(upbeat music)- Hey, hey, welcome to another episodeof "Marketing Against the Grain"I'm your host Kipp Bodnar,the Chief marketing officer at HubSpot.I'm joined as always by myco-host Kieran Flanagan,who's the chief marketingofficer over at Zapier.And we're coming to you,mine's blown out of breath,completely excited about what's happeningin the world of AI,auto GPT and Open AI aresending the world reeling.We're gonna talk about what Auto GPTs are,the impact they're gonnahave on growing your businessand doing the marketingtasks that you need.How they are forcing companies like Googleto reinvent themselves.We've got some fresh,fresh data about AI andAI adoption from scale AIthat we can't wait to share with you,as well as some rumors aboutwhat's happening with Meta's push into AI.There is so much we haveto talk about Kieran.My head is spinningwith what's happeningin the world right now.- Oh, Auto GPTs are the hot new thing.Everyone is going to betalking to you about these.We aren't just happy with having like onesuper intelligent assistant.We humans want to like threadand connect all these thingstogether because we're solazy we don't even wantto have to come up with the prompts.We just wanna like have all these thingsdo all of the work for us.- So Kieran, first I wantyou to tell everybodywhat Auto GPTs are,high level, how they work.And then I had a terrifyingrevelation over the weekend thatI will share as a follow up to that.- We're gonna get into that'cause I actually have itrunning on my laptop right now.And so I was listening to a,an incredible interviewby an AI scientist,a very well known AI scientist.And he talked about the factthat for years and years and years,researchers have said there islike three fundamental thingsthat humans should notdo when they're decidingto develop like AI superpowers or just AI.And he said you should neverteach the AI about humans,because it will learn tomanipulate the humans.You should never teach the AI how to codebecause if they learn to code,that unlocks like theability to self-improveand you should never give AIaccess to the internet, right?- We have failed.We have failed miserably.- Auto GPTs like take allof that which GPT like threeplugins had access tothe internet could code.But it takes all of thatand it accelerates itbecause it gives it memory as well.And so you can give Auto GPT a goal,actually you end up giving it five goals.You tell it to, you describe it,you give it a kind of missionand then you give it fivecore goals to complete.And what it does is itstarts giving itself tasksand it starts offloading those tasksto other GPT agents whoactually run for you.So I've got a version of itrun on the laptop right nowto build us an outline for this show.And I give it a bunch of different goalsto go research things,find expert point of views,like go tell me their LinkedIn profilesand it's given itself,recursively given itselflike all of these differentprompts to go do that work.And so you can imagine in thefuture I can just give thisthing a goal and then I don'treally really have to worryabout anything else after that.- You're just gonna go sit onthe beach, you know, hang out.- Yeah, sit on the beach.It starts to like work withall of the other GPT agentsto actually fulfill that goal.And I will say that a lotof what I'm seeing online,like in terms of the example, so I really,I think our cherry pickedbecause I can tell youit is still very hit or miss.- Hold on, hold on.Before you get into the examples,I want people to understand,the Auto GPTs, it's not a company,it is some code thatwas released open sourceon a platform called GitHub,which is where developerscan basically store code,share code, build on each other's code.And so somebody posted thisAuto GPT code as an experimentand it quickly skyrocketedto the most popularcode repo on GitHub.And so it's not like you can go in,you can go use Auto GPT from some company.It's like you have to be adeveloper, install auto GPT,run auto, run the code,do all those things.So it is still far,far away from prime time froma user experience standpoint.And you're also I thinktrying to make the pointthat a lot of the use cases that peopleare showing are still like verycherry picked and there's a lot of stuffthat auto GPTs aren't evenclose to being good at yet.Is that where you're going?- Yeah, I gave it sometasks over the weekendand it got stuck in recursive loopsand also the output is like wish washy.But there are some reallycool ones like someone builta to-do app and this is really coolbecause you add your liketo-do item to the listand as soon as you add it to the list,Auto GBT kicks off andstarts like completingthat task for you.So you just like, it's aself complete and to-do list.One of the other one.So I saw one where,which is kind of likegoing back to what I said.Now it has access to the internetand it can actually store memoryand write Python scripts.I saw one where it was recursivelyimproving and debuggingits own code and I'vedone that for writingwhere I've told it to writeme an article or write me,edit me something and thencorrect itself and do betterand it will do that.And it's doing that for code.I've seen examples for customer service,like not implementedyet but like mapped out,which you just give it the goal.Like if you give it the goalto give everyone a five starcustomer service experience,then it will start to recursivelygo and like read throughtickets and get back to people via emails.And like start to actuallyensure that it has like giveneveryone that adequate of a service.So some of the potentialuse cases for this,if you thought Chat GPTwas pretty disruptive,like this is pretty incrediblebecause you actually are going,you are actually going onelevel up in terms of the waysthat you can set those goalsand you can give it biggerand more complex thingsto actually go complete.- I love that.And when people talk about Auto GPTs,they also talk to them as baby AGI's.They talk about it asbasically the front runnerto real artificial intelligencewhere the AI can think,feel, have emotion, allthose kinds of things.And so I was like thinking,I was reading aboutthis stuff this weekend,Kieran and I just,I got kind of sad causeI was like I don't know,do we need machines tohave that much emotionand maybe it's just me as human,I'm not sure that emotion is a featureand not a bug in humans.- Well, did you see theSundar CEO of Google?60 minutes.- 60 Minutes just PR tour.It's like dude you're late to the AI gamein terms of commercializingthe technology,because I think you're scaredand so you're going ona 60 minutes media tour.Yeah I saw that.It was kind of crazy.- But I think he is scaredand let me tell you why he's scared.So first of all, he did say,and he says something likevery interesting within thatinterview which is he was asked like,you do not understandhow these things work.And his reply was,well we don't really understandhow the human brain works either.I'm like, well, the human brain mayor like for the most partin our history has not yet destroyed us.And we do not know if thisnew entity will or not.But one of the fascinating things I heardabout over the weekend in relation to AI,there's this essay called"Meditations Moloch" and it talksabout like it's a metaphorfor destructive systemswhich really kind of is howincentives leads to sub-optimalresults for all, andso I'll give give you.- That's how I ruined my life.- Yes.But this is an incredibleexample and again I'm stealing onit from that interviewthat I saw on Freedan,which is Instagram filters.He gave a great exampleof Instagram filters,which is everyone was onInstagram and then they integratedfilters and then a certainpercentage of influencers woulduse filters to start to make themselveslook better to grow their audience.And then everyone is forced,that leads to sub-optimal results,'cause then everyone isforced to use filters to runthe influencer play and then everyone,it's all normalized and theworld is a worse off placebecause now we have theseunrealistic expectationsand what people should looklike to grow their following.So everyone loses in that situation.And I think what's happeningto Google is they may actuallybe like, this is way too quick.Like he actually had lotsof parts of that interviewwhere he's like, this may be too fast.We're trying to be a little bit thoughtfulbecause we don't think humans are readyfor this but they'rebeing forced into the gamebecause Chat GPT has completelytransformed how peoplewant to use those tools like searchfor things and that'sGoogle's core business.And so people are gettingdragged into the raceand hopefully it willnot be like sub-optimalresults for humans.- Well what you just describedalso is marketing in a nutshell, right?A marketer figures out some strategy,some tactics that works andworks really well then everybodycopies it and then stopsworking as well, right?- [Kieran] Exactly.- That's basically what our society isand that's why you and I obsessabout being first movers, right?Because when you have afirst mover advantage,especially in something like marketing,you can take advantage ofas much time as possibleof that new strategy,that new tactic beforeit normalizes, right?And what we are arguing is thatthere might be some extremenormalization and downside use caseswith AI potentially, we don't know.But what we do know isthat with this technology,I believe it is goingto completely transformhow we think about marketing,sales, customer support.You already gave likethe goal of Auto GPTsbeing able to be customer support agents.What I wanna hear from youKieran is like what do you thinkif you're a marketer today?Like what do you thinksomething like auto GPTsare going to do for you overthe next 6, 12, 18 months?Like what are the big things.- I think they get better?So I give it a project overthe weekend to research a topicto break down that topicinto like different subtopicsand kind of go through all ofthe places that I could createcontent that would be differentiatedthan what existed todayand gimme summaries forthose pieces of contentand then tell me like oneunique thing I should tell,I should have in that contentto make it share more shareable.And the results were pretty awful.But actually I think in thefuture if that gets better,it's just replaced a bunch ofwork that has to happen acrossa bunch of different teams today.So you can like thread things togetherin really cool ways.And so I think inmarketing you can go from,first of all you can go frompersona research like reallycraft the persona to like thestrategy for that persona,to the summary of how I buildthose different playbooks.And then I love that actuallyto how I would implement them.And the other thing I startedto do was I said to it likegive me the top a hundredways that famous SaaS Primeshave grown their business,like top of the funnel forcustomer acquisition tactics.Then I said to categorize them by tactic,then to summarize itand then I said give me the top 10and find me 10 people on LinkedIn.'Cause I wanted to see if itcould like understand people,find 10 people on LinkedInthat I can call upand ask about those tacticsbecause they have done them beforeor they're experts in those things.Like provide me the LinkedInprofiles and then write mea draft email to be ableto send to that personand introduce myself andactually ask them for some advicearound these tactics.Now, I will tell you.- I was gonna say, how'd that go?- I think that you'llactually be able to do thatwithin the next year, maybe withinwithin the next six months.How did it go?It got stuck in a recursive loop.And the weird thing is itstarted opening up like lots of browserson my laptop like,like the Safari and itstarted like scrape,it looks like it's likeanalyzing the contentand I'm like what does it doing in here?And you can see like it'slike opening this thing,tried this thing.But it's like you can threadtogether entire workflowsin a very seamless way.- That is the magic,if you are running a business,you're running a marketingteam and you're thinking aboutthe future of AI, especiallyauto GPT like technology,what you're saying is AI isgonna be able to string togetherlots of different tasks in a workflowand automate those in a waythat really could only everbe done by a human beforeand it'll be able to automatethem and in some cases I thinkdo it much faster than a human,which is pretty magical.And I think Kieran gave yousome great examples aroundresearch positioning,I think like generatingrough drafts of plans,all of that kind of stuff.'Cause at some point we are going to useAuto GPTs I believe to figure outthe new horizons where we wanna invest.What is everybody else doingand what is a viable strategythat's not what everybody else is doing,which is something that Kieranand I believe deeply in.I think we are going to endup having to use Auto GPTsto differentiate fromall of the normalizationthat they create, whichis gonna be interesting.Kieran, I wanna get your feedbackon something as it relates to this.I sent a tweet out earliertoday and I wanna see if youagree with me or not.You ready?- Hit me.- I said marketing nowversus in the future.And this is as it relates toAI and Auto GPTs etcetera.I said, today we think alot about content marketingin the future we're gonnahave ultra micro webapps that replace a lotof the content we create.Today, we have templateswhere we start with a PowerPoint templateor a template for our budget.All of those things thatwe either use or give awayto our prospects, those aregonna turn into prompts.I said, right now we obsessabout graphical user experiencewebsites and that's gonna turninto chat user experience websites.And today marketing's allabout how do you personalizethe experience and in the future marketingit's gonna be all about how you createsegments of one that because of allthis advancement in this technology,you're gonna be able tocreate a very specific thingfor each individual person thatyou are trying to market to,versus some kind of roughhigh level personalization.You agree, disagree?Did I leave anything out?What do you think?- Templates are prompts, yes.One to many, so one-to-one, yes.Like everyone gets theconcierge experience .Chat websites.Let me throw in more crazierversion of that which is.- Yes, please I'd love it.- Which is, I think that you want,you may want the combinationof both those things,but actually if you lookat some of the examplesfrom Auto GPT,what you actually couldget is a chat experiencethat could actually build you the web.Like could just say, okay,can I look at the homepageand it will just like build whateverhomepage you want for that websitebased upon what your needs.So you would like say, oh,like I have these questions about thatand I would like to see it in a webpageand it could just create the webpage onthis fly for your brand so everyonecan get a one-to-one webpage, right?- What you're saying is that the segmentof one and the website'sidea could emerge.- Are the same.- And you could have a milliondifferent versions of your website.- Of your website because if look at.- I completely believe thatand that's mind blowing.- [Kieran] It's mind blowing.- You could go from a websitethat a whole team of peoplework on for most companiesnow for like one homepageand say, ah, it's I have one homepage,I'm gonna have a million homepages.- Like your hubs, the hubs hot website,the Zapier website couldbe whatever that personneeds that website to be.So they're in the chaton the right hand side.They get the website built,the webpage built and theyclick around and say, okay,I don't want that webpage,I want a different webpage.There are tools that are building that.- There's a world where thewebpage also that homepageor the product page changes as they chat.- Right.- Yeah.- The chat interaction with the AIis a type of prompting and thecontent dynamically changeson the webpage based on that chat.- Exactly.- What you're arguing is thatwe're gonna have a chat userexperience and a graphical user experiencein parallel with each other.- In Parallel.- And so it's all withinthe same experience.That's probably better.- And it's personal to you.The one that I don't knowabout but you can explain to meis content to micro apps becauseI wanna read my newsletter.Why would I replace that with a app?- Well I think it's not a,it's not a hundred percent,but I think that there's alot of times that people are.- Just one quick thing?- It's opening up myfucking Nord password thing.- What?Auto GPT is is ruining Kieran's computer.- I just wanna make sure it's not loggingin to my fucking Nord password.Shit, I think it might be, you know.- Auto GPT's are not ready for prime time.- No.- Please experiment at your own riskfor learning purposes only.- Let me just tell it to stop here.Live on the pod.I need to tell my Auto GPTto stop opening stuff, okay.It stopped itself, okay.Because you can tell it.- Folks right there waswas what Kieran sounds likewhen he's panicked.There's a different tone of of his voicethat happens when he'spanicked right there.- Things open and magically in your life.Okay, weird, let's shootback into the content.- Okay, so now that Kieran'slaptop is maybe secured.Well, what I was saying Kieran is,is that there's a lot of timeswhere we are creating content,when actually a web appwould be better solution.So like here,let me tell you how todo this thing versushaving a web app to do that thing for you.And the reason we couldn'tdo that historicallyis 'cause it's too expensiveto pay somebody to build thiscustomable this thing.- Ah, yes.Yeah, yeah, yeah, yeah.- And now that the cost ofweb apps are becoming zero,we can do that.And so a lot of that likeinstructional based content willgo from this is how you do it,to we have done it for you.That is a big shift.I wanna repeat that again.It's a big shift in the world of AI.We are gonna go from how youdo this, which is what we,how we think about mostthings today in marketing too.I will do this for you.And that is a big, big shift.- It's somewhat interweavedinto the templates and promptsbecause all of the educational contentcan just be templatized prompts.- Correct.- But you know what I mean?Like you actually don't needany of the educational content.- The the two overlap a lot, okay.We have so much stuff to cover.Kieran, have you seen this,the new AI readinessreport from scaled out AI?- Yes.- That just came out.Okay, so they surveyed 1600execs and there's a bunchof interesting stuff andI've got some saucy takeson some of these things.And one of the things thatare interesting is the debatebetween open source and closed models.Like are we gonna have big opensource models or is a closedmodel like Chat GPT whereopen AI controls the model,the data, everything, what's gonna win?What do you think,which model do you think isgoing to be more prevalentin the future?Open source or closed source AI models?- I think for enterprisecompanies, closed source.I think for small businesses open source.- Fricking love that,you and I are on exactly the same page.Here's the chart thatfirst made me lose my mind.Only 37% of businessesin the next three yearsthink that it is highly criticalto invest in AI for their business.- What the hell are people doing man?- I thought that.- I don't know how it's allnot highly critical to everybody.- Yeah, but I still think that's,like if you actually think abouthow things have changed,like six months to see the majorityof businesses say thatthis is somewhat critical.- But it hasn't been six months,it's been six months sinceopen AI blew up, yes.But AI's been happening for like 12 years.- Yeah, but it's not, it wasn'treal to those businesses.I know we had, open AI had their APIs,but even Sam Altman said like,no one is really using these things.- But the fact that we are sub 50%as highly critical is surprising to me.- I think that changes really rapidly.- I agree.I am just putting out that if that,if you're a business andyou were watching this,I would put it in the high critical bucketis my advice.What's interesting is thata lot of the boring industriesare gonna be leaninginto AI, insurance,supply chain management,financial services, healthcare.And this goes to my big, big thing.If you look at the tops use cases,you know what they are Kieran?Their money's saving, not money making.If you are listening to this show,watching this show on YouTube,I think the biggest opportunityin AI is not money saving.Its money making.All of these claimsprocessing, fraud detection,they are trying to saveyou money, reduce risk,reduce expense off your balance sheet,because that's where theeconomic climate is right now.I think AI has much biggerpotential to transformhow your sales team operates,how your marketing team generates,leads, demand, customers,all of those things.And there's much moreopportunity in making moneythan saving money.Do you agree or disagree with me Kieran?- I think it's both, I think.- Well it's both, but I'marguing that making moneyis the bigger of the two opportunities.- I think AI is potentiallyincredibly deflationaryand the way that companies will bemuch more profitable asto make their companiesmuch, much leaner.- Are you agreeing withme or disagreeing with me?Are you thinking AI is betterfor saving money or for making.- I think in a lot of casesit's better for saving money,but then there's like use casesand jobs that you cannot dotoday or things that you cannot sell today'cause they're not possiblethat AI will make possible.There's new businesses thatwill be created through AI.And they're money makingand then there are existingbusinesses that will integrate AIand they'll be money saving.- Look, I think that eventhe existing businesses,the ones that spend more oftheir time on money makingversus money saving willhave the bigger upside.It doesn't mean that you'renot gonna save money,it doesn't mean that you'renot gonna get leaner.But those are,the problem with moneysave using AI to savemoney is that you have afixed set of expense, right?You know you spend a million dollarson this thing a year and you're like, oh,I can use AI to save half of it.So you have a cap on what you can achieve.You could only save half a million dollarsin that example, right?When you cut,when it comes to money making,your potential is uncapped.You can.- But I'm struggling to seehow you can differentiate those things.Let me give you like,I have a hundred personsales team and I'm like okay,well, I want AI to make my sales teammuch, much more efficient.I think what I'm gonna endup with it with is a smallersales team and a better sales funnel.- Well, you could argue there'stwo ways of look at this.You could argue, hey,I'm gonna meet my existingsales quota with less cost,because my sales reps aregonna become more productive.Or I believe I can make allof the sales reps I have threetimes more productiveand I'm making more revenue, right?That's was a perfect exampleand I think the latteris the bigger opportunity.- So I agree with that for first movers.This is the my point thatI made in a previous show,which is for first moversthere has never been a potentialbetter time to integratethis and land grab,the laggards,by the time the laggards integrate AI,I don't think that opportunity existsbecause it just becomes, again,come back to the very thing I startedwith the incentives reduced for everyone.And so everyone is like leftin a worst situation andthat's what marketing,that's what a lot of go to market tacticslike they reduce it every,someone gets an advantage at the startand then it gets leveledat over the longer term.I think that's what's gonna happen.So I agree.If I was in the first mover,which I think we are,I would just focus on likehow do I have 110x engineersversus having 10 engineerswho can do the job of like a hundred.I would keep my a hundredengineers and I'd be like,okay well I can actually justlike try to win this marketand land grab really rapidly here.- I love that.One last piece of datafrom the scale.A report,and this is a shout out,I'm got a thread pulled up herefrom @nonmajorPete on Twitter.But one of the things that wekeep talking about pushing onwith AI is that AI isonly as good as the datathat the model has.And companies, are reportingthat they still have major,major data issues and thetwo biggest data issuesare data quality.So they have data but the data's not good,it's not accurate, it's not cleaned up.And then data collection,they're just also having problemsgetting that data, right?- Billion dollar business idea.I'm telling you, oh billiondollar business idealike company comes in and basically sucksin all the data and form.It's like a CRM foryour data for AI models.- [Kipp] Oh yeah.- So it can come in,suck in the data and organize it in a waythat's gonna be useful for AI models.- I think GPT labs and a bunch of peopleare working on those,like a lot of the data datacompanies in the tech spaceare working on that.But like whoever leadsthat is going to be.- It's huge.- It's a huge opportunity.They're going to be very, very successful.Kieran, what's crazy about this,we've been talking already andwe haven't talked about maybethe biggest story of the last couple daysin that Google is reinventing itself.They have put a brand new teamof people to try to rebuildGoogle search in an AI world,what the hell's going on?What do you think about it?Gimme your take.- I have been right about thissince the very first time.- What's that sound everybody.Oh that's Kieran taking a victory lap.That's him running around.He's taking a victory lap.- Everyone was like, oh, like it's gonna,that's a toy and they're gonnaintegrate like these featuresinto the search engine and Googlewill win and Google may still win,but Google now realizethey're not in a fight,they're not in a search engine war.They are in a natural language platformfor the internet warbetween the and open AI.- I like that, that was good.- The aggregator of theaggregators war and search is likecommoditized in thatbecause it is like partof that experience.It is not that experience.And so what they did wasthey were like, wait,they were like Samsungcancel them and said yo,we might just put Bing acrossall our devices because youare like the old school born search engineand that's a $3 billion a year.- Yeah, so explain that,.Explain this to people.What did Samsung do?- Samsung, rumor is itthey are contemplatingripping out Google asa default search engineand add in Bing becauseBing, because Bing is cool.I told everyone Bing was gonna be cool.I did the graphic withthe undertaker coming upand rising from the deadand put Bing's name on itand that is a $3 billion ayear contract for Google.Now apparently that really shocked Google.Imagine it.They're sitting up intheir ivory search tower,no one has ever questioned their throne.They're like,they have the iron throne andthen suddenly someone comesand goes, yeah but youaren't that cool anymore.We're gonna use someone else, Bing.And so.- Well, hold on, hold on.Kieran, you're missingthe second part of it.Samsung picked Bing over Google,it's not just the searchdistribution that Google loses,Samsung is running Androidof most of those phonesand it basically puts the entire Androidbusiness at risk too.- I didn't think about it actually.- Right?It's a double whammy.And you know Satya atMicrosoft's being like, cool,well Samsung gimme that searchbusiness and then I'll putWindows on all of your phonesand we'll just kill not justGoogle search but Android too.- It's a signal.It's a signal that Googleis like, okay, wake up.And so they put 160 peopleon and you search experience.But I wanna be really clearin the language they usebecause this wasn't, weare going to integrate AI.This was, we are going toreimagine what search is.- [Kipp] Yes.- And that is the point thatI have been making repeatedlythat the blue lengths click, click, click,is a historical waythat we used to retrieveinformation and is notgonna be the way we retrieveinformation in the future.But coming back to like the.- Kieran, I still think our best marketingagainst the grain swag ideais just a hoodie that says legacyand has blue links below it.'Cause it's just likethe old way of doing it.- If you want that jumper hoodieadd a comment into the YouTube- If you want that asas the first MAT swag,let us know.- And so, they wanna reimagine search.And so the big thing about that is,we talked to Neil Butel,we had him on the show.You can go back andlisten to that episode.It was a great episode.We went through Chat GPT.Does it crush, search and advertiseand Neil is on the side that Googleis going to weather this storm and SEO,everything is all good.But the one thing we didtalk better than that,that is part of this PR releaseis there may be a differentway for Google to monetize itsexperience outside of search,which is through transactions.Because none of them can bethe aggregator of aggregators,which means like booking.comand Kayak and why do you needany of this when you have anatural language platform thatcan do all of this stuff for you.I can just tell it what to doand it goes and does it for me.- We talked about the solution with Neilwhere bids are gonna move from biddingon advertisements tobidding on revenue shareand the transactions.And I think that is a bigpart of the future of search.So Google, Chat GPT, Auto GPT is Samsung.There's a lot of these marketpressures that are forcingGoogle to reinvent search and Kieran,there's another angle to this.You know what I have beenhearing rumors about?Seeing rumors about?That Meta and Mark Zuckerberg,they're making some big AIprogress behind the scenesand that they are going to be makingsome big announcements soonaround their own AI developments.So it's not that just thatGoogle has to contend with openAI and Microsoft,Meta is being very aggressive in this gamefrom everything I'm hearing.And I think you and I aregonna be doing an emergency podsometime soon on all ofthe meta AI announcements.- Do you know the meme where the partyand you have the guy with the drinkin the corner and they always have like,they don't know like I'm a Bitcoin traderand everyone's ignoring it.I always feel like that's in Facebook.Nah, that's like the Metaverse.It's like they don'tknow I'm still workingon the Metaverse, it's like.- Win AI so that then Ican then I can go backto the Metaverse.But in all seriousness,Facebook has one of the mostunique data sets in the world.And so they already havethat advantage in AI.- [Kieran] Huge.- And so that's what makesanything they do on AIso critically important andso important for anybodyin the business and marketing world.Like that will be a must watch,Kieran and I will have to like wake upin the middle of thenight and do that pod,whatever we need to do.Because I think that is goingto be what are the biggerpieces of AI news overthe next couple monthsis going to be what Facebook does.And then I think nextis gonna be how Googlehandles Google io.What is Google going todo at Google io around AI,I think are gonna be twovery interesting plot linesof the next couple of months.- Well, they're gonna launch,apparently the word in the streetis they're gonna launch Magi.Project Magi.Which is.- [Kipp] Magi.- Magi.Is it Magi?- It's okay, everybody watch.You just know Kieranpronounces everything wrong.- Magi, Magi.- And if you try to lookat anything based onhow he says it,you're gonna find some weird, weird stuff.Please tell your colleaguesthat you were watchingKipp and Kieran and you've heardall about Project Magi, all right,so they're gonna have these Project Magi.- Still can't say it.- I can't say it.They're gonna give amillion people access.And so that is like whatapparently they're gonna give likea million people.I'm sure.- I'm sure you and I are not gonna beat that million people.- Well, I can categorically confirmthat I will not get access 'causeI'm based in Ireland andwe don't get anything.We're not allowed to have anything.So that's what they're gonna do.- I feel like I'm out ofbreath from everythingthat is happening in theworld right now and the,the just pace of whichwe just went througheverything on today's show.I think our big lessonshere is that Auto GPTsare a glimpse into the future of what AIis going to be able to doand how they're gonna beable to transform businesses.They're not ready for prime time.Trying to log into Kieran'spasswords and do all kindsof sketchy stuff.Even on the show.The data from scale was super illuminatingon just where businesses' heads are at.And one of the mostinteresting things to meis that data is still a hugeproblem for most businesses.And it's going to be one ofthe things that delays AIadoption for larger companies.And that's going to give the upstartsand the disruptors some advantages.And we are beyond interestedto see what happens with Googleand Meta over the next couple months.There's lots of rumors shaken aroundand we cannot wait tosee when stuff goes liveand share it all with you.Please hit subscribeon the YouTube channel.We love that you takethe time, watch our show,leave us comments ifthere's any future topicsor any questions you have.We'll see you real soon on"Marketing Against the Grain."(upbeat music)- This data is wrong every freaking time.- Have you heard of HubSpot?HubSpot is a CRM platformwhere everything is fully integrated.- [Speaker] Whoa, I can see the client'swhole history, calls,support tickets, emails,and here's a test from threedays ago, I totally missed.- [Narrator] HubSpot, grow better.(upbeat music)
