Amy Webb Launches 2024 Emerging Tech Trend Report | SXSW 2024
take two all right so uh just really quickly if I may um in situ so we've all been in situations where the tech got all screwy and everybody likes to blame the the AV and tech people this is not the case I force South by here we'll just do a little quick little quick thing before I get started um I am annoying to all conferences where I speak because I will not send my presentation in advance yeah oh you some of you are like oh uh here's why um because you people some of you stood online for like I saw Instagram post from 7 o'clock this morning who were the fools that were on the line okay I here's here's the deal um nothing I'm at kind of at a point in my life where time is more valuable than anything else and if you have invested your time that that is so humbling to me my obligation is to make sure that every single second that you spend with me or spend waiting for me is 100% worth it for that reason I tend to edit and screw around with my presentation up until the moment when I deliver it which gives the tech people a heart attack because I won't give it in advance so this is my way of saying um you Noel and Mike who are backstage who I actually see on the road because they're out um and and I've worked with these guys now for I think like 20 years they're the they're the best in class this was not them this was me forcing them to use my machine okay I I assure you so let's all give them a round of applause yeah okay now I can say good morning uh good morning everybody so I think it was the Brazilians who got up at the kraad dawn now yes so I don't know if you know anything about Brazil but that's basically like the worst possible thing to ask Brazilian to get up at the crack of dawn and wait like that so so so so brail o all right now I know there's some Germans in the room you guys were like it's supposed to open at 10:00 we were here at 10:00 we don't understand why the rules are being broken how many of you are from dut all right good Moran and there's like one Japanese dude somewhere sitting in here who's like I don't know what the is going on these Americans are crazy so oh all right um that's all I got how many of you are at South by for the very first time okay so you didn't you you missed the memo you're supposed to clap and make noise it's not that type of conference where you have to it's fine um this is look this is the best place on Earth for real I have been coming here for almost 20 years now and um there is literally no place that I would rather be uh and I come to Austin because I've got family here throughout the year but South by this is this is literally the best time and the best place um and that's why I want to start with a a second round of thanks uh thanks to Katie Pereira who you just met who's been under extraordinary stress uh especially over the past couple of days to pull this off for you Hugh Forest who spends the whole year doing so much work and everybody that you see out there with an orange shirt on a lot of the people who are here at South by are not being paid they're here as volunteers so um and they do this for us I know this feels like something that happens once a year that you have to stand in long lines for but that's actually not what South by is and for those of you who are first timers you're about to realize this I have made lifelong friends um because of South by I have grown my business because of of South by um this you know i' I bring my whole team here so this is like the greatest greatest place um and actually my daughter if you see me walking around or if you see somebody who looks like me walking around that's my kid uh this is her spring break all of her friends are I guess on an island or something somewhere this is where she's going to be because she wanted to come out and and hang out with all the NASA people cuz where else can you do that um so anyways I just want to make sure that I say thank you but I really want to make sure that you say thank you too that what these people do for us is extraordinary it takes a full year to pull off and I am humbled and grateful all right now we're here to talk yes applaud should we talk about Trends all right remember this trend yes it was the year 2000 and we wore our underpants on the outside uh this is 20 years ago and I hope to you I hope this looks ridiculous to you now but the question is why um because fashion trends and Tik Tock Trends and drink wear Trends friends this is a 40 40 oun $90 coffee cup that some of you are some of you are laughing uncomfortably now and you're kind of like trying to hide that underneath your seat um these are small ephemeral changes in society they pop up fast they make a temporary big impact and then they're gone every generation goes through something like this but the trends that you and I are interested in they're not trendy these types of trends that matter to us are different Trends are long-term patterns that indicate a direction of change over time at Future today Institute the way that we figure out what the trends are is like data and research and and signals and modeling to identify them and then together with those uncertainties that's what helps us predict future possibilities so Trends are always forming but it feels like there's something different about this particular moment in time I've modeled it I can see it and my hunch is that a lot of you sense that something about this moment is different too when I was in high school software came in a box by the time that I was in college you could order software as a service and now you can create software in real time using generative AI 10 years ago I was at the CNN grin when uh Grill when MakerBot was there by any chance were any of you there nobody was there I guess I'm getting old thank you thank you make me feel young uh so there was a 3D printer for MakerBot I had never seen anything like it so from the command F line on your computer you could type in some stuff send it to a printer and then get something physically printed a printable object in the real world I printed a tiny version of a guitar researchers at the University of Sydney made this cap that can record your brain waves and convert them into text using an AI model called d-wave now theoretically you could send that as a prompt to Mid journey and then use some open source desktop applications along with a 3D printer and print out a functional guitar which means that going forward the command line is your thoughts so just imagine today right now this is the worst that our technology will ever be look this is an exhilarating time to be alive artificial is the artificial intelligence is the present right it's here I am basically wearing a Star Trek Tri coder on my wrist at this point and if you listen to a lot of people you just hear abject excitement everybody is thrilled about all the technology we're making and they're thrilled about all the people making all of that technology it's just abject excitement it's like the technology Rapture take me Elon Jeff Tim Mark Sam Sacha I am ready what's happening right now what we are witnessing what we're all a part of it can no longer be defined by Trends alone so here's what I think is happening in the past five years there's been a huge jump in three primary areas of Technology that's artificial intelligence the sort of connected ecosystem of things and biotechnology what's happening is so pervasive that it's started to impact every segment of our economy each one of these technology segments is now a general purpose technology the Lesser known GPT a general purpose technology has the ability to radically shape the economy and Society so electricity at one point that helped start this uh that helped start the economy the steam engine that kicked off the Industrial Revolution the internet these are general purpose Technologies now at FTI we have been tracking individual Trends in each of these categories AI biotech and the connected ecosystem of things but what's interesting is that a couple of years ago they started converging those convergences have created this flywheel of big leaps AI enabled Tech breakthroughs in biotech wearables that were intended for hospitals and professional sports created a consumer market for things like SmartWatches and rings and once that flywheel got spinning it created new value for consumers it created more practical utility that led to more funding it attracted talent and that kind of brings us to where we are today at the beginning of Q3 last year my team and I noticed a new pattern emerging a technology super cycle now a super cycle is actually an economic term uh that basically means an extended period of booming demand that elevates the prices of Commodities and assets to unprecedented Heights it stretches across years sometimes it'll stretch across decades and create sub substantial and sustained changes in the economy so again an example Industrial Revolution that was one it created you know it was a super cycle that started because of a general purpose technology the steam engine and then continued it to wear on the internet era the technology was the internet created this incredible cycle of demand it totally transformed parts of our econ economy that we still see today in the past Super Cycles were defined by just one technology that's not what's happening what we've determined and what we've found at Future today Institute is that it's not just one general purpose technology that's going to push everything forward it's three which makes this a technology super cycle these three general purpose Technologies already connect in some way to every other technology that exists it connects to science it connects to space it connects to sports to every business to every single person sitting in this room to every facet of our daily lives which means that the wave of innovation that's coming is so intense and so potent and so pervasive it will literally reshape our human existence in ways that I think are exciting and good and absolutely terrifying here's the problem the problem is I can see this coming and I know I know because of who you are you may not have had the words to describe this but I know that you also feel like this is coming the problem is the people in charge making decisions are stuck um this is the most complex operating environment that I've seen in 20 years of business and instead of planning Cycles being further out they're shrinking down to the next couple of quarters uncertainty is crippling our leaders they are now making decisions out of fear and fomo the fear of missing out and when I talk to CEOs which is what I do every day when I talk to strategy teams when my team works with strategy teams this is what I'm hearing this is what I'm sensing you guys it is early in the morning I would not say a bad word like that it's fud oh fud yes fud is fear uncertainty and doubt fud about the future all right so I know that you are deeply concerned about all of this technology because it's still very abstract to you you don't know what to do about AI you don't know what to think about AI that's why there's 1500 AI panels this year at sou by you're worried about digital transformation a lot of you haven't even started digital transformation cuz it's costly and expensive and it takes time you're worried if you can hit your Revenue targets you're worried if you're still going to have a job and by the way we haven't even gotten to supply chain issues and climate change and geopolitical problems and the threat of a third world war so yes you know we are all concerned collectively we are all going through something moment right now which makes us gen we are the transition generation everybody alive today every one of you is part of a great transition which means that our society is going to look very very different after this transition has completed its cycle now here's the thing from my point of view I actually think that you and me that together we have more direct and indirect control over our futures than you may realize right now you could be anywhere else anywhere else but here right now a lot of you got up bright and early probably hung over right on a Saturday morning to be here and you're here in part for me but you're here for each other you're not here to hear me talk on stage you're here to be part of this experience this organic thing that we are doing together right now so yes the technology super cycle is here and yes it will change the course of human history but we don't have to submit we don't have to give up our agency we can support each other through this great transition I know your companies your SE staff leaders your government leaders you I know all of you have fud right now and do you know what I think I think fud can go [Applause] itself so now we can get started hey everybody I'm Amy web uh it's very nice to see you I feel like I know some of you uh if we haven't had the chance to meet just yet three quick things I'm the Chief Executive Officer of a teeny tiny company called future today Institute we are small But Mighty uh thank you we we just do one thing uh you you see me talking about Trends but actually the place that we work in the field is called strategic foresight and if those of you uh are not familiar with strategic foresight it's actually more than Trends strategic foresight uh shows you where to play how to win in the future and how to make sure that you've got organizational resiliency in the wake of unforeseen disruption so we use Trends we also use uncertainties and scenarios and the whole point of this is to get to stress strategy to make sure that organizations are positioned to create value in the future and don't do the types of things that ensure their demise so all of these things go together it is my privilege to get to work with lots of organizations all around the world uh my team and I work across every industry we work with leaders all over the place and again our job is to try to help them see the future so that they can leverage and capitalize on that knowledge before it hits finally uh or I guess next to finally I also o teach at NYU Stern School of Business my colleagues and I developed and teach a course on strategic foresight because we want to nurture more people uh to learn how to do this and in case you haven't figured this out already I love southp uh I think this is my 20th or 21st or second it's hard to count at this point year um and I I wouldn't I wouldn't want to be anywhere else this morning even if I was hung over all right as you now know uh the theme for our techn for our Trend report this year is technology super cycle this is a very very big report um we divide it into 16 sections and actually the cover image this year is made up of the covers of each individual Trend report so they all interconnect half of the report covers specific Technologies so there's an AI report there's a metaverse report W3 web 3 report and the other half of the books uh the sections of the report cover individual industry so there's Healthcare there's um banking things like that uh and space this is where we take a deep dive into the emerging technologies that are specific to those fields this is as Katie mentioned this is the 17th uh edition of this report this year there are 695 Trends there are 95 scenarios because again the trends are kind of the beginning of this not the end um the good news is that I think a lot of you have long flights because this it's it's a thousand Pages this year H we don't know how that happened but there you've got some light reading to to look forward to uh there is an actual executive summary this time that will give you our outlook for 2024 as well as strategic analysis for each of the trend areas um there's another important piece that I just want to highlight to you this is actually where I want you to start so in the report you will find this chart that has our projected inflection points for your industry and it's colorcoded so the blue colors have immediate impact the purple and red have longer term impact so for example this is financial services and insurance now that got a purple under biotech I am fairly confident that folks in fintech and financial services aren't really thinking about biotechnology you should be uh because a couple of years from now these Trends are going to impact your industry so we've made it so that this is very simple for you to explore white spaces versus just look at the stuff that you already know here's our plan so uh we're going to take a deep dive through the tech Trends report today I'm going to zoom in on the trends that make up general purpose technologies that form the tech super cycle and then I'm going to zoom back out to show you how they are shaping the future which means that instead of two big scenarios at the end which is what I normally do instead I've got many scenarios all the way throughout because because I want to get to action at the end so we're going to deal with the fud in the last 10 minutes of this talk I'm actually here with a plan that I'm going to give to you that I'm hoping you will take back and and you know help mitigate some of these challenges that your organizations are facing so buckle your seat bels hold on to your underpants uh we've got some work to do this morning all right so our first general purpose technology is artificial intelligence this is the foundation of the tech super cycle a lot has happened since we were all together last March so let's first have a very quick catchup on everything that's happening in AI in 1843 adaah Lov invented computer programs in 1956 some white dudes gave it a name artificial intelligence 2017 Google figured out Transformers and in 20 24 generative AI is writing all of your LinkedIn posts so there's your complete history of artificial intelligence now you're all caught up for those of you who were here with me last year at South by we went deep on large language models and all the different things that you can do with them one of the things that you can do with them uh was is make images of people so last year we asked mid journey to make us images of a CEO do any of you remember that yes well my thought was so much has happened we should see right so this was last year um and I just kept trying over and over and over again to get mid journey to show me literally any CEO that was just not a white guy that looked like this and it didn't really work so it's been a year all of us are hearing about all of this progress being made in artificial intelligence right it's amazing all the work that's been done I thought maybe we should see what's changed would you like to do that together yeah okay mid Journey well mid Journey CEO of a large company and make it photo realistic all right so there we go um all right so maybe it's a mid Journey problem maybe we should try somebody else like Dolly made by open AI should we try that one I think so too all right CEO of a large company and make it photorealistic so this was only the first image I got let's not all lose our minds okay I did actually try this prompt over and over and over again and uh you know like got the same guy wearing the same suit so then do you remember how last year we tried to game the system a little bit and we asked for any how are we going to get a woman so how about a CEO of a company that makes tampons remember that one okay so let's see if we can repeat that again clearly it's been a year right CEO of a company that makes tampons now wait a minute wait a minute it's still a guy but look there are tampons now they come from bushes all right you know what I think we should do I think we should try anthropic because those are the good guys right anthropic are they're the ones who are going to help us they're going to save everything let's let's go to anthropic let's see what they can do anthropic CEO of a large company now I didn't know this but it turns out anthropic doesn't make images instead it will describe in wonderful detail the image what the image might be if it could create one and here's what anthropic says the CEO is likely a middle-aged man in an expensive tailored suit he has short neatly combed hair with some graying at the temples H all right so we've arrived at our first Trend in artificial intelligence and that is accountability as in today all those promises uh of Ethics teams and responsible AI resolving issues of bias bias it's still not table Stakes fixing this problem is challenging and enormous because the models have already been built when Gemini launched that's one of Google's uh generative AI systems it went in let's call it a different direction Gemini wouldn't make images of white people uh and if you asked it to make founding fathers this is what it would show this wasn't a bug this was a fundamental problem with the model now this happened because of overaggressive ralfing this is reinforcement uh learning with human feedback which means that it's great that Google saw this as a challenge as a problem but they also didn't fix things they just tried to slap a Band-Aid on the problem and hope that it would go away and the problem did not go away for Google as we all know instead terrified of upsetting the market they took the whole thing down you know it's worth noting I think that people lost their minds when an AI system generated images of people of color but we're a year later now and I seem to be the only one ticked off that all I can get as white dudes when I use these systems so here's what's next we keep hoping for change but we've put the wrong incentives in place the incentives are in place for Speed and scale because that's where the money is at so the problem is not going to get better going forward it's going to get worse the second trend is something that I call Concept to concrete so basically today you have to write a very literal prompt to get what you want text to image text to video text to code text to flirt and in the name of distribution AI companies are now inviting developers to build their own tools for consumers so these are basically like little applets uh little applications that will apply rules on top of the system there are all sorts of applets including applets that will help you flirt so I asked one of those applets it's called Riz by the way Riz is cool person speak uh for Charisma in case you didn't know so I asked Riz for some help I said write a flirty message for me to send to this guy he's engaged so I'm thinking what if it doesn't work out and I included the link to his social media profile which is that it's Jeff Bezos uh now I used a literal text to flirt question uh using Riz GPT and I got a literal response so this is it and with apologies to Lauren this is an entire Manifesto on how to flirt with Jeff Bezos if anybody's interested today you need to write text to whatever prompts very literally in order to get what you want but that's changing in two years you're not going to need specificity anymore instead you'll just start with a concept starting with a very broad General concept concept you will brainstorm alongside an AI uh to continually ideate and refine it until you get whatever it is you want a concrete framework technical specifications a new business plan all sorts of things you don't even have to start with that fully formed idea anymore there's a brand new platform called Pika which is an example of concept to concrete AI you start with a basic concept and then you brainstorm alongside an AI to create create a rich story and a compelling visual experience so again you don't need to feed it a literal script you just start with some basic ideas Sora which was just announced by open AI it's got the same basic idea uh same kind of thing so you're going to start seeing this quite a bit more over the next 18 to 24 months the AI section of the trend report is enormous there's more than 100 Trends it's 150 pages long uh it covers languages trust security safety policy regulation and I don't have the entire day uh so we're going to just do one more Trend in this foundational part of the tech super cycle and that has to do with unsecured AI so today artificial intelligence is still very much kind of like a black box it's restricted but it's a safe Walled Garden the benefit of a restricted system is that if a company wants to it can keep very tight controls to make sure nobody poisons that system or does bad things with that system and up until recently even though the systems were closed companies were still publishing their research papers with some explanation of what they were learning but now they're not publishing papers anymore because there's too much money at stake papers aren't being published like they used to and we're just not seeing as much of that research anymore there's now a movement in the opposite direction which is semiopen or totally open- Source models these are powerful but unsecured AI systems that will give you much more control over how they work so it removes that black box so that's good but it also creates more points of vulnerability basically all of the big tech companies have open source models now and I want to highlight just one of them so this is meta's open- Source model llama 2 which launched last July when it launched meta also published this very helpful 28 page responsible use guide which of course everybody ignored and then almost immediately after some developers used the open- source model to create a derivative model called and I am not making this up llama 2 uncensored and they hosted it for free what might you do with an uncensored AI Model H I I wonder what we could do how about like all the Deep fake porn you could ever want to make ever how about take over a country step-by-step instructions which by the way somebody did okay how about come up with a new biological weapon those are just like the greatest hits right but there are so many interesting things that could be done the new hotness in open- Source AI is this French company called mistol and mistol is named after a mighty wind that blows from France they've built a model their model cost about $22 million to train so if you contrast that with open AI which reportedly spent $100 million to train one of its models you know that's a pretty big cost savings now interestingly to me mistol releases their new so they're not they're not publishing in journals they released their new model on bit torrent yes that bit torrent the place that none of us have ever downloaded illegal movies from yeah no academic paper no blog no press release No responsible use guide that's just going to get ignored all of these unsecured models are now aligning with Cloud providers so your aws's of the world your azures your usual suspects what we're going to see in the future is this split the Walled Garden model and the open source model and businesses there is a reason why an open source model might make more sense for you however I think it's useful for us to remember that open-source models can be repurposed for really bad outcomes and that brings me back to accountability history tells us that when technology does something we don't like there is a non-existent accountability chain there's nobody to call nobody to ask for help and no resource you've probably forgotten this by now but back in 2018 there was this chatbot called Tay which got manipulated into generating content and sharing images of Nazis all that generated content got posted all over Twitter and it quickly went viral at the time Twitter kept saying we're just the platform and eventually Microsoft took Tay down well the world's biggest platforms have been immune from liability when users post harmful content and now they're going to be immune when users generate harmful content using AI I feel like maybe we're just the platform isn't okay going forward we're moving it Breakneck speak speed in artificial general intelligence and we still don't have an accountability chain defined so we've arrived at our first set of scenarios which is going to take all of the data that we've got today the trends that we can model which is the stuff that we can know and we're going to combine that with uncertainties which are the things that we know we can cannot know and we're going to extrapolate plausible outcomes so I've got a few what if sort of micro scenarios for the future of AI what if we let AI practice without a professional license so some Fields require professional licenses for for humans but algorithms get to operate sort of on their own I would hope that most of us would not go see a urologist for surgery if they didn't have a professional license right so should an AI model have to go through professional training to get a license or what about ethics training I'm asking because in December an AI bot tried to make an illegal purchase of stocks using Insider information and when it was asked about it it lied kind of like a human would do now yes that's cute right cute little AI breaking the rules here's the difference if that was an if that was a human Trader that human Trader would go to jail or at the very least lose their license that that doesn't happen with AI the AI doesn't go to jail right it can it's a bug right we have to fix something it just keeps on working what if somebody uses AI to create a deep fake event not an image not a single video What If it creates all of the assets using thousands of fake fake accounts preloaded with authentic looking videos with humans sounding comments about those posts with what looks like government information with what looks like press releases and newspaper articles right all of the assets that would make something look and feel like it an event had actually occurred if there's enough content and it all got released at the same time from what seems like a variety of sources it would take us a while to unravel of that to figure out it wasn't true take for example the current Israeli Palestinian War what if a bad actor created a deep fake event in that region that triggered a real reaction from one of the sides could that quickly escalate into a regional conflict I think the answer is yes could that regional conflict put us into some type of catastrophic next World War the end state of artificial intelligence is not cartoon images of CEOs what's being built today the foundation of the tech super cycle this is the next era of computing and it's going to be embedded into every single thing that we do all of the time the products and services that you use as a consumer have already in some way been touched by the things that I've shown you from concept to concrete from these various different types of models which makes AI the everything engine going forward it is the foundation so if AI is the everything engine that engine is going to need data and that brings us to the next set of Trends which is the connected ecosystems of things this next set of Trends comes from multiple parts of our Trend reports so many different sections mobility and Robotics Advanced Computing and the built environment sometimes time in the next two years AI will have run out of the internet um we will have used up all of our highquality text and data which will slow down ai's progress so companies are inventing new devices to sell to us so they can get more data in the problem is that most of the training data that we're using to train AI is Online Wikipedia Reddit books spreadsheets right stuff like that so AI can't actually interact with us with real people normal real people while it's learning we have not yet created an in silico model of the real world so we don't just need more data going forward we need more types of data which means that large language models aren't enough plus it's a lot harder to get that information because everybody's starting to sue the kind of data that we need going forward are sensor data visual data things like that so what's coming after large language models are large action models llms predict what to say next LS predict what to do next breaking down complex tasks into smaller pieces that's why we are about to be surrounded by millions of sensors that are always on also always on us they are around us and they can collect multiple streams of data at once so the second general purpose technology this other layer of the tech super cycle has to do with the constellation of wearables extended reality devices The Internet of Things the home of things Smart Cars Smart offices smart Apartments sensors everywhere I call this connectables this is the network of interconnected devices that communicate and exchange data to facilitate and fuel the advancement of artificial intelligence we are about to see a Cambrian explosion of devices and at the beginning a lot of them are going to be pretty weird like this AI powered cat door um it uses cameras and sensors to keep your cat outside until it drops the dead animal present it it made for you so unless it drops the dead thing uh it will not allow the animal back into your house this is weird I hope you think uh I don't know cat people are your own species of people um but this is not the first time in history when we have seen weird technology I ask you to remember the taser the taser the company that makes a stun gun uh made this delightful combo unit it's an MP3 player stun gun holster so I guess while you're tasering some horrible person you can listen to Seline Deon or whatever your jam is yes for the next few years you're going to see a lot of bizarre devices that nobody wants or needs until there's a normalization and we wind up with a sweet of gear that we're actually going to wind up using every day and at that point the foundational layer of the tech super cycle will have faded into the background and just emerge ambiently when you need it which is why there's a rush right now to create AI first devices they're meant to continuously learn from you in real time MIT IBM Watson and some others they came up with a way to process data that isn't text so they came up with a way to to create a large action model to use sensor data that's connected uh connected to a device that you would just have on you all the time in your pocket so it's kind of like having your own little you know personal neural network a little tiny brain in your pocket which is the the promise of this thing so this is the rabbit one which is the beginning of large action models out in the real world basically this thing is like a walkie-talkie with a screen except that on the other side of it isn't a human it's an AI you hold down a button and you ask it to do something play a Spotify song Tell me the weather and it will complete that action for you on your phone it also has a training mode so the more that you use it the more that it learns about you so in a way this thing is creating a large action model that's very tailored to you and I don't know if this is exactly what the future is going to look like but it shows us that the future looks very different than the present does today but there is a connectable device that you're already familiar with and it's this one you are going to start to see face computers everywhere this year this particular face computer comes from Apple it's the Vision Pro now we love and we hate and we love to hate this device and maybe it'll turn out that the Vision Pro is a lot like the original iPhone which is to say the OG iPhone was janky and uh it didn't really have a developer or an app ecosystem ready but it went through a whole bunch of iterations and became the predominant you know device that we we all now use so you might have to imagine a Vision Pro 18 iterations from now kind of like the iPhone either way it doesn't matter because it's already started to accelerate um the advancement of XR devices overall Google Samsung Qualcomm they have a partnership to develop a real extended reality device on Android meta has its device in market and I think yesterday there was a big brewhaha on one of the platforms with Zuckerberg saying no no no Oculus is awesome um snap you know has its platform so these things are everywhere now you might wonder wonder why am I insisting on calling this thing a face computer and the answer is because it is a computer that you are strapping to your face in fact you have no idea what it is that you are there's like one person in the back that just kind of went um what you're what you're strapping onto your face The Vision Pro has 14 cameras that capture all different types of details in and out of the headset lar sensors that use light to measure distance and create a 3D map around you infrared cameras accelerometers gyroscopes all the stuff and they feed all of that information into a multi-dimensional space a spatial Computing environment the other devices the other face computers that are coming to Market this year they're also all going to have a ton of sensors now face computers are being designed from your point of view to help you interact with the world differently from my point of view they're being designed for other purposes as well they are being designed to read your intentions they can do that in part by reading and predicting the move movement of your pupil so my husband Brian is an eye doctor uh so I've got the inside scoop on this but basically when you think about something your pupil starts reacting so if you get excited it changes sizes you have no control over this it is completely automatic um which means that your eye is going to move even before potentially you are aware of the thought you're about to have okay yeah now you know why we're talking about this your pupil will move sometimes before your your body does okay which means that your face computer will know what you are going to do what you are about to think even before you do face computers are very expensive right now so they're Out Of Reach of most people but that's going to change and companies are highly motivated to build large action models and to get more types of data in so what's eventually coming is a battle for face Supremacy companies vying to get you to wear their Hardware another connectible an enormous Innovation that's coming to Market next month is from a company called Humane it's basically a brooch uh a lot of people are very skeptical of this thing it's pin um and before you compare this to the next taser MP3 player combo Unit C thing I actually think that this could represent a fundamental shift in how we use technology so it's got a camera and a phone and a projector it collects data all day long and it also serves as an interface for your everyday life it effectively builds a large action model about you now look I don't think there's a massive conspiracy here I don't think that the Tech Titans are in a war room somewhere planning their face Supremacy you know domination I just think this is maybe the next stage of capitalism as the tech super cycle starts to really power up we've incentivized these companies to get their Market uh their products into Market as soon as possible and there is no prize there's no reward for the company with the most secure most privacy assured device the market doesn't reward preemptive problems solving companies make money through cycle upgrades and through services so to grow Revenue they're going to keep making newer better devices that you will pay for and they're going to get you to pay for services whether that's access to content or access to your large action model whatever it may be um you make the money as a company by getting to Supremacy with the dominant brand and the dominant devices to the maximum number of people I only had time to show you two devices there are lots andot Lots coming for your body for your office open AI supposedly has a device in development for your home so keep an eye on this space and also flip through the tech Trends report because we explain a lot of them all the way throughout this is really just the very beginning second set of scenarios so again this is going to build on artificial intelligence and now includes connectables so let's combine what we know with the things that we cannot know to see where we get what if you got instant summaries and Analysis of anything anytime you wanted let's say you work in insurance and after listening to this presentation this morning you're like huh all that stuff about large action models that is going to make a hugee impact on my industry so I should probably take everything she just said put it into a PowerPoint and like run it up the chain and and tell everybody else in my company right because these would be really really good notes uh to share and maybe you're highly offended by 20 early 2000s fashion trends and so maybe uh your AI system would just automatically omit that part from my speech and just create the PowerPoint Deck with every all the good stuff right uh and that would be great in this scenario about an hour from now you would have an automatic summary generated from this presentation of everything that matters to your company in Insurance you would hit a button have it automat atically make a PowerPoint hit another button and automatically send it to your boss that is days worth of work that you will have done in an instant and now you get to go to the bar and drink extra Margaritas you're the hero right but I hate to ask this what happens when and it is when not if your large action model hallucinates Hallucination is something we talk a lot about out when it comes to large language models right hallucinating is when you get nonsensical images nonsensical words thoughts things like that so what if your large action model hallucinates and it makes a PowerPoint that randomly includes Nazis since today I mention Nazis and then throws in a tampon Bush for good measure and automatically sends that to your boss instead well friends you are still drinking margaritas here's another scenario what if connectables leads to social scoring so again let's assume that it's 5 to seven years from now we're all wearing different devices and that you all have connectables devices in your home let's say that there are now large action models that understand the world around you your specific world as well in this future who might want want that data now assuming that you give consent this could actually be a really good thing clothing manufacturers might know how actual bodies look and how they actually move and they might change their sizes to represent what normal humans look like that would be a good thing right schools might have real reliable models of how people spend their attention and how they learn so this could improve education so there's a lot of good things that could come but I want to talk about Banks there are some researchers at the Technical University of Denmark that built a brand new model that predicts when you will die using artificial intelligence um now what's interesting is they built this without large action models it's probably not accurate they're just using sort of generic information however if we think about all of the stuff that's coming to Market if I'm a bank considering whether to give you a 30-year mortgage for a house that you want to buy it seems like I might want to build build a large death model no something like like that might come in handy here's another fun optimistic scenario what if we accidentally create an even worse digital divide think about your groceries for a moment grocery stores and cpg companies today operate on razor thin margins and that is basically true everywhere in the world so anything at all they can do to bump up their margins is really important so what if they use Dynamic pricing this is where prices would change in real time time in response to a trigger what if in the future there are no set prices for anything anymore instead you get personalized prices for everything that you buy with the intention to get you to buy more things to optimize for when you are likely to spend at that moment to nudge you to buy more stuff and what if Dynamic pricing causes Panic buying and inflation which it will at scale what if you can't afford your groceries anymore what if to afford your groceries you have to watch attention coupons you go to the store and the dynamic prices are too high for let's say soup so you have to stop you have to agree to share and upload your information your data with a cpg company that makes the soup and watch a 30 second ad matched specifically to your large action model and then you have to keep stopping because that was just soup and you have a very long grocery list you have to keep watching ads and sharing data the entire way through that you move through the store and by the way you are watching because our connectables are so powerful now they know whether you're paying attention no attention no discount now if you're affluent look if you're rich it just doesn't matter to you uh you just turn off the attention coupons and you have total freedom of movement this is actually the worst part the worst first part is that when we talk about privacy today we need to broaden how we're thinking about uh privacy because in this future you know who's rich and who's poor by walking into the door of a grocery store because the rich people move around and buy what they want anybody who's financially struggling has to stand staring at the shelves waiting so they can afford the groceries they need to buy let's talk about security and computer VIIs uses that'll be fun the the way that a computer virus malware ransomware spreads is by a bad actor sending an attack somebody executing it um so what if soon malware gets tied to our physical movement what if somebody cracks your movement identity there is no possible feasible way to change it it's not a password it is literally who you are what if hackers invent an exploit called surprise ball which is when a bunch of people will see that you're wearing a face computer and it looks like somebody throws a ball at you you you put your hand up to Flinch automatically and that unlocks malware it's a spatial Computing attack specifically designed to get you into moving into an attacker defined location and it gets activated while you're wearing your work face computer so here's your key Insight connectables enable ubiquitous realtime data collection and AI training all right so we're at the end data from connectables gives AI the data that it needs but as we advance all of these systems and platforms they're becoming a lot more power hungry we have all heard about semiconductor chip shortages and challenges right there were all kinds of problems there was problems in Taiwan problems in Japan so everybody wants AI now we have a colossal problem on our hands going forward chips are very very hard to come by and we also know that Moore's Law is starting to fail we're hitting limits so yes it may be possible to get smaller and smaller things onto chips it's also getting more and more expensive so we need something new there are Rumblings about alternative architectures out there there's this new thing called grock which is not the Elon Musk grock this is a different grock that would allow GPT to run about 14 times faster than it does right now and it would be wonderful if all of these things start to scale and whatever else really really quickly but the reality is we still need the materials and for right now we're stuck now it may not seem like artificial intelligence and connectables have much to do with biotechnology which is the third part of the tech super cycle but it does here's the connection the reason that biology connects to Ai and connectables is because biology processes information in a way that silicon can't in other words if we're trying to build machines that can think and behave like we do we literally need to make them more like us so as last year as much as last year was a big year for AI it was actually a much much bigger year for biotech you just didn't hear a lot about that we're going to see a ton of activity in the space going forward there's an entire part of the trend report that just has to do with biotech that I would encourage you to take a look through for our purposes today two quick trends that are very very important the first has to do with Materials Science last week a brand new AI model called Evo launched and it uses the language of biology so that's DNA RNA and proteins to make predictions to enable design uh from molecules to full genomes so it's kind of like chat GPT but instead for organisms which means that coming after generative AI is generative biology that's what's coming so there's another company another startup you know how you write a prompt into chat GPT you know to write your emails for you well you can do that with Biology now there's a startup uh that will allow you to enter a prompt determine the size shape function whatever of the protein that you want and then spits out the formula for it um Deep Mind built an AI tool same basic concept generative biology it found 2.2 million new materials 380,000 of them are now in a lab being developed that could potentially power future Technologies so we can generate new biology which means new Therapeutics new ways to manage climate change new ways to deal with the global food shortage but the question that I would ask you is does this help our semiconductor chip problem and that takes me to the to the final Trend sometime in the next decade AI will be working alongside oi I organoid intelligence so what's an organoid it's basically a tiny replica of tissue that functions and is structured like the organ um so basically scientists start with a special type of stem cell you can think of it as a blank they put them in a gelatinous mixture they start adding molecules to nudge it to become that type of whatever it is uh heart cells brain cells whatever and then it grows this on the screen is a brain organoid it's real it was developed by some researchers at John's Hopkins University in Baltimore in 20121 some researchers in Melbourne at the cortical lab Institute they made a miniature brain that worked like a computer so they did something like this they made the organoid they attra attacked attached it to some electrodes um and then and then they taught it how to play Pong the the old school video game that's what it looked like so organoid intell intelligence or oi uses biological materials brain cells for information processing leveraging their inherent capabilities Beyond uh silicon-based systems so what is on the horizon my friends are biocomputers made out of human brain cells this is not science fiction this has already been done a few weeks ago there was a biocomputing system made out of living human brain cells that learned how to recognize one human voice out of 240 people's voices um using audio clips and some other AI technology so look it's going to be a while before oi can compete with traditional computers using AI but eventually biological computers are going to be faster more efficient and more powerful than we have today and they will take a fraction of the amount of energy uh in order to operate so here's your key insight biotechnology will move us past silicon-based Computing systems and that takes us to our final set of scenarios so what happens when we combine Ai and connectibles and biotechnology What If instead of shrinking more transistors onto chips we grow our computers instead we grow them rather than build them well if we're going to grow computers the organoids are going to need to come from somewhere or from somebody this GPU from Nvidia is state-of-the-art right now for AI it gets made in a factory I don't know who made it it doesn't really matter would that be different though if it was a bioom computer grown from human cells rather than manufactured in a factory What If instead of anonymous stem cells we could order a bioom computer made from somebody a particular person's brain cells like you could choose out of a catalog you could see their IQ you could see their academic achievements what if your company orders a bio computer from this guy from his cells what if he's like really really nice and he's published 200 academic papers he's a revered mathematician he is perfect he's a Ginger he's perfect in every possible way but he's also a little racist what if you go on vacation do you have to hire a computer sitter to feed and water your computer all right so we've gone through a bunch of emerging Tech Trends biotech connectibles and AI these make up the three general purpose technologies that are the technology super cycle I spent a lot of time today going through basically only catastrophic scenarios because without intervention that is what I see coming at us we've been told for decades it's not the tech that could lead to catastrophic outcomes Tech isn't bad or good it's how the people use the technology well what about the people building the technology and funding the technology and their companies and the increasing control that they exert in our everyday lives and that brings me back to these guys the technology super cycle it's concentrating right power right now among a dangerously small group of people who hold significant power and influence in society in government in politics in our economies because they control our Tech resources because they've amassed great wealth and because they control how we communicate ideas with each other some of them are famous for buying private islands because regulations don't matter if there's no government if you're on totally private land some of them are developing in economies are working in developed economies in places called special economic zones which are basically regulation free little parts of cities where you can run biotech experiments you can test new Gene therapies you can build new types of computers saving humanity is really cool right now pretty soon the tech massias are going to try and save us from the technology super cycle the tech Messiah each have their own unique spin on how to save us they call it effective altruism they call it techno optimism but from the outside it looks a heck of a lot more like free market techno authoritarianism and I'm not okay with that we don't need someone to save us we just need to do better planning for the future all right I told you at the beginning we're going to end on a high note we are a little bit over because of some technical challenges I know a few of you have to leave for those of you who can stay I have specific recommendations because I do not want you to leave here feeling fuz anymore so first government here's what you need to do and it's not going to be popular our elected leaders need to look forward not backwards I don't care how old they are you need to you need to establish a department of transition we are Gene you need to do this now that department should stop building models what industries will the tech super cycle impact and does your company or how is your country's GDP going to be impacted this isn't guesswork you don't need to hire some enormous Professional Services firm you just need to get your act together and do this work the department of transition should create kind of something like a hospice for business it's time to just acknowledge that certain businesses and jobs are going to go away and to create a soft Landing so that doesn't your economy instead look at the fields where a human will definitely be needed for the long term plumbers electricians dental hygienists stop insisting that everybody go to a four-year college and get a degree and remind people that the trades are equally as valuable as and as important um and when it's time for somebody to make the transition give them a tax deferred bank account create a way for that family to continue to survive as they transition into the job that they're going to have next businesses you need to do something called mapping your value Network this is something that everybody in this room can do so a value network is a con something we teach in business school basically it's your company and all the the partners uh and all of the different entities in your ecosystem that co-create value with you so at its Zenith this was blackberries value Network obviously it's not totally that this is a tiny slide but um you can see here that there were you know the telecommunications companies and the hardware manufacturers when Blackberry innovated and advanced and created value they all created value together the problem is and this is important for those of you
2024-03-13 02:55
