to begin with um of isos and they would keep those really low and till right short of IPO and then then a company would IPO and they'd be worth you know you'd be done right in in a heartbeat you'd be done it's it's much harder now when I came here you know the rule was you know one startup might pop really well out of out of say 10 you know two or three might kind of linger a while and three or four would just kind of fail and and but you knew in a couple years you knew two or three years whether the startup was going to fail you can go to the next one and if you happen to place your bet at one of the big ones the ones that big you're kind of done and and that's a kind of a role of the dice whether that can happen or not because I've been at startups that I really liked I thought they were going to do well and they didn't who knows why um but if if you're if you're if you're trying a bunch of them eventually hopefully you hit on one that that happens to do really well and if I had come out here earlier I think it would have increased my probability of not having to work at an earlier age which is kind of a golden thing I think so um how long you've been at cerebrus um I am employee number 30 I've been there uh seven years now seven years so you said uh within two years you'll probably have a good idea of whether or not this startup is going to Su well so so let me let me let me when I came to the valley in um 92 it was what I was describing for in terms of its startup now it's nothing like that the Valley's been through a couple of expansions and traction expansions contractions and um VC money is much much uh scarcer in a way um the the probability that a company's going to IPO has gone way way down uh more companies are being acquired at the time and and you know if you can if you know go back to 92 you're in a startup you get 100,000 shares the company goes public at $10 a share they go to $100 a share you're not get stop working um that that that doesn't that mentality doesn't doesn't happen here now you're not likely you know maybe one in I don't know 50 or 60 will will will pop and and do something interesting and the rest will get acquired or or or fail um and now we have uh companies and and and venture capital and stuff gaming the system they don't necessarily care that the fundamentals are good if they can cash out you know um early regardless of what the fundamentals are they'll do that right um so taking a step back you mentioned uh cerebrus doesn't uh make as many chips on the scale of like apple Nvidia AMD um for tsmc you guys are relatively small customer um you guys focus on a specific Niche specifically trading um what what what do you think the future is is this your retirement plan when do you think cus will have a big enough uh Market to say okay let's IP who knows who knows I mean you you know I think our CEO made a really wise comment he said every every CEO at a startup should have a price in the back of their head that they will sell the company at If somebody walks in and writes a check with that many zeros on it they'll take it so you do what the market tells you to do right you do what what's right you you know in order to IPO you need to have fundamentals down I think and I'm I'm still a believer in that you need to have you know some number of quarters of profitability you need to be able to look to the look to the Future and see that you're going to be profitable you have done more than just one customer there's a bunch of things fundamental things I think you should have in order to IPO and not tank right you want to IPO and stay sustainable so what what is that number for cerebras because I think right now yeah you guys are maybe uh at your last funding round valued at something like two to four billion dollar or something you would have to be psychic and look into the CEO's mind to figure out what that number is either you just have to add a couple more zeros to that but if you if you compare that to nvidia's market cap they're like uh a thousand times bigger than cerebrus and that seems like would even begin to guess I wouldn't even begin to guess that's that's just not something I can speak to I I think that means there's a lot of room to grow though the Market's big enough for a lot of players actually I believe that's fundamentally true the Market's big enough for a lot of players and I certainly don't big grudge envidia their their market share they were clever enough to to to to do Cuda in the first place not even even aware that this was going to become a thing they were a crypto for that matter they just saw that hey this is an interesting thing if we release it it's kind of like a you know feel the dreams thing build it and they will come and and they did um they came and in and in in enormous numbers but I think we were talking the past and you mentioned that you can run Cuda code on cugus Hardware no no oh no no oh okay we've got our own custom Assembly Language oh I see okay okay gotcha gotcha um but like if I had add are things uh in Cuda I could like convert it over to um no oh so essentially models are described it's interesting because the the the and and again I'm not an ml guy right I'm just kind of associated with that all guys my understanding is that models are described in tensor FL pytorch as a series of layers with certain operations that happen at those layers and really the very top level of the of that is uh is describing Network which can be abstracted and broke into smaller pieces which are then mapped to kernels that do those operations and are balanced on the hardware that you're running on like like for instance a part a big part of the problem is if you've got a layer that is twice as fast as another layer right it's going to finish its work um before the next layer is ready for it so maybe what you do is you um yeah you use half the course for it you make them do them twice as slowly so that when that layer is done it exactly matches the the the throughput of the layer after that and so this is a job that software does in balancing kernels not just the selection of the kernels but actually how many of them and where you place them so that the the data flows maximally through the thing with with little bottle few bottle necs so I I know Nvidia has had Cuda for a while um but I think that's still kind of a a little cumbersome to work with for software developers and I I was reading uh about bend a new um highlevel programming language similar to like pythonesque where you describe this concept of uh threads and it just handles all this parallelization the background good for them not just with a single uh GPU but across like massive uh workloads so what what is uh I guess cerebrus is equivalent to Cuda and how do you guys make it simple I know you weren an ml guy but if you had well we have I mean we have we have just like them we have Kernels at map to our assembly code we have operations at maap to our kernel assembly code and you just have to do that mapping um and I guess the difference is that um it's not our a our assembly code is not open source it's not something you can just look at and write your own code for um all of it's now controlled by by us um at some point probably they'll have standard developers kits out there I think there is a standard developers kit that you can use to um I think you can actually just become a developer and download it but I'm not 100% certain of that but anyway there's um you know I mean they're not going to stand still they they're going to optimize that's you know that's that's great for them and the whole Market making making programming easier is is a job everybody is kind of working on I think so what's your stance on the whole closed Source versus open source both with the I guess the software stack and the hardware architecture design both have their their their places where they're useful I mean it's like am you know let's let's take a more abstract question like C versus python right they're two very different languages but they both serve very different purposes there are some things that I would absolutely want to code in C or C++ and there's some things I would absolutely not want to code in CF C++ and want to do in a scripting language like Python and so it's the same kind of kind of problem um there there there are things that you want to do one thing with and there are things that you want to do with a different tool depending on what your use case is and how fast it is and so open source great stuff it has enabled you know like like let's just take an obvious example you want to run a server and you want to run software on a server well let me tell you all of them to running is some kind of free version or free version of of of Apache and Linux on them and you want to run compute it's probably all running Linux and there'll be some layer of layers of open source code on top of that on top of that to to get it run VMS or or what have you um and so that's all really great but you know a long time ago I worked at um rendition it was a Graphics Company and ear one of the early Graphics companies and and you know back in those days drivers were incredibly different were and difficult like people probably less frequently hear the acronym BS s OD or blue screen of death um Microsoft Windows would crash all the time with this thing called the blue screen of death and it would just be a crash and it would be fatal and you'd have to reboot and they got they got enormous amount of heat for that and 99% of the time it wasn't anything that Microsoft wrote it was a third party driver that did it was your lound driver or your graphics driver or some other kind of piece of driver that was and and so they created this whole certification methodology called wickle to help certify drivers so they wouldn't be shipping things or ultimately computers wouldn't crash right and that's kind of a goal even though it's not their software they want to enable so that you have to lock down some stuff to prevent instability I mean and that's kind of where closed Source stuff happens is you lock it down to prevent instability I mean that's one of the big reasons why I've loved the Mac uh for a long time yeah I use M CU it's got Linux underneath you know and and I can open up a shell and I can do all the same stuff that I could do pretty much on on a Linux shell and and even even though Microsoft has tried to create a shell to do that they have some very fundamental like piping you know piping in a Linux shell is piping between multiple commands but in a in a in a in a in a a Windows terminal window it's not you can't just pipe stuff the same way you would you know it says oh this command doesn't accept that or argum and it's really like everything that you want to do you have to relearn for their shell right and I'm not saying it's wrong or bad I haven't used that that much I think it's just that uh most computer science students learn to work with Linux when they're when they're because it's free and then transfers easily to Mac versus window and that was that was that was Apple's Brilliance when the Macs I remember when the Macs first came out which is dating myself I remember seeing that printer I remember seeing a Word document and it had pictures in it now maybe hard to believe but that was a really hard thing like when I when I started my career you you handw wrote your document you handed it to the admin and she would type it up wow right and then if you had figures you would reference the figure and you would have an attachment with all the figures at the back and they would staple it to it and they would copy it and then copier right and and when Max came out you could Engineers could then write their own documents and embed the figures right in the document so you'd see it right there and it would print out on these crappy little dot matrix printers but at the time it was like oh it's all embedded it's and so like the admins went from one admin per five people to one admin for a whole department because all the engineers were typing their own documents and editing them and and going through cycles of editing much much faster and and and much much quicker and so the the whole the whole sort of like admin pool went you know decimated by that so computers took their jobs pretty much but I mean think about think about how I mean think about you know PE typewriters right you know imagine your imagine your IBM electric vendor right the moment the moment computers came out all the typewriters went away too I'm sure it wasn't an overnight uh it took a well it was pretty close to I mean a couple years all the you know once once computers like the Mac came out all the type artists went away do you ever worry that with uh these increasing technology Gams with like AI that um were a lot of people are just going to lose their jobs um like it could be all of us that's a great great question and that's kind of one of the things I was hoping to talk about on this podcast is some of the more optimistic views on where AI might be taking us and and one of them that I used just the other day um um was you know a l a while back they did this um paper with they they got together some few dozen Radiologists and they were looking at uh brain scans um and they were trying to identify U tumors and then they trained an AI to do the same job and they got a few dozen Radiologists I think got a score of like 96% but the a got a score of 95% right the difference is that you know Radiologists take a long time to train and and have a very large salary and they need to be on staff at every hospital and at small hospitals it's not possible to do that but if you have an AI That's just almost as good as a real radiologist every doctor everywhere in the world can run a copy of that Ai and for zero time and very little energy and and cost can have an expert's view of of an x-ray and we're talking doctors in you know poor countries doctors in rural neighborhoods small clinics can get access to worldclass Radiology uh interpretation yeah and also I think that would just help uh doctors and radiologist just Radiologists just save time as well right because uh before if it takes you a lot of time to I don't know I'm not a radiologist but like I would assume that um it probably takes some amount of time to just you know look at the image and just try to like make sense of it I don't exactly know like how long that takes but maybe a radiologist can do it in let's say like x amount of time yeah yeah minutes uh yeah minutes probably maybe like tens of minutes um it it's a b complex but I would think that uh the AI might be able to do it within seconds or like milliseconds as well sure sure and and and and so um you can still it won't eliminate the need for for for Radiologists right because there'll be some questionable ones but the AI it's a little uncertain and so you kick that back to your team of Radiologists and also I would assume that um the Radiologists do more than just interpreting the images because like uh they have to go and you know run all the Machinery make sure they don't like burn you with ultrasound like you know give you like cancer from all the x-rays right I'm not sure that's the Radiologists I think the the nurses and the radiologist CL are doing that uh yeah I think that's true uh although I my understanding through talking to some like Radiologists is that like there is quite a bit of training in order to like actually run a lot of that equipment yeah so I think you run the danger of making of of Deval devaluing certain jobs because it can be done better by Ai and and so and then you like if nobody goes into radi ology you lose the expertise right and that that's that that could be problematic um and so there's so I heard this interesting quote the other day it was that and and I I kind of I don't believe that AI is going to replace jobs wholesale the way there fears of I think though um the quote I heard was um it won't replace jobs but a person using AI will replace a person who doesn't use AI for the same job and I think that's true AI has its best um I I think today um it's best use case in being supplemental to work that people already do to make them more efficient and make them you know may maybe take some tedious Parts out of their job so they can focus on more creative and interesting parts of it and that's that's my hope yeah I I I think that that is probably true um for a vast uh amount of jobs although I think that uh depending on the job it it may just completely eliminate it so for example um not necessarily like um AI like all jobs but like for example I think of like uh truck drivers right so it's like if your entire job is just uh driving from point A to point B you're like uber um taxi driver right you know self- driving cars then like that uh a complete U population of people may be out of a job uh and be replaced by like a wh self-driving taxi right maybe maybe at some point yeah but but you know number of desks will go down and and things like that on the road because I think AI will be better at driving ultimately and and maybe that truck driver maybe you'll need to have a truck driver in there in the in in the vehicle just not necessarily driving um and maybe the truck driver will have a a train of two or three vehicles behind him which is a company that I think has been you know in the works before that you know years ago um and that truck driver can then do something more productive um during that time when he's watching them you know they're Manning the truck yeah I think that's true and I think that like in the short term we'll probably definitely see something like that uh I'm just thinking like you know let's say not in the next five years or 10 years but let's say the next like 50 years or something like that when I think but like we'll probably be able to agree that in 50 years from now like probably many people won't like ding might even be like a lost skill yeah practically in the next all bets are off at that point right you know um so like what do you think like if the AI becomes smart enough to let's say replace all the jobs like what jobs or like what do you think people will do uh in 50 years uh from now like you know if or like 100 years whatever like sometime like when the AI is able to be smarter than like any human to like the point of being like a god um which I think could be the case in the future maybe well let's just go there a little bit let's just go there a little bit and and and and I'll outline a couple of really interesting scenarios one is everybody will be then free to pursue whatever they want like imagine imagine people like AI will probably be running universities imagine people learning to get degrees in philosophy and music and and Math and Science and all kinds of really interesting things that they have time to study and pursue as Hobbies rather than now clocking into a 9 to-5 job you know maybe somebody wants to spend their time making music or something like that then their all their needs are taken care of and they can just make music um if the I isn't making it for them um I think I think the biggest fear is though that things get done because people have influence and money to do them and the the people traditionally in our species the people with the influence and money have always wanted to control the technology and the things they done and so you get a very stratified society and so if AI can do everything and say a lot of people are out of work um who's going to provide for them right like suppose a robot you need a robot right but who's going to are the people that making the robots with their energy and time and influence and whatever counts for currency of the time are they going to just give you that robot or they going to demand something in return for that robot right and and and so technology has traditionally found ways always found ways to stratify society and so we as a species have to find ways to counteract that tendency that we feel everything is so scarce and so um and so dangerous that we want to we want to protect ourselves by accumulating as much as possible by having as much as possible by being as as you know having as much you know um let's call it defense right would be that money to hire people to literally defend you or big houses where you don't have to talk to people or or any number of things have you know or enough financial gain to hire lawyers or whatever people have always been really afraid and so they accumulate with empowered to defend themselves against things that aren't really Zero Sum games you know against things that you know are as a society we can figure out but but deep in our Reptilian Brain we're programmed to believe that things are scarce and that there's danger around every corner and so that leads people to be very defensive and be very and want to hoard things my hord I mean wealth and power and other things at the same time and so we have to we have to crack that nut in society I think too and that's a nut that you know philosophers have been trying to solve for for Millennia yeah it's it's it's interesting because uh you mentioned um like wanting to hoard wealth um and power and whatnot uh but it's in a certain sense I can Envision like a world of abundance like a complete abundance uh with AI but um that's the St Trek world right where everybody all their needs are met they can do whatever they want to do um and I think it's it's really hard to make sure that um like not to say like that abundance will be distributed like equally necessarily to everybody but um to like potentially like raise like the base quality of living for corre everybody yeah and that's and that's kind of the thing if you can find a way to um it's it's not enough to raise the quality of living of the lowest 50% an increment where the quality of living of the highest 5% gets raised um a million times that you know like you could you could say well Healthcare has been better for everybody on the planet for everybody it's gone up for everybody therefore let's not begrudge the the um the economic discrepancies that are occurring for other reasons um but I'm like no well I mean this the question should there be billionaires should there be trillionaires is there something you know um that like like if it's okay for people to accumulate wealth and power is it okay for one person to have it all like if you take that to this ridiculous thing is it okay for one person to have it all everybody else is poor and so right now yeah I I I think that we probably agree that like taking to the that extreme like if one person uh just had all the wealth uh that would not be okay and I think if that happened then all the other billion people or whatever if there's one person and the 99 people had nothing right they would just go beat the crap out of that person and be like it's well the counterargument is if that one person controls all the robots all the wealth all the technology then the 99% can't do anything that's and that's the problem right and and we see societies throughout all of time being being that way from Kings to popes to um you know um oligarchs to autocrats to you know we've seen societies go through that same struggle all the time where the one person get makes a grab for as much power and wealth and subjugates the the the rest of society look at North Korea it's a perfect example of that and you know what North Korea's been there for 50 years right and it hasn't changed and I don't see it ever changing right I mean people are not going to rise up I think uh authoritarian societies are awesome when they work well uh you're trying to say benevolent marar is better than yes any other system the problem is you can't always pick the benevolent Monarch yeah right you have to find some way of succeeding like like imagine so here's here's the scenario imagine you know North Korea and imagine the whole world that way the whole world right there's no Democratic country trying to balance that out or trying to reverse it and and that's that was the whole I think fear for the red for the Communist scare was if the whole world is that way everybody's repressed and there's no way to turn it turn it around if you deprogrammed everybody and you've you know you've given them such propaganda all the time and you've beaten out of them any thought of resistance and the whole world is that way um what can you do and so so that's why this enormous amount of resources was put in to try to try to go up against I think at least I'm not political scientist to go up against communism because they didn't want the whole world to be that way I guess democracy prevents the worst case scenario but uh there is a lot of friction and especially with our capitalist version of democracy the income inequality is just like rapidly expanded especially in since the pandemic I think the amount of wealth that was created in the top 1% was just in the double digit trillions I think it's been 30 years it hasn't been it hasn't just been since the pandemic I think although that ex a little look at a short time frame of 2 three years their wealth doubled if not more the we should get back to AI but along those lines there was a time that I understand when the the the multiplier between the highest paid employee and the lowest paid employee was like 30 so highest to lowest now the multiplier between the average employee and the highest plate employee is 300 400 so it went from highest to lowest which means everybody was in between of of 30X to average the highest of 3 400x and and yeah I I do consider that a failure of our current current economic system but that's that's a I'm not a political scientist so um same uh but the reason I was talking about that is because that is the one end of this extreme um where the other end is yes things have gotten way better Healthcare has improved uh we have access to amazing technology um and you've been here in the Bay Area since '92 so I can't imagine the amount of progress you've seen here um and how quickly things are moving now and to give uh the you know billionaires credit they are trying to make access to these Technologies as easy as possible I mean GPT 40 is now free for everyone everywhere in the world same with the Google's Gemini so I I don't know if we could have imagine a Time 20 30 years ago when we would have had this relatively um super intelligent being in our pockets for free the the vast majority of people working have no need for anything like that you don't think it's helpful day-to-day not to the to the made cleaning houses or to the the garbage collector picking up garbage or The Farmer in the field or none of that stuff it might be helpful to the corporation that has farmers in the field but not the guy out there picking not as an employee but just as a as an individual I think uh I found myself turning to these llms more often than going to a search engine so just for my personal information I think these are incredible tools yeah but people don't people I think most people don't care they don't they're not searching for stuff right like we're that's that that's one of I think the the conceits of tech people and I and I have it too is we think that because it can make our life so much easier and better um why is it making everybody's life so much easier but it's really hard for us to put ourselves in the shoes of you know the the plumber or the or the who don't care about that kind of thing don't doesn't help their job hardly at all so what do you what do you think is the point of life then what what is the purpose oh man uh um and can L help well and let's let's get to there because that that gets us into some things I really did want to talk about um my personal philosophy is one of of you know the point is is to decrease suffering in the world that's part of the point at least that's one of the edicts I go by because it doesn't require any kind of Supernatural belief it's just something you can see and act upon and I think there's a lot of opportunity for AI to help decrease suffering and one of the ways you do that is by um trying to um filter out information that is detrimental to the to the advancement of society and uh I've talked about a couple ideas on the in the in the Meetup on how to do this um one is to um like suppose you trained an trained an AI to U recognize um when someone's trying to convince you of something that is a um they're using either logical fallacies or using some kind of cognitive bias trick um maybe we can train AIS to detect hey what is the real AR argument here that this person is making and what is this sideways attack to try to to try to win you over or get you to believe or follow them um that they're making that has absolutely nothing to do with the things that you want in your life it's just a A demonization or a a a a an inciting of fear or some kind of some kind of anger uh to get you to believe something so that they can convince you to to to vote the way they want to vote or to to act or support who they want you to support and I can see like a future version on a phone call similar to how in Gmail you get an email and it marks it as spam so if you're talking to someone and they they have a conversational pattern when they're trying to manipulate you it marks that as okay this this guy being manipulative watch out yeah no exactly and and and and really I'm the goal isn't necessarily just to um to tag people and punish them the goal is to get people better at recognizing it so that they can support the people who are more um um less disingenuous that are more straightforward in what they're what they're saying and how they're how they're presenting their argument and and and one that doesn't trigger some kind of anger or some kind of uh um um disgust or something like that like disgust is a really powerful way to for a politician to demonize a group of people and there was a recent study that talked about how just calling someone a name just saying something derogatory about them makes it more likely that the people who have heard that will believe something worse right so if you say oh that guy and it may not be true and it doesn't even have to be true it just it just you know because of our cognitive biases it just primes our brain to to to to to think to associate that person with something negative so the next thing you said that negative we associate with them and you can just ride that all the way to the point where you're hating somebody and and and I think that's that's a that's a trick that evolutionist paid on us because it favors um treating everything is in a negative way because you survive better but we don't we we're not animals anymore we don't need to do that but our brains still don't know that right so you know if the grass is rustling um 99.99% of the time it's not the lion right it's just the wind in the grass but if you always assume that it's just the wind in the grass the one time it's a lion you die and your genes get out of the gene pool and the next guy that assumes that it is a lion more frequently Will Survive more frequently and those genes will carry on and and you'll have you'll have um you know you'll you'll you'll you'll have a a tendency to an evolutionary tendency to to give more weight to negative things than positive things and that's kind of the way I think our brains are coded so you talked about a bunch of things that I wanted to touch well one of those uh important conversations which is pointing for everyone I think is the upcoming elections and how people can use llms to maybe generate more propaganda AB do like Mass manipulation at scale um create like misinformation because it's it's so easy to generate content both uh text and images and now soon maybe like video footage too and we won't even know what's going on because they will send it directly through the platform to the person receiving it and that was one of the that was one of I think the um takeaways from the whole Cambridge alga thing was that I think they discovered that upwards of 3 to four% of the population is prone to believe in conspiracy theories and and and we now know and back then they were saying well you know could it be that bad but we now know that it's actually worse than that you know that up to 20 to 30% of the population can believe in conspiracy theories without any without any problem we also know that it only takes one they also prove that it only takes one to two% of the population to swing an election one way or the other and they practice a number of countries including brexit including the the 2016 election and while I don't think you can say that like it made the election if it had not been used I think the election would have gone the other way but there's a number of things that you could say that about this is just one of them and in this upcoming election you know I'm not seeing a lot and I don't know why and I'm hoping it's not because I'm hoping it's because there are more guard rails that that technology companies are putting on to prevent that kind of behavior and I'm hoping it's not because they've just got better at hiding it and and but we do see on social media platforms this just enormous amount of fear and rage and hatred and all these things that are stoked by partly by you know information that AI could classify as hey look this is not this is not in your best interest this is not in our species best interest let alone our countryes or our democracy's best interest um let's say you were uh narcissistic megalomaniac how would you take take over the country if you were going for going after elections um a narcissistic me megalomania you you'd have to so like if you look at the you know I don't sure I shouldn't tell you this well I think maybe there's a way to to to to to prevent it yeah and and I would encourage your listeners to to to look at this and and and maybe look for ways to use this to prevent it but if you look at Wikipedia there's some 230 cognitive biases that are listed in there and you know um marketing people have been really good at getting you to do something because of cognitive bias and like for instance think smoking right you know smoking was made cool by having movie stars and and and movies do it and having television ads saying it's really cool and all the cool kids are doing it and stuff like that they marketing fake you know candy cigarettes to kids and and and and and whatnot to make it cool um and that was all marketing and it was just to be fair smoking is also just it sells itself it it just feels so good in the short term and the longterm uh would you even start if you knew that it was long-term detrimental why would you even start and the reason is that are long-term detrimental like I eat like a dut because it tastes good but I don't you absolutely so but but anyway my point is that these marketing people have long known that they cognitive bias you know you put a if you look at some of the ads from the 50s and 60s you're selling a a radio and you put a sexy woman on it look at the airlines the way they used to dress stus right there's they know that these cognitive bias things can influence you in favor of what they're selling or pitching or whatever but imagine an AI that uses not only that particular cognitive bias but all of them you can trade it and not only all of them but they have a profile view so they know what works best just for you and they can emphasize the ones that will convince you of something stoke your fear your hopes or your whatever um playing on any possible cognitive bias that they can find that does it um and convince you to do literally whatever I mean convince you to you know storm the capital right to believe that that the election was stolen and that you should go fight for democracy I believe that a lot of those people really believed that they really believed that there was the election was stolen despite enormous amounts of evidence but because the the the tribe they belong to and the leader of the tribe they belong to told them that it was true and they're loyal to their tribe um they they they they did that in in in in a lot of them in good faith you know believing that hey this is wrong we need to we need to turn this around and and so um um we need to find ways to to far it out that kind of Truth uh or the possibility of falsehood in those narratives and AI is uniquely positioned to see like I said when I was talking about cognitive bias to detect when an argument is being made an ad hom of attack or a straw man attack or or some other kind of enraging thing is being made that isn't really the point but it does get you inflamed then I would I rather than just say these people are bad they're doing these things I would say hey it's the population that votes for somebody you guys need to learn how to recognize this stuff so you're not ConEd right so you're not you're not C there's a science fiction story where one of the statements the guy made was people will believe either what they hope is true or what they're afraid is true right and often times the truth is neither of those two things and maybe somewhere in the middle but if you can Stoke their fears and hopes you can get them to believe what you want them to believe and that's the promise of you know a lot of religions a lot of politicians you know they hope it's true that this is happening or they afraid it's true so they they they they do that and and and we have to get I think we I would like to see people learn how to recognize what the truth really is and and and understand and maybe AI can give them some guidance as to how they can recognize that what to look for um if you had a speech for instance and you could you could mark it up you know or even a website where you point a URL to the website and the website would mark up all the things that were you know cognitive bias arguments and even just the difference between editorial and and news right you know news is a news is a presentation of The Facts of the things that are happening right there are things that you can prove or disprove there was a hurricane there was a riot that you know there's a fire you know all these things are and they may be false but they're at least things that you could prove true or false right from editorial which is you know there was a fire because you know PNG p is a terrible company or something crazy like that right um those are editorial they're not but and if you could if you could if you could run text or spee through uh um this AI filter and you could see that 90% of it are things that could be TR proven true or false you say that's news whereas 90% of it is things that are just you know opinion or just something that's neither provable then it's editorial and two you know it started way back in the days of Rush limo where he was he was presenting as news not only editorial but whenever he was taken to task on things that he said they're wrong he said I'm just comedian this isn't News why are you guys taking me seriously so under that guise of being a comedian he would be pretending things that people were taking as news which weren't and we saw that on Facebook as well you know Facebook didn't want to be you know a lot of people were spreading news and I air quote the news um that they thought was really news but it really wasn't it was editorial or was some other kind of opinion that they would claim is truth when it when it really really wasn't so you could have an AI kind of like Tred to differentiate between editorial and news and maybe people would learn to detect what's news and what's editorial and put it in a proper weight instead of just going on and consuming media thinking it's all truth when most of it now is editorial much less of it is actually news and most of it now is editorial and news is a much more important way to be informed because it's based on things that happen facts and whatnot yeah you know I think that this would be a really interesting uh project to build so uh for those that are listening that want like a fun little uh weekend project I I feel like this would be um a thing that you could build maybe nice little Chrome extension uh where you can go and uh look at all the Articles uh that you're watching maybe like I don't know uh if you like as a bonus it could even help like interpret YouTube videos or whatnot to look for the cognitive bius is there um I think that this could be a project that could probably be built within a couple days uh if you know you're doing uh AI you know and and if you want some more Exposition on this contact me through shash Shanker Mark and I'll be happy to work to to to kind of outline more of of my thinking on how this can can benefit our Humanity actually I'm going to I would say democracy and culture you know we have a ton of obligations in our lives we have obligations to our family we have obligations to our to our city to our high school to our College to our to our state to our country to the world to humanity to Nature and everybody has to balance some of those obligations out but I think the obligation to humanity which is a little different than the obligation to say nature is to try to get people recognizing ground real truth in things and being able to move Society forward rather than rather than backwards you know I I agree I think that it's really important that we can all at least get the same fact right um so um you know if people are disagreeing on the fundamental things that have happened um then it's really hard to have a meaningful conversation if you can't agree on that um and I think AI could you know help uh get people at least on common ground and avoid um any biases cuz like I know for a fact that I am like impacted by biases all the time and like it's really really hard um so like I I think that you mentioned that like a large percentage of the population uh believes in conspiracy theories um and I I think that like to to defend them for a moment um I think that it's it's really easy uh to get swept up in um potentially conspiracy theories because the problem is is um like there's a lot of things that I think like in the past were told as like the truth uh but now in like hindsight we realize uh that maybe that was like wrong right so to take like an extreme example right like um like let's say like 500 years ago like slavery was maybe like seen as like a like a normal thing and now like we realize that like okay like I don't think anybody's like trying to defend that right uh but like you know at the time it was just like a normal part of life and like you know nobody was questioning and I think that like in order to um so like once you find out that like one thing that you found um was maybe untrue um then like you start to think like well what else is you know also not true because like um you know your your bandwidth is limited right like if you are you know working at 9 to-5 um if you're uh you know busy you have obligations you have you have kids you have to like worry about your bills um you have all these things and like people don't have time to like actually go and like you know vet to see like if things are like actually true right and um uh if somebody you know makes like some sort of like YouTube video or like makes some sort of like a video saying like hey like I think um that like the government is full of lizard people um like right like I mean you might start to believe it right it's it I think also way down in our brains it's it's more important for us to belong than us for us to be honest right and and because survival is depending on belonging to your your clan or tribe or whatever and it's not dependent on you telling the truth often right so if if your tribe says you know we're going to kick you out if you don't believe this you know you have this tendency to want to believe it because you want to belong right and that's independent of you know kind of as a science Tech person you know I kind of like I'm a fact-based person but you know facts are often times you know uncomfortable right and you know it might lose you friends it might lose you family it might lose you all kinds of things to be talking about facts where you go it's not that important to me to take a stand this political stand if I don't get to talk with my aunt anymore or my uncle or my grandfather or my son or my daughter or whatever I would rather talk to my son or daughter and believe these uncomfortable things and and and and participate in their lives than not and so that's that's a thing we that that again one of those things that is as more as you know as we if we want to prove that we're more than just animals we have to fight against that tendency to be tribal over more important goals for our you know for ourselves for our for Humanity and for the planet yeah yeah and um I think that hopefully AI will be able to help with that because you know life is short and while it would be nice to you know go and um be logical about everything like sometimes it's just hard we need to take like mental shortcuts um well find find ways I mean like like people out there find ways for AI to help with this and I had this idea of a social media where a priority you tell people that half of the participants in the social media are avatars they of some sort of bots but they're they're designed to be um um affirming they're designed to be you know not hostile they're designed to um to to be to be interested in you right and a lot of people won't care you know and if there's a if there's a bully or if there's a troll the AI is designed to to to engage with them and slowly wean them out of the conversation and try to get them to be better citizens and the beauty of AI is like like TR R always like to have the last word but AI can always have the last word it it never gets tired it never it never breaks character never gets mad it never gets fatigued and it can be and and I'll say it has to be you know take a big stick and whack them with it it it can be very non-judgmental and really engage with a person to find out why they believe what they believe and and how they can maybe believe something different that will make them more productive in the in the community that they're engaging with um and and that that would be kind of like a a social media thing where you know friends and people can get together and some of the agents will moderate and maybe be do provide a little bit of lightweight therapy and and and actually pull out destructive people rather than just Banning them right pull out destructive people and try to coach them to to to be to to um to engage in more more more productive way that's a really interesting idea for a um social media I I also think that um it'd be interesting if you didn't know who was Bots and you're interacting with so so I thought of that question I've turned it over and and and a lot of people have said that well if they didn't know who was Bots and but I think trust is tantamount yeah in this and you would have to tell them a PR because if someone feels betrayed you will never ever get them back they'll be enemies forever so have to not betray their trust what I'm saying is like um imagine it was kind of like a game where you said like when you signed up and like was super clear uh it said like got like a big banner at the top of the page like 50% of the interactions will be Bots U and then like you don't know who it is uh so like it it's kind of like a game to try to like guess like it's sort of like a they're doing that so they're doing that there's a there's a and I really don't like this show it's called the the circle I saw that yeah they C people in a building and and and it used to be that you would there would be a couple people catf fishing they wouldn't be who they said they were and they would try to get people to believe in you know they would I guess both people offer something in the tradition style and this this season or recent season one of the catfishes was an AI right and and they it would it would be fooling they told them that it was an AI and it would be trying to fool people into believing it was a real a real person or or real Persona of something wait so they said that like I nobody knows who it is so they said like one person's an AI like so what would it be like 10 how how does this show there's like I don't know 15 or 20 people and everybody knows that there are a couple catfishes there and they know that they're there look they're just trying to figure out are you just like talking to like like a wall well so you're texting each other oh you're texting each other T right there pictures and some of them are are not real right so there are pictures of everybody in and and and the reason I don't really like it is because well for one it encourages people to be dishonest about things and it makes that into entertainment and and for two um entertainment is the worst of humanity right it's you know it it's like um to promote people being the worst of human ity as entertainment feels kind of gladitorial to me it's kind of like let's let's just I don't know it just rubs me the wrong way ethically yeah when you said entertainment is the worst I don't know if I agree with that but I do think reality TV which brings out the worst in people and Records them 24/7 forces them to be in uncomfortable situations and make mistakes as human beings naturally do and showcase that on TV for the best to succeed in this program you have to be dishonest yeah right deliberately and and you know to some degree the I I kind of like feel worse about the people who watch it right because the people are in it know it and they know what they're doing it's like kind of a boxer right the boxer knows he could get beat up and it what not but the people driving Joy from watching somebody get beat to a pul I don't know but the interesting part where they sent an AI to pretend to be a human being right it seemed very successful so I don't watch the show so I I watched a couple episodes of of that season uh cuz I was like oh this is this is uh it's like a touring test fun yeah it is a touring test and the AI was able to construct this Persona and I think it was some kind of an agent running in the background that had some system prompt that said okay be a likable contestant in a game show manipulate lie do whatever you can to get people on your side and have them vote other people off off and you know you got to be in the game make alliances uh be non-threatening uh throw shade on other people yeah so that was their loss function like don't get don't get disqualified whatever they did a really good job oh no no doubt no doubt because it's I don't think it's hard to do I don't think it's hard to trick people because people are inherently want to believe right the things that people say I guess to be fair at in that reality show setting everyone was primed to be a little suspicious of everyone else already and statistically you there's like 10 people one AI it's going to be hard for you to you know I think I just don't like I'm a little squeamish about normalizing that as acceptable behavior right you see something in a show and yeah it's just Hollywood yeah it's just a show yeah everybody knows but it normalizes the behavior so that if you see it in real life yeah you don't think twice about it oh yeah I was on the show it was funny I like the show I I enjoyed the show quite a bit yeah it was bad for this person to do that but I'm used to it agreed uh this is going off uh topic a little bit uh but I I I do think to maybe salvage some of uh some good out of this reality TV World um people do realize that they have to live their lives outside of this television show and carry on that reputation that they build in the show so I think uh some of the recent contestants are realizing that okay they can't be absolute Psychopaths they need to be a decent human being treat other people well because it seems like they're all uh on average workingclass people who would really benefit from winning the prize and helping their family their kids but but um but but holding that up as a prize you know you know for for this for this period of time you have to be a horrible person and we'll give you this prize it'll make that's a choice that's a choice you had the option sure yeah I guess but in a certain sense like people don't necessarily have the choice cuz like you know I mean if if you have no money and you have a bunch of like debt cleor calling and somebody says like oh if you're a horrible person we'll give you $100,000 it it's a better choice than you know going out and trying to rob a bank or a liquor store to be sure but yeah Society should F find better ways to provide for its less fortunate so I I had something uh I wanted to ask you with uh one of your previous points you mentioned we have obligations as human beings to the rest of humanity to society to one of them nature and I I would say maybe those are lower down the list of your most absolute critical priorities closer to home maybe is you know your immediate family your partner your kids and so on so it remind me again I think you have a son who's in middle school yeah and he's in high school high school now okay what advice are you giving him to be a capable member of society in this rapidly changing world to maybe use these new technologies and what what is he going to study uh what are you encouraging him to learn um what tips or things are you trying to get him to you know it's it's first of all Let Let me let me talk about the obligation hierarchy that I sort of thre out it's not static Sometimes some of those obligations outweigh others at different points in time and you know oftentimes they're conflicting with each other in order to help my family I have to do this negative thing for the environment right or or or or or something like that so there's there's that um and then as far as the the advice you know he's he's his own person the advice is the same as you would give kids throughout Millennia right don't lie be a good person be compassionate uh look out for less fortunate you know maybe more specifically since you're uh someone who's worked in Silton Valley in Tech um if he's interested in thinking about his career and going to a college soon Tech is a good career tuch tuch is actually one of the few careers that going to college makes more sense these days you know colleges have gotten a little bit ridiculous in terms of the value proposition for for students going into them and Tech is one that was is showing to be actually worth worth the education it's it's less expensive than some others and you you know you don't have to go to a Stanford MIT to learn how to be a really good coder or or some other kind of tech thing and they'll probably be a job waiting for you that'll that'll cover your your college expenses so your debt um and so you still think so when by the time he graduates maybe uh five years or something from now I think the university and banking system has let students down and so I'm going to encourage him not to go to a hyp expensive top tier school because I don't think it's a value proposition I think half a million dollars in debt even if you do get a dream job you know you're going to be you're going to be a you know um an indentured servant voluntary indention serve it for the rest of your life and that's not a value proposition well the public schools here in California are not too bad right the UC system is relatively affable they're all good no the I like the UC system especially if you're a resident it's really really excellent if you can get in the the the the price is good for residents the the problem is that the that there's there's a a million International students that pay completely willing to pay exorbitant tuitions because they have the access to the money and they're more of them than there are people that are residents and the colleges love them because they pay three or four times a tuition that in state person does so there's a lot of pressure on state colleges and universities to increase the ratio of non um resident students or decrease the ratio of Resident students and increase the rate of non-resident students because they can pay for all the really expensive gyms and food like I went on a college tour recently and I can tell you the food was a thousand times better than when I went to college when I went to college was like a was like a high school cafeteria you you walked in and there were like maybe three different ENT trays you picked and you got your tray out and they put it on it they had like 15 different stations with different kinds of ethnic food or seafood or or or and there was a grill and there was there were literally 15 different stations they could get food at you know and I'm like they got to pay for that somehow right and and that's I think the increase in tuition is covering that cost because because it makes them competitive right and so you know students what's really funny is when you want a loan from A bank usually as an adult you have to have some kind of collateral you have to prove that somehow you're worth it and in the past banks that would lend money to students I think did so because they were they were banking on the future career the student would have as being the collateral for the thing so there's no real collateral there's no real understanding there's no like a car or real estate or anything like that behind it so it's a future thing and because of that they could make it anything they wanted right they could make it you know we think you're going to get a job that's going to pay a lot of money so so the banks would would give bigger and bigger loans the universities were only too happy to and I'm not saying this a conspiracy theory I'm thinking this is just the free market in action right Banks would get bigger and big loans colleges would would charge what the students are willing to borrow and they just go up and up and up and up and up and up and up and it seems an allous the health insurance ins industry in in a way this yeah there a lack of transparency of of uh transmission of the of the real cost to the consumer to make an informed decision but Banks can't make an informed decision because they don't know what kind of job the students going to have so if they get no job then they're totally out of luck right if you if you if you if you spend a $250,000 on a college tuition for a degree in economics and there isn't a job waiting for you well now you've got this loan that's big than most people's house loans that you have to pay off through your entire life right people spend 30 years paying off their mortgage right uh when I got out of college within two years I paid off my student loan it was the student loan was literally a fifth of my starting salary and now the student loan is two to three extra starting salary yeah you know I I think that that's true with um you know the the increasing student loans with the banks but I also think it has to do with uh the government uh giving loans as well because you know there's like both like I think the federal back loans and then there's also like the private loans uh so I think like there's a lot of money from there as well well the banks like to give loans because a lot of them are backed by the government right yeah so that it's it's like the Savings and Loan scandals of the past right if they're guaranteed by the governments then banks are only too happy to make you take let you take as much risk as possible because it's not their dime yeah especially like if the government if Uncle Sam is gonna gonna back there are some things there totally I mean I'm not saying that the government is good at that or that's the government student loan programs are all really good but they do actually have a cap right they don't they they they don't go up to $400,000 they they're you know the the government the M you can borrow from the government is really is relatively small compar I think some of the doctors and lawyers loans do get up there uh sure but are they government backed loans are they government loans are they I don't know that's a good question I I think government loans have a cap I don't know I'm could be totally wrong you listen I was talking to someone from uh one of these loans refinancing companies Sofi or Ernest or something else and they mentioned the most expensive uh clientele or doctors lawyers who are still repaying their loans in their late 30s in their late 30s and probably for you know I called it I mean the credit card companies did this too I called it U voluntary and dented servitude right you're you're choosing to get the credit card you're choosing to run it out to to its maximum am it's not something that someone is making you do you're choosing to go to the school you're choosing to take these loans out right and so people the institutions can hide behind the fact that it's a choice that an individual is making it's tough to call it a choice when there is not a good alternative there are you need to build credit score you need to get a big name College get a good job so it's stacked right so that I was going there and it is you're choosing that that's what they hide behind but some of the you know when when they when they give you a credit card and they tell a college student well here's the minimum amount you can pay off but they don't tell you that if you pay off the minimum amount you'll never repay the loan you'll never repay the credit you'll never pay it off right so so regulation needs to set that minimum amount so that you actually do pay it off in some period of time or at least you're the person is made clearly aware of how long it takes them to pay off like like if you buy something with a credit card like a like um like a television a $2,000 television and you just pay off the minimum every month you'll end up paying $10,000 for that television set if you ever pay it off at all right and even if you do pay it off in Fairly big trunks you might end up paying $3,000 for the television set that normally cost $2,000 and if you're the kind of person that can't resist just buying whatever you want right because marketing is at your face you know buy this big TV buy this computer buy this car buy this blah blah and buy all this stuff then you will rack up 20 to $30,000 in credit card debt and you'll never pay it off you'll never be able to get off from under it and that's partly I think our system's problem not just the individual lack of willpower yeah that's true cuz I mean it's really hard to to fight against that I mean like you have some of the smartest people in the world uh just like doing everything they possibly can to just like you know sell product yes yes to sell a product to you and and they will play on those cognitive biases that you have to to have what your neighbor had to fit in with the Joneses to to buy the big TV at the computer they'll play on those yeah it it's a it's a losing battle uh it's it's so hard uh I mean like I think that's actually a really interesting uh regulation proposal to uh raise the the minimum credit card payment um to say like hey like maybe like the credit card payment is like $20 it's like oh I can handle $20 it's like yeah $20 a month for the next 700 years right right right right I think they did change the regulation such that the recommended minimum payment had to pay it off right because for a while there the recommended minimum payment would last in in right yeah I mean that is just I think that's kind of predatory that's terrible yeah but I mean you know kab the let the buyer beware it's a person's choice just because they didn't read the FI print they're still liable for the decision they made I mean for sure yeah uh but sometimes like the fine print is uh you know really like you need to have law degree to to interpret it sometimes also what is the last time any of you guys were to find PR I mean you probably the wrong guy I plead the fist um but actually one thing that I've been to bring the conversation back to AI a little bit Yeah so uh one thing that I've been doing a lot is whenever I like you know click agree on the terms and conditions I'll throw in the terms and conditions to chat be summize it yeah and be like hey like what do you like anything that should be aware of and uh you know typically it's like the standard leue of Lees uh but sometimes you find some good stuff but yeah I would recommend doing that um just kind of know you're agreen to um so so I think Beyond like I I want to make people better at doing that themselves right I want to I want to train them to recognize when these things are happening I think AI can be really useful as a a tutor or counselor or advice or whatever on decisions people might make or understanding problems so that they learned not to make those mistakes again now I have a lot of faith in humanity and maybe that's a little misplaced but um I I really think AI can be used as an equalizer that way yeah I I I think I think so too uh I have like massive optimism for the future uh like I think that for sure uh there's a lot of problems and it's important to recognize those problems so like we can we can fix it uh but I think that you know the problems were created by humans and then I I think like humans maybe along with the help of AI like we can solve those problems too uh so like I I don't know I think that like the future um has the opportunity uh to be extremely bright because the thing is like we collectively as humans are we build the future right so like you know we can make the that we want I would argue that our and this is an argument I made 30 years ago is that our ability to make ethical decisions around technology doesn't keep Pace with the leverage that technology gives us to make bad decisions and I just finished this book called um the alignment problem and one of the last things he says in the book is is um luckily we're really incompetent right because because the technology we've created um if we had employed it perfectly would probably have destroyed us but the fact that we make mistakes and we Bumble on things and and whatnot has prevented it from actually doing the more damage than it could actually do so maybe uh trying to end the conversation on a brighter note what are you most optimistic about what is the best case scenario uh that you're looking forward to I think there's a ton of people out there that feel as I do that that are like Mark are optimistic about and who are trying to drive uh Humanity to better better positions and I think AI can do do that but it's a tool like any other tool can be abused can be you know it's just a tool um it's in people's hearts and and minds but I think there needs to be you know unfortunately the the the incentives to use it not for evil per se but for personal gain which causes certain bad things to happen just because everybody looking out for themselves doesn't mean that anybody's looking out for say nature or anything like that right the the those things are are um really hard to to change but AI can kind of maybe level the playing field somehow can can play on our like imagine you could have ai play on our cognitive biases in a way that gets us to do good rather than in a way that enrage us or makes us afraid right and I'm not saying trick people into doing good using their cognitive biases but you know recognize when you know one things like um um the desire to be with a tribe is is is is a you're doing it because you want to be with your family and you're you don't change who you are based on that you can still recognize hey I've got facts I've got information on my side that says that this and and maybe you'll be able to fight the good fight and nudge them to a more honest an honest position and one that looks at Humanity in general and some of those other obligations other than the ones to yourself so shashan I know you're trying to like semi- wrap this up but I I I want to ask one more kind follow on question which hopefully won't turn into too much of a rabbit hole but um I'm stop me then so um you kind of mentioned how like a may be able to help society and whatnot and I'm curious do you think there's uh ever going to be um AI in government do you think that there could be uh maybe in the future um I don't know exactly how we would be able to do this with the current political process um it's going to be in Weapons Systems yeah so but like when I say like in government I mean like a leadership in government like do you think we could ever have like an AI president maybe starting small an AI Governor California um I think a highest AI an AI Avatar has the potential maybe not as a a governor or president but as a um as a voting member let's say with a certain amount of uh um Power like say human 51% AI 49% right humans have VTO power ultimately but as an agent they can um so the thing about AI is it doesn't get offended right it doesn't have the emotional things that happen in a lot of government positions it can look at the information and the data it can look at what's been effective and what's not been effective and it can recommend things that it thinks are going to be more effective or less effective and it justes there's no BS right it'll just if it's if it's if it's programmed to you know solve a problem say homelessness or or or or oral hunger or something like that well or could actually probably balance all of the problems at once it could certainly balance all the problems at once you know it can it can it can handle more data simultaneously what you're getting at than a human could exactly and it wouldn't have the same it would have prejudices based on the data it was trained with but if there's enough people understanding some of the prejudices I think some of the prejudices we have are very um very sort of isolated and personal to a particular politician or a particular thing you can just like Radiologists you could average out all those things to an ideal of what we want a a a governor to be or legislator to be or a law to be and it can it can tell you what that ideal is crunch all the data tell you how successful it can be within that system um with with real information that people can follow or not follow then there won't be any be there won't be any spin there won't be any this you know is capitalism inherently evil or is it inherently good yes there should be millionaires those are should you could say well I want I want everybody to have a a minimal lifestyle how do we achieve that and maybe it'll say no billionaires right and and and that would be the solution maybe it'll say like here's one that I think would be really radicals no taxes throughout all your life except at the very end 100% inheritance tax right everything goes to goes back goes and goes back to the government and so that way you know there's all this complaint about people on wellfare and and food stamps and STA and all those programs taking advantage of the program and not working and being lazy well you know the the the The Offspring of wealthy people don't have to do any of that stuff they they they actually cannot do anything right so how do you incentivize them to participate in the society with everybody how do you prove to to wealthy people well look let's Rize the plan Feld not just the the unfortunate people but everybody and so in your life you can't count on any of your parents inheritance you have to make their own way in your life right um and then nobody's taxed right so all the money you ear you keep you want to be a billionaire if you can get there you get there but at the end of your lifespan none of it goes to your children at all goes goes into helping Society become better I think maybe Marcus might be thinking about something that I'm also thinking about that only works until we start living forever yeah exactly well yeah that's that's a long way out and and I think that's a long way out I think that's a longer way out than other problems we have yeah um I think it's coming and there's some great science on this this there there's a number of problems that if we solve I think it's like 10 or 11 problems that we solve with human biology and and you can make longevity be in the many 100e range or augmented human computer interfaces sure sure sure I think um then maybe we get into the definition of what it means to be a person if you're living on a chip maybe a cerebrus chip that you know and this is where technology will statisfy Society right like if I've got if I'm augmented if I can afford to be augmented with a like say a super brain right that gives me access to information about everybody everywhere and and and and calculations I can do in my head and all this information I'm going to be a better employee than most people only because I have this superb brain and so I might be able to do the job of 10 people I might displace those 10 people right um and so there's that technology danger too again technology is it's really like like the the end of this this book that I was reading about my ler problem talked about technology as being this this this stallion Galloping across the across the field and and our ability to to ethically handle that technology is being this weak limed fold just trying to find its footing right and and and it's it's it's kind of true the the you know a thousand years ago we didn't have the technology to destroy the world accidentally no one person could probably do it now one person or a handful of people can do an enormous amount of damage including possibly eliminating Humanity as a whole right by just by making mistakes by using the leverage of technology and one could argue it might be happening in tiny chunks by the Technologies being used for instance for finance for politics for um you know fake fake news and creating all these all these all these things that are really believable those tiny chunks are bites out of out of Humanity's probability of surviving right and people need to be out there and that's why I would encourage people to looking out for how to counteract that sometimes not even conscious erosion that happens just just by people trying to make the best for themselves yeah