Sharp Tech with Ben Thompson - Appropriate Fear of Microsoft, A Workplace Efficiency Counterpoint, Artifact and Substack in the Modern Era

Episode Date: March 20, 2023

What Microsoft Copilot could mean for SaaS businesses and CIOs and Google, the midrange jumper and questions about long-term office culture optimization, and a few more thoughts on regulation in the wake of the SVB collapse. At the end: Artifact and Substack and the challenge for text-based businesses as the world moves to video.

Hello and welcome back to another episode of Sharp Tech. I'm Andrew Sharp and on the other line, Ben Thompson. Ben, how you doing? I'm doing well. I'm doing well. Last week was actually our 50th episode of Sharp Tech, a nice sort of synergy with Fernando Onzo getting his 100th podium ahead of Mercedes. So just, you know, everything sort of tied together.

It's very, very lovely. There we go. Well, we delayed today's recording session so Ben could watch the end of the Saudi Grand Prix. And all I can say is what I need. Winners keep on winning. What I need more than anything else in my life right now is some sort of scandal to hit Red Bull and Max for stopping. So if anyone out there has any compromising information we can use to bring those guys down, you can reach me at email at sharp tech.fm. And we'll get that

over to the FIA and get the wheels in motion. I like it. This is sort of the same strategy of Lewis just complain in wine and look for it. Rules-based edge. It's more constructive than what Lewis and Mercedes are doing because that's just like pure passive aggressiveness. I'm looking for real hard facts that I can use to take these guys down. But yes, it has been 50 episodes. We've only talked about Formula One on like 48 of those episodes.

We had a nice little winter break as it were. So we should probably get back to the tech. Let's get to it. And we are going to be. in Ben with an addendum to last Thursday's episode. So, as you recall, about two hours after we finished recording that show, I texted you with a warning and said, your Yonnasink board seat is now under formal review after you compared him to a chat plug-in from Microsoft Office.

And at the time, you hit me with the crying, laughing emoji, blah, blah, blah. But then 12 hours later, you came back dead serious and said, you're going to take this back. The demo from Microsoft was unbelievable. So I went back and watched the demo over the weekend. I think most importantly, Microsoft deserves a lot of credit for keeping this demo short. It was a tight 36 minutes. We very much appreciate that. But what amazed you so much about the demo such that you're now willing to lean into the

comparison between Janus, the best basketball player on Earth and a chat plug-in for Microsoft Office. Well, I think you should go first. I'm the one that is like hype this up. So you came in dubious. What was your response? So I am a little dubious still. It seems to me that co-pilot isn't really like a new product.

Like it's mostly just a feature that adds on to all of Microsoft. 's existing products. And so the idea that this is going to, like, disrupt how everyone works, I need more convincing on that front. It definitely seems like a sustaining technology that is going to allow Microsoft to maintain its edge and maintain the supremacy for Microsoft Office and everything else. But, like, I didn't watch that demo and come away thinking, like, if I didn't already use office, watching people use co-pilot would make me say,

oh my God, I need to get Microsoft Office as soon as possible. Like it's not quite that transformative, at least from my standpoint, but maybe I'm missing something. No, I think you actually nailed it, but you actually yelled it completely, to be honest. So just to back up, Microsoft demonstrated, the first thing was basically not chat GPT,

everyone calls it a chat GPT, A remarkable victory of branding. It's like Google is like there's no search, there's Google. There's no tissue. There's Kleenex. There's no large language models. There's just chat GPT. Everything is chat TPT, which is a nice, nice work by OpenAI there.

But you have this large language model that is based on GPT4. They're calling it co-pilot, which by the way, as far as branding goes, is a phenomenal brand name. I mean, just like it comes from GitHub where co-pilot was, it sort of helps you, you know, it helps you code. It's an auto-complete sort of functionality there. And I think a great branding extension by Microsoft to put it in all these different apps because it really captures the sort of idea. It's not writing stuff for you, but you give it the right prompts. You give it an outline and it fills it out in Word, for example. Or you're in Microsoft Excel and you're trying to like analyze some sort of data or do some scenario analysis or what, you know, if this variable was changed,

what would look like instead of like duplicating a sheet and then going in and changing some variables and doing new graphs, you can just ask it, what effects would have been different? And it will just generate a different graph and do this sort of things, you know, sort of from that perspective. Or you want to make up a PowerPoint. You have a document, say, make a PowerPoint out of this document. And it makes the PowerPoint. Now, obviously, we're looking at a demo where everything is perfect, right? You have all the right assets in place. There's already the right images for it to draw. You can imagine just throwing like our show outline and saying make a PowerPoint demo. You're going to get a, you know, it's going to be a bunch of text on slides.

But this idea where it's this augmentation of your capabilities in these apps is just, it's very straightforward. It's very compelling. And the fact that you're unimpressed, your sort of take is, yeah, duh, of course that's what it does. To me, that is the great victory of this demo. Unlike other demos where it's like trying to. to suggest here's a new use case or something new you might do that is impressive in its own right. And in the grand scheme of things is arguably a greater opportunity. But what I found so compelling about this is the seamless way where it was just like, oh, duh, of course that's

what it does. And to me, that's actually what is compelling. Because you can imagine this, this is going to be the AI product that gets massive adoption right off the bat because it's just going to be super obvious, I think, in sort of how it fits in with what you do. And it's also, that's number one, number two, to your point, everyone uses office, it's not everyone, but people use word processors, they use spreadsheets, they use presentation software. And it's going to resonate, I think, in a very clear way where, yeah, that would be awesome. That would be really great to have. It sort of exhibits the capability of this in a compelling way. And so I thought that the, the boringness from a certain perspective to me was actually what was very compelling.

Yeah, well, and it all starts with the name. That's actually a good aspect of it to highlight. Co-pilot, I was watching the video and I was thinking to myself, you know, that conveys exactly what this product is and it conveys it in a friendly way because there's all this angst that, you know, AI is just going to replace white collar workers across the economy, and that may still happen, but co-pilot at least sort of like articulates how this is going to fit into people's workday. Now, in terms of actual work, I don't know, because like I'm of two minds. Like if you're in consulting, I was talking to my wife about this

and you're competing for work, the ability to go back through every successful proposal that the company's had in the past and pull out a similar framework and update details where applicable. Like, that's all really great. But then some of the stuff in terms of like productivity after meetings, like these tools make it so easy for one team member to generate content after a meeting. But that also creates more content that every other team member has to review. If you're dealing with like a co-pilot fanatic on your team who's sending everyone the action items. And so I just wonder whether the adoption will be. as, you know, whole hog as you're envisioning here.

Because I think some people, like, particularly like ultra type A type people could watch this video and be like, oh my God, this is going to make me 10,000 times more productive. And then a lot of other people who work in offices are going to say, this is going to create 10,000 times more annoying BS that I have to sift through on a daily and weekly basis. Well, wouldn't it be nice if you could have a. an agent that could sift through all that stuff trivially, right? I think that's sort of the other piece here, which was you had the obvious sort of, quote-unquote, plug-ins to the existing applications, but then you also had this business agent.

I can't remember exactly what they called it. Obviously, a less compelling brand name, whatever it was, but where it basically will operate across your entire corpus of data. And so you can ask for notes from the meeting previously, or all, our interactions with this client or whatever might be. And it pulls it in. And this one, I think in some respects is a less compelling demo because it's not that, that is the new capability that was really shown here.

Like you have this, you know, you have a share point or wherever it be and all this sort of stuff and your CRM systems and all these different pieces. And you can access that basically with a question. But that also really puts, you know, you can imagine a CIO. who's making a top-down decision for his company's tech stack saying, like, yeah, that's what we need. We need to be able to sift through all this data. And we need to make it a situation where the generation of more data is an asset and not a liability, which means we need a way to be able to search through it.

You know, just a bit of a regression. You go back to Google in the sort of early days when Google came along and just crushed it. You know, the search engine was so much better. The old original version of a search engine was like an index. Like you had a list of links that went into it. And you go back to the old Yahoo your homepages. And they would actually have in parentheses the number of pages that were under that category. And it just kept getting bigger and bigger.

And it eventually sort of fell over under its own weight. Whereas Google would just give you the answer, right? And Google, by virtue of leveraging the link, by figuring out how to rank stuff by who was linking to what, the link was actually the core driver of Google's sort of accuracy and liability. And you had this situation where the larger the web became, the more links you got and the more accurate Google became. So they have this real bifurcation with traditional search engines where search engines were breaking under the weight of the exponentially exploding web. And Google was getting better because it was actually getting more information that I could

act on to give you better results. I think this is a similar dynamic here where Google, to some respect if you want to complain that Google searches have gotten worse. That's not because Google's stupid. It's because there's just a bunch of crap online, right? There's a lot of bad web pages. Yeah, I think Gruber made that point a couple weeks ago, and it really rings true. The problem is like the web generally, not really Google. And it's been like 25 years of folks trying to game the Google algorithm and figuring out

how to get their stuff up top and understanding sort of how it works. The brilliant bit about these large language models is they're like, yeah, the more data the better. Like, give us more. And we'll sort of much better at finding that needle in the haystack. And actually, the more hay there is, the better, because it's not finding the needle. It's constructing the needle, right? It's like, I'm looking for this needle and I'm going to run this like molecular process of bringing in all this hay and transforming it into the needle you're looking for that may or may not have existed before, but it's pulling in all these different pieces and it's sort of a piece of acron information. So you can, Imagine a future where now there's interaction with your corporate data. And from a certain perspective, it's, yeah, generate more meeting notes. Like capture every transcript. Put all this stuff in there. And our expectation is not that other people in the company and all have that much more crap to read.

It's we have that much more material to create the needle as you need it. And so you think about that. The impact of this is not just that this is an interesting tool in its own right. You're sitting there as a CIO saying, look, every single tool we use needs to plug into this system. So, hey, that CRM system doesn't work with our business agent, switch to the Microsoft one or switch to the one that has a deal with Microsoft to make sure they plug in and incorporate so their data can be served.

Like, I think this is a real serious threat to lots of SaaS companies where they got a way of Microsoft sort of keeping their eye off the ball. The big ones are like the horizontal ones. They're trying to be like broad-based, like Slack or whatever, or notion or these sort of things like, We're just going to capture all this data across the company. Now it's like just use the Microsoft alternative because the payoff is not the content creation. It's the content discovery.

And they have this tool to do discovery. And if you're a vertical SaaS product that's in like a specific sort of application of work, it's like you better figure out how to plug in to the Microsoft ecosystem so that your content can be discovered because, you know, there's two ways to grow a product. You can be bottoms up and small teams adopt you or it can be top down. The CIO says you have to use this product. this demo is going to be very compelling to a top-down CIO that is looking at the efficiency level. It's like, how can I actually get more production out of the current employee base? And it's like, look, if we switch stuff to Microsoft or Microsoft endorsed products,

I believe we're going to get more, I believe as the CIO, we're going to get more productivity. And by the way, if the CIO makes that decision, it doesn't matter if it's true or not to some extent, right? Like the power of a demo can be just compelling from that sales perspective. So like there's an aspect. Silicon Valley was terrified on Microsoft in the 90s, just absolutely petrified. And whenever they had moved to the space, everyone will sort of freak out. And then Microsoft goes into this 20-year lull where, you know, not only were they behind the ball, but then they completely dropped the ball on a lot of SaaS applications, in part because

they were trying to like protect Windows and their entire business model rested on like Windows server and all that sort of thing. and they finally move into the SaaS sort of space with office. And now there's a bit where they are firing on all cylinders. They did not pull back because of Bing. They kept pushing forward. And I was talking to someone in the Valley on Thursday about this demo. They're like, what demo?

And I think on one hand you say, well, Microsoft should do a better job making it aware of what's going on. And we can talk about that. They like it wasn't on YouTube. Like just a like just a bunch of dumb Microsoft stuff. It was on LinkedIn. Like talk about a strategy task. But let's talk about the dumb people in Silicon Valley who are not paying attention to what Microsoft is building over there. I think there needs to be a much healthier degree of fear, to be totally honest.

So let me ask you, though, in terms of their advantages, Microsoft's advantages, that is, in terms of what they are bringing to the table, am I correct in saying that the reason they're going to have long-term advantages is because the corpus of data they're drawing from is just much broader from. from Microsoft Office users than someone who's like if Slack were to develop its own AI plugin. Like Microsoft is going to be better because there's more applications to pull from whether it's Word or Outlook or anything else. Is that right?

Well, there's a few things. The number one advantage that Microsoft has, has always had and will continue to have is existing business relationships with the vast majority of large companies and small and medium companies, frankly. Everyone already uses Microsoft. That's what I'm wondering.

Like is Microsoft so entrenched at this point that the technology isn't necessarily going to like create some new advantage here? Because ultimately everybody's already on Microsoft Office. Like it's sort of the workplace version of the iPhone where they have so much market share that like new features are definitely helpful. But it's not going to really like change the dynamics we already have right now. Yeah. No, this is an affirmation of my thesis in sort of the, you know, a few months ago about about what's the. business impact of AI. And my takeaway was it's mostly going to help the incumbents and especially Microsoft. Like that, I mean, that clearly seems to be the case. And so, yeah, just like right now,

if every single company rolled out the exact same AI capability tomorrow, that's good for Microsoft because they, they already have the existing customer base. They already have, to your point, most a lot of the data, they call it the Microsoft graph, but all this sort of documents you've say all the email that you have is our a huge portion of it is already in the Microsoft system. It's obviously going to work well together. And if something lifts all boats, the biggest boat is going to get sort of the, the most benefit. That's definitely the case. You add on the fact that, number one, Microsoft is shipping or not, we'll hear the shipping

bit of good, at least they're demoing it. It's not available yet. But no one else is. So they have a timing advantage. And then number two, because it's the only one we've seen, it's the one that seems to work the best. I mean, Google talked about it and did a little bit of a demo, but again, not available anywhere. And what, you know, what's been my ongoing critique of Google for the last several years? I think Google is utterly and completely dropped the ball in office productivity.

Yes, they have the Google apps. And they took a big chunk out of Microsoft, a good 20, 25%. And, and, but what they never did was you had this explosion of SaaS apps from all, all these vertical SaaS apps from Silicon Valley startups. And someone needed to tie them together, to have them work together, to have data easily shared across them. And one of the reasons why Microsoft has stormed back and taken back share is there's just an aspect you can get all the Microsoft stuff. And on every individual product basis, it might be worse, right?

Teams for chat is worse than Slack, right? One drive is worse than Dropbox. So you can go through category after category. But if you level up and you level up and you. You say, look, the job of this business is not to chat more effectively. It's to actually get stuff done. And to actually get stuff done, it helps if stuff works together. I don't have to operate as an IT integrator trying to make stuff work together all the time.

That's always been a big Microsoft selling proposition, which is, look, we're not going to sit here and claim everything we do is the best in class. But if you buy the whole thing, it's going to all work together. You're not to have to deal with hassles and XYZ. And oh, by the way, you've been buying. from us for years. Like you're we already have this existing relationship like teams teams. Teams. Teams is already part of Office 365. We're throwing it in for free. Right. Like like like why not just use it? And oh, by the way, there's a slowdown or a potential recession on the horizon. Do you really want to be spending extra on all this extra SaaS software when we offer you a good enough alternative that works

together? And now you layer on top a potential meaningful differentiator where it actually ties it all together and works better. And that's a very tough thing to sort of fight against. And Google, they're going to pay the price for insufficient, what Google should have done for the last several years is they should have been building the alternative office suite. And it's not just Google stuff. It's like, why isn't Google deeply integrated with Slack, for example? Why isn't Google deeply integrated with Notion?

Instead, they just try to do their half-ass sort of like similar features in Google Docs or whatever. And instead of building up this alternative ecosystem and say, look, Okay, whenever you have sort of modular, it's going to be a little rough around the edges. But when you can get best and breed products up and down the stack, it's worth it because there's so much better. And we're going to tie it all together for you. And oh, by the way, here comes our AI agent that actually works across all these properties and can discover all the content.

And do you like, come on, let's be honest. The Microsoft stuff, it's all mediocre, right? But Google is not just late to deliver this stuff. They did not lay the groundwork for the last five to 10 years. to be ready to take advantage of this opportunity. That's what I'm wondering. So it's more of a 10-year mistake because I look at how successful Microsoft has been with Office. And I think, you know, Google has better AI, like Google Assistant is vastly superior to Siri,

but Siri has so much adoption across everywhere that everybody thinks of like the AI assistance as being bad because Siri is an inferior product. Very effective branding by Apple to just ruin the whole space. I mean, but now it's like it's capable because these largely, like there's this whole area, those are because of natural language processing. And it's funny, like, because it's not just that Siri is bad, but these LMs are embarrassing every single chat agent out there. So it's so much better. We've all come to know AI as this like bumbling idiot assistant and now suddenly it's flipped entirely. But that's the analogy I was thinking of watching the Microsoft videos.

Like, all right, so even if Google had nailed this, Microsoft still has. has all these other strategic advantages. So I'm not sure how much market there would be to like claw back in a scenario where Google's 20 or 30% better because they invested heavily in this over the last couple of years. But you're saying the real mistake was to not build up like a little competing universe of SaaS companies that they could offer as a bundle. Yeah, I mean, I've been complaining about this for years on strategy. and sort of broadly speak, you know, on various podcasts.

Just, and it's sort of in line with the overall critique of Google, which is they have this unbelievable business model with search. And it sort of makes you just sort of fat and lazy everywhere else. And yeah, they did Google apps and, you know, really, like really took a nice chunk out of Microsoft. And, you know, much to Microsoft's consternation and, you know, frenzy, like nothing gets Microsoft more worked up than I think. the whole Google Docs area, much more so than search, because that was like biting into their

actual sort of sort of business. And they just sort of stopped. Like there's just been no real meaningful development in the, and not just Google Docs, but the, the broader ecosystem. And that's sort of what happens when it's not a core driver of your business. Like, like it just, you know, you're not going to do the hustling and do the, you know, the grinding to get all this stuff working together and to lay the sort of groundwork when it's an infinitesimal part of your business. and there's no real like real gain to sort of dominating it because you have this thing on the side that's just making you a whole whole bunch of money and it's almost it's very easy to like paint this as a it's obviously coming across as a criticism of Google but I think it's more interesting to think about it as just a fundamental failing of large companies like like and you can go back and everyone will point to a million things that Microsoft missed or was slow on XYZ back in the day and that's because they were in the same situation like that's what happens when you get big and fat and you have one you know you that just drives basically all the outcomes of your company. It's tough to sort of walk in elsewhere, particularly when that might cost you margin or

it might cost you, you know, you have to move quickly. You have to adapt to the companies that you're working with. And I think they're just going to reap, reap that pain a bit going forward. Yeah. Well, a couple final thoughts here before we move on. As far as the demo was concerned, another thought I had watching. the way this is theoretically going to be implemented on a daily basis. Like a lot of the work that is going to be wiped out is research intensive, basically

like shut your brain off grunt work that is perfect for automation, but is currently performed by younger employees. And so one way this could disrupt work would be rendering all those younger employees jobless. Yeah, I mean, this is basically like the. world's greatest intern is kind of what the demonstration was. Totally. And so I'm very curious to see how businesses handle this because obviously you and I have talked about several different times.

Like the premium here, the differentiation is going to be idea generation. Like what prompt do you feed the AI or whatever? Or how do you win business, this, that, or the other? Like the question, though, is you're still going to need to nurture and like retain younger talent who can one day fill those roles. And so I just wonder how companies plan to strike the balance between investing in employees who aren't necessarily going to be all that efficient in a vacuum and are probably going to be less efficient than AI. And that's one of the questions I'm curious. That's one of the things I'll be watching as this actually like hits the economy.

Yeah, it is very interesting because I think one of the the ongoing critiques. of Microsoft that I've leveled is, okay, we get it. You have great distribution with existing businesses. Like, how do you actually get new companies sort of using your stuff, right? And one of the reasons why Silicon Valley is oblivious to Microsoft is known in Silicon Valley uses Microsoft products, right? Like, it's sort of a, it's a natural sort of blind spot. I do think on one hand, this actually does give Microsoft a selling point to actually grow share, to actually acquire sort of new companies and do this way. But the overall, I think, critique and question remains to an extent and not just about

the battle versus Google per se, but just the younger generation in general, you know, are they going to even use office applications, right? Are they even going to be making presentations in PowerPoint or should they all just be making short form videos to do the, I don't know, like, but there's, I think that's a legitimate question, too. It's going to accelerate this bifurcation between the older generation, the younger generation. I think it's a good one. And maybe that's a question within companies that use this.

But also, if young people really adapt to this capability well, and not just the Microsoft capability, but just the figure out how to use these large language models in work generally, is that actually a push for more disruptive companies that actually do approach problems in completely different way? and are just totally orthogonal to the existing things. Like maybe there is a real line between any company that was founded, say, before 2015 or even 2020 and once after. I think that's an interesting thought. The other question I had for you, did you see the section of the video where a VP from the office product group walked us through a scenario in which co-pilot helps plan her daughter's high school graduation party?

Yes, I did see. that. So for anyone who hasn't watched the full YouTube video, we'll put it in the show notes. I just need to say that co-pilot drafted an email to this lady's friends, a PowerPoint presentation that was going to be played at the graduation party that drew from family photos. It created a to do list for different party planning needs and drafted a speech for this woman to give at the party. itself. And Ben, I cannot emphasize enough how badly every big tech company needs a VP of normalcy to just double check these videos before they go live, takes it from fun and exciting to creepy and dystopian real quick. And you should just have like a big red button that the VP of

normalcy can hit to delete all of that. It was pretty bad. I was pretty bad. Number one, it was just this complete, like, who's throwing 100 person parties, like birthday parties for their kids? I mean, it was, it was an unrealistic scenario that was completely out of touch, number one. Number two, the whole writing the speech bit was rough. I can't believe it, man. I'm going to demonstrate what, how much I don't care about my child. It's like, that's one of the only things. Look, if you're auto generating a thank you note, auto generating a note to your team, fine, do your worst with AI.

If you're writing a speech at your daughter's high school graduation party, I want you to use 100% of your brain for that one. Just take 20 minutes and hammer something out. Is that too much to ask? It was it was bad. And then the great thing was it wrote this speech. And it includes like notes about like when you should like make reference to other things or XYZ. And then it's like, yeah, this is too long. Can you make it shorter?

Yeah. So this is the future that we're all going to and have it together. It is kind of crazy, though, how quickly we've gotten here. I feel like we say this on every podcast we have about AI. But like we were sort of spitballing about a lot of these features back in the fall. And now it's about to be rolled out by Microsoft for real. Yeah, I mean, the one critique is it's not shipping. And this is, you know, now it's like we're back in Google Wan or let's give a demo.

and a couple companies are testing it. We'll see TBD when it's determined. We also don't know what the business model is going to be. I mean, this stuff is expensive. I think one thing that's interesting is you saw a couple times where they did a V1 and they asked it to like sharpen it up. That's what's probably happening there is it's using a more efficient, less accurate model that's probably cheaper on the first run.

And then when you ask it to like do it again, then it's like using the real stuff to some extent. And so, but that speaks to the fact this stuff costs money. It costs money to actually execute. So I think this will probably not be a team style, oh, here's this new tool that ties it together that you're already part of the bundle you're already paying for. It's probably going to be an add-on plan. Microsoft's not in the business of, Microsoft's in the business of making money. And so I think this will probably be something that is charged for, which does leave a potential

opening for companies that can figure out an alternative model to sort of come in underneath it. So I would bet from a Microsoft perspective, the first payoff here is going to be able to increase their average revenue per user. So they have different Microsoft 365 plans. Like there's the small business plans. There's the enterprise plans. The most expensive enterprise plan is called E5. And it includes all the security capabilities and all the, not just all the apps, they all include the apps, but like just lots and lots of different stuff that you might need. I wonder if there's going to be like an E6 now, which is like, oh, it includes all this sort of AI bit.

So their first payoff will be getting more customers. Their second payoff, I think, will be driving more usage of their apps because you want to have your data in this universe. Maybe number three is having a deeper influence on other SaaS companies to, you know, there's already this push to plug into teams, right? And a lot of people will gripe about this, that Microsoft gives you a very little nice little. a box that's for you to plug in. It's not like being a Windows back in the day where you can make

any sort of Windows application you want, right? All the capabilities are available to you. No, you have these defined capabilities in teams that you can use. This is probably going to be even more stringent in that regard. And then benefit number four would be actually acquiring new users. So I think to your point, there is a relation between my being impressed by the demo and the sort of skepticism about how much big picture difference does this make? Because what makes it impressive is the way it plugs in with what exists. But to your point, we were already in the state that exists. Where's the completely new stuff we've never seen before? And that's sort of a fair sort of pushback. There you go. Well, I will continue waiting for the truly disruptive technology to emerge from this dystopian six months of AI hype. But the sustaining technology on Microsoft side is all. This is. This is all. Dropping the Christians in terms. It's good. I like it.

But I think that remains the AI question is to what extent is it sustaining versus disruptive? And you go back to computing generally. There was this sort of initial wave of computing by and large made existing companies more efficient. Like they went from a room full of accountants to a computer program that like did the bookkeeping. Right. And there's those old pictures, like everyone like doing like basically spreadsheets by hand for all sets of purposes. That's all gone, right? Like computers did that.

And so you had a lot of companies like big banks, like airlines, like all like in large companies and ERP systems to manage your supply chain. All computerized sort of very, very early and all to the benefit of existing companies. It took much longer for there to be the rise of completely new kinds of companies that started out with sort of like different sort of, different sort of assumptions. And you can saw the internet too, right? You start out all the initial pages like, we're in newspapers,

put our stuff online. And it's like, wow, we have way we're readers. This is amazing. In that case, it accelerated very rapidly to crap. Everyone has all their content online.

We have no competitive advantage. And we're all screwed, right? The question for AI is when and where does that second step come in? No, we're actually creating stuff that was not at all previously possible. Sure, AI made all these things. better and more efficient.

But now there's something completely new that's orthogonal to what's available or it's fundamentally changed the competitive parameters of sort of the space that we're in. The possibilities of what can be generated. Right. And I think that that remains. One of the biggest questions about AI is when will this new capability be possible? Where will the capabilities come from? Like everyone's building on the Open AI API.

Like is open eye going to like come along? Like there was an extent where GPT4 comes out and it does the job that a lot of startups building in GP3 did better because it's just like better at different stuff. It's like, well, that was fun. There's a bit where, oh, a lot of Silicon Valley companies were building capabilities that Microsoft just demoed. And oh, by the way, we just demoed it within the apps that all your prospective customers are already using. Right.

It's wild. And again, it extends the humans because I think about this in a law con. context. Like, you learn how to be a lawyer by doing legal research and doing a lot of grunt work that, you know, computers will do more efficiently. If not now, then certainly like three or four years down the line. But then how do you generate new partners? If you're not allowing people time to, like, figure out how to actually practice their trade, there are certain industries that will be mostly automated before too long and are probably headed that way already. But more broadly, it poses this question of like how much inefficiency is actually healthy for the long term if you're nurturing talent.

It's a great observation. I really, I really like it. Because I mean, you see this, like, there's real benefits from inefficiency that are impossible to measure, right? Like, there's a, there's a broader observation here. You and I talk a lot about sports and like how much we, the analytics boom in the NBA drives us up the wall, not because there aren't insights to be gleaned. Of course there are. But the reality of basketball, the beauty of basketball is so much of it is about the interaction between the players on the court. And by the way, there is defense and defense matters. And there's an aspect of things that are not measured. And the temptation of when there is a measurement is you will vocalize, oh, we know there's stuff that are measured. We're aware. You still need to watch the game. But the way it actually manifests is you have people that they're scared to state their opinion.

about something that is not measurable and they retreat to just caring about nothing but what can be measured. And I think the analogy here that you're sort of hinting at is look, and this gets to my bit about this demo being particularly attracted to CIOs, which is they're going to look,

say look at all this efficiency we can gain. Look at all these jobs we can eliminate. And to your point, what things that can't be measured are going to give, erased with this. And this idea of you're going to lose this pipeline that's not just, it's not busy work. Like there's actual knowledge being gained along the way. If you're the lawyer that can leverage large language models, do all your research, your ability to interpret

and contextualize the return comes from your many years of sort of fitting their style. Now I'm just repeating what you said because it's such a great point. I'm, I guess, trying to steal it for myself. But it's a very good observation. Yeah, so we'll continue monitoring all of it. I will say to put a finer point on what you're talking about with the numbers, it's smarter to take threes if you're a 33% shooter versus mid-range jumpers, if you're a 45% shooter on mid-range jumpers throughout the regular season. But if you need one bucket in the playoffs, taking the 45% 2 is much better than taking the 33% 3.

but that's only available to you if you're practicing it all season long when it is clearly the more inefficient way to play basketball, but it's the better way to win championships when you need that bucket at the end of a playoff game. So that's a little digression there. I just didn't want to be too vague in terms of what we're talking about with numbers. I think the bit about you have to practice. You have to get better at it. I mean, the bucks ran up the largest point of versus in the league two years in a row

and flame out in the playoffs both times. And then they take a foot back. We're going to start like practicing playoff scenarios during the regular season, even though it's going to hurt our numbers. And oh, by the way, then all the numbers got, oh, guess Janus is an MVP anymore. Guess what? He's an NBA champion.

Like that, like, there's a bit about this. And it's going to be tough because like you think about it, like, go back. I think the law scenario is actually really interesting. Because you have, you do have a weirdness of incentives in law, right? where more partners mean more profit share, right? Like, West for me, right? There's, and it works out because you're forced into the training of the next generation because you literally need them to do the job.

And is there going to be any sort of incentive to think about the long term when it's not just that you can save money in the short term, but like, do I really want to be training my replacement, right? I don't know. I'm looking at the outside. I haven't worked in a law firm, but I can imagine there being these weird incentive issues that, again, I just come back to the internet when you introduce zero distribution costs.

There are all these weird effects that no one realized. Newspapers thought they were different because their content. Turns out they're different because they own printing presses, right? Like that's sort of a real splash of water to the face that the fundamental differentiator of our business was not what we thought it was in the way we managed for a very long time. It was something completely different. And oh, by the way, it's gone. We're screwed, right? How many industries are going to have a similar wake-up call when this stuff sort of pro,

proliferates. There you go. Yeah. Well, we'll all find out together here. That's part of the fun. I mean, it's like, like, is Lewis Hamilton actually a great driver if he doesn't have the fastest car? Oh, my God. I really wish Ben had started watching Formula One before last season, but it is what it is. He's working from the training corpus that he has to keep it moving.

SV says, I love last week's episode on the SVB Bank Collapse. and one point that Ben made was that there will be increased regulations on local banks after the FDIC guaranteed SVB's deposits. Are we sure about that? Whether it's political gridlock, lobbying from banks, or some unrelated crisis that diverts the government's attention, I could see a situation where there are no meaningful changes to how banks are regulated, and that could lead to an even bigger bank failure. So, Ben, I'll let you respond to that question. and also just curious whether a week out you have any general thoughts on the way this conversation around SVB has progressed since the whole world lost its mind.

Yeah, I mean, the crisis is certainly not over. I mean, we'll see if First Republic even opens its doors tomorrow. It's interesting, they actually had a much more like a real asset issue. Like they wrote all these mortgages that are deeply underwater as opposed to being sort of government bonds or whatever. So the crisis is not over. I think there's a certain, I mean, I'm as cynical as anyone else. That's not true. There is a deeper level of cynicism reflecting this question that I think is not quite right.

The U.S. government does, for better or worse, gets stuff done. It gets it done in an ugly way and it's often attached to these huge omnibus bills. But like, changes do happen by and large. They happen in 2008. in response to the crisis where the tradeoff for the big four banks, where it was basically made clear they're too big to fail. So they're by definition safe. And there was, not just that, but equity bond, you know, equity shareholders got a huge haircut.

They ended up making a much of money out of the deal because when the prices sort of went back up. And that was a real bell out. That was a real ballot. And there were meaningful regulatory changes in responses where they have to go through these stress tests and they have to have this certain amount of, capital on hand that was significantly higher than someone like Silicon Valley Bank needed to. And so that, and that change happened.

And it was an appropriate change given the implicit or explicit sort of guarantee that sort of came out, came out of that. And number one, number two, I would say that the one area where the U.S. government does tend to be on the ball is managing things like this. when it comes to money. Like that is the like the U.S. understands, I think at a very deep and fundamental level,

the extent of control it exerts over the world via controlling the dollar, controlling the flow of money. Like that's the whole sanctions regime. All this sort of stuff is all predicated on the U.S. actually being in control of this stuff. Now, there is a bit where a lot of that control is sort of technocratic. It's in the Fed.

It's in the Treasury Department. one of the hangups now is Congress came back after 2008 and said, okay, thanks for saving the world economy. You're never doing that again without authorization. And actually one of the hangups right now is the Fed actually cannot put in an equity backstop to these banks. That would actually be what would solve the crisis right now. They can do this weird guarantee of deposits and they can say we're going to open up. this borrowing window that you're sort of be okay. They can't actually infuse equity into the banks to say this bank is by definition

solves we're putting in. That was explicitly forbidden by Congress after 2008. And so that is one of the challenges here where Congress sort of like appropriately in a democratic system took back that authority because it is a pretty big decision. But now they have to actually act if they're going to sort of shore this up. And it's not clear, politically speaking, that there is the stomach to do that, to sort of act to support the bank. So there is a bit here where maybe the cynicism is justified precisely because the best way to stop the current crisis does require congressional action. And that's going to make it tougher.

But I think in the long run, the long course, you just fundamentally cannot have a system of. basically infinite insurance on one side and no real restrictions on risk taking on the other. And maybe it will take a couple years or a decade. But this will shift over time. I actually think you can definitely see a scenario in the long run where the logic for, I think I might have mentioned this, but the logic for like a national bank is pretty obvious at this point. Like, look, this is where it's going to be completely safe.

And you're not going to get much return, but that's because it's completely safe. and maybe that nationalization because the U.S. does this stuff sort of weirdly, right? Like the U.S. has like universal basic income, but it's like Social Security disability. It's like,

that's... Well, and don't we already sort of have national banks if we're talking about like the big four banks? Exactly. That's the analogy that I'm making, right? We kind of do have national banks, which is they're just called J.P. Morgan, City Bank, Bank of America and Wells Fargo. And so that may be how it ends up.

But there is a bigger, broader takeaway of this. And this is a piece I'm working on my head. I'm going to write it at some point. I got to it in the last podcast actually. I think there's a lot worth teasing out here where it's like a physics reaction. This bit of like

an action triggers a reaction. And the reaction to the speed of social media and the sort of like, you know, I call it the AMC thing. I meant to say GameStop. Like this bank run is like, was like akin to the game stop bit a few years ago. where it just rises up super quickly.

But AMC was a meme stock also. So the reference lands. It didn't land as well as it could have. So I'm disappointed in myself in that regard. But you're going to have a response of rigidity elsewhere in the system just in the pursuit of stability. And this is one of the outcomes where you have destabilization and changes in one place. It doesn't end there. Newspapers, yeah, we get readers everywhere.

It doesn't end there. Right? Like it actually is going to play through. And if you don't have a structure that is stable, like governments don't have that much control. The system is going to work through this instability to a stable spot. And ideally with enlightened leadership, you guide it to a potential equilibrium that is all and all better or better than the alternatives to the extent that SB is right and we don't have enlightened leadership to guide us to that state.

that doesn't mean we're not going to be moving to a different state, right? Like we're not going to stay. This current equilibrium of effectively infinite insurance is not a stable equilibrium. And so it is going to shift. How that will shift will be interesting to observe. Okay. Well, we will see. And, you know, fingers crossed that the crisis doesn't get any worse over the next couple of weeks.

It's a little bit unsettling to talk to people in finance and have them come back to you and say, not actually sure how this is going to work out. So again, let's all just knock on wood and hope for the best here. Yeah, I mean, yeah, I mean, I haven't a few sort of finance chats. It is disconcerting. It's like, I don't know this stuff. I kind of wish you did. Exactly.

So I don't know. Back to tech, though, pure tech question. Jeremy says, I enjoyed Ben's interview with the Instagram guys talking about their new news reading app artifact. I enjoyed that interview as well. as a fan and purveyor of the written word myself, I'm happy to see somebody try something new in text, but if I'm not incorrectly reading in between the lines

as to Ben's instincts, I think I share what I presume is Ben's skepticism of a successful outcome for this app. What I fear is this app lacks any avenue to spread. Given that it's only app-based, it's trying to get you to adopt a new reading habit. For me, I'm trying to do less reading, my mobile devices. My email has slowly become my RSS feed, thanks to the newsletterization of

writing, and thanks in large part to Ben. And between those newsletters, my podcasts, and occasional Twitter browsing, I just don't see a way to fit in artifact, and I'm the ideal user for it. Can artifact succeed? And Ben, if you want to give a little primer as far as what artifact actually is for anyone who's unfamiliar with it, you're welcome to, and we'll link the Stratecta interview in the show notes. Yeah, artifact is from the founders of Instagram, Kevin Sistram and Mike Krieger, and it's a algorithmic

list of stories you might be interested in. And I actually think it works pretty well. I mean, like, now, I've been motivated to use it because I did this interview, because it's sort of an interesting new thing. But over time, the more you read it, the more sort of fine stuff that you will be interested in.

And this is just broadly speaking, a capability that recommendations are just better and better, right? And so it's not a new idea per se. It's just that the capability is so much higher today that, you know, it can do a better job. And I think that's the case. And we got into the interesting things about the cold start problem, right? When it doesn't know you, how does it find the right stuff to service to you? It takes on the order of like a thousand quick-throughs for it to really have you, like, dialed in. And so that's a real tough problem. But I think Jeremy is right

about my skepticism about this being a big outcome for a few reasons. He's exactly right. How do you grow, right? Like, how do you actually, what's the customer acquisition mechanism where you're sharing stuff and that triggers people to want to use it themselves? That is an open question. I think the fundamental market for text is not large. We've talked about this in context of Twitter.

Like videos and photos, particularly videos, they're just more compelling. That's something that people are generally sort of more interested in. There is a bit where I think artifact has a, has. a garbage in, garbage out problem, which is there's just a lot of crappy websites out there, particularly the free ones. There's ads everywhere. They're just like, like I, you forget how bad it is out there, particularly if you have, you know, when you're clicking through this, and they have a reader mode. They're, they're slowly making it more and more high level so you can see it. But there's attention because they're basically taking this content for free.

I mean, and, you know, are, do you ruin the business model of the, of the, of the, of the, of the, that you're taking it from, are they going to like block you or whatever it might be? So they have a real tension in solving that, but that doesn't change the fact. A lot of articles suck. It's still susceptible to quickbait sort of things. Like you're just seeing headlines. So you might want to sort of go through a thing and click it. So there's there's there's there's a lot of sort of difficulties here.

And oh, by the way, you could just see a leapfrog where why do I actually want aggregated crap articles instead of just an AI agent that. just tells me the things that I'm interested in. Like, give me the summary of what happened in F1. No, no, no, no. We're not done reading. The next generation can't just be done with reading all together or at least incapable of reading anything that's not in like four or five bullet points.

With all due respect to Sam Altman, I want better for the future of humanity. Yeah, but that's the natural state of humanity, right? Reading is the aberration. Like that fundamentally changed. It changed society. It changed how we work. A lot of the ability to build the structure of civilization depends on the ability of reading and writing and being able to pass on ideas and traditions and all these sorts of things. But you go back, like we are fundamentally sort of oral oriented creatures of not passing down specifics of just gathering vibes.

from, you know, and so you have Homer, absolutely, the king of vibes. And you have this shift broadly just with social media in general and even with Twitter, but particularly when you get to these video apps where the conveyance of information is moving more towards or back to a more sort of oral tradition as opposed to the idea of permanence. And, you know, like you start out, it was books. It's like you're setting this stuff down. And there's an inherent editorial aspect to that because who decides what books get printed? And you have to actually put in the effort. And so there's a distillation and a concentration of information that's inherent to the text process.

And where you fast forward to today, there's no limitation on publishing. Everyone has a camera on their phone. Everyone can publish wherever they want. And on one hand, this also gives rise to the necessity. and usefulness of AI chat robots that can go through all this stuff and pull out what's relevant. But there is a withdrawal from this distillation process. Like the distillation were by necessity handing off to AI. Because to your point, if we're getting meeting notes generated from every meeting,

like someone has to actually, like, who's going to actually read that stuff? It creates this weird dynamic where you're generating content with AI. that then you need AI to process. That's right. But you realize that that is how progress, where by progress, I'm not saying capital P, like small P in more a bunch of stuff changes.

That's how it happens. This kind of goes back to the bank point. You like progress in one area does not stand alone. It has a systemic impact that ripples through everything out. And so in this case, we're talking about society level impacts where, yeah, if you can generate a bunch of text, the reaction is going to be,

you need something to serve through that text and surface it. And you sort of play this out and how that impacts the way people interact. You look forward like, oh, that's going to make a big deal in six months. You look back six years or 60 years. The fundamental structure of society is totally different. We can see that now with the internet, right?

Just the way the world works is so much different today in 2023 than it was in 1980s. or even 1999 through in the internet came along. No single change stands alone. And part of our job is to try to forecast and figure out what those are going to be. But it's also very hard to, it's in some projects hard to predict because it's hard to break the future. But this is also why you can make predictions. Because you think through the systems and changing variable over here has an impact on the variable on the way over here.

Yeah. I'm going to bet on the written word because you mentioned. it's an aberration. It's like a several thousand year aberration. So I feel like it's been pretty durable. And I wonder, I'm not a psychologist, but I think, or I guess a scientist, I don't know. I feel like people retain. You're a podcastologist. I'm a podcastist, for sure. I think people retain information more effectively when they read it, but I could just be speaking from my own experience. And maybe all of this is going to evolve. And I'm just like King Boomer here. on the podcast about to get passed off, passed by.

I'm not saying the progress is better. Like, like, the normative, the normative statement is a separate one. Yeah. Yeah. Yeah. So we'll see. The artifact question interested me, though, because I heard your discussion of artifact on dithering like a month ago.

And I sort of came away with the same takeaway where it's like, I don't think Ben really believes this can succeed, but he's enjoyed playing with it for a couple days. But you keep sending me artifact. links like every day, generally trolling me with Formula One news, but clearly something's working there. Yeah. I do find it useful. It is a nice diversion, particularly, you know, as Twitter's usefulness in spreading links

as diminished over time. And I think the long run, the optimistic take here is this is actually sort of a backdoor takedown of Twitter. And so it's useful because it's already more useful today. day. Like, you have an F1 list that you shared with me for a while, and I have another one. And by and large, it's not very useful. Like, reading F1 news and artifacts is actually way better. I can get in and in 15 minutes, I can basically have read everything that's happened that I might even care about. I can share a couple links with you. The bit about I know when you open a link sounds very creepy, but I actually love it in practice.

Oh, wow. I didn't know that. Yeah. So if I send you, you know, some sort of link about Lewis whining, then I see you open it and then I can drop in the WhatsApp. chat, you know, two minutes later and taught you about it, right? It's very, very efficient in that regard. So, so this is the, but that, that's a positive bit. Like, wait. That's the bowlcase. Yeah. It's providing real value to me. And so then you can imagine layering on, why do we have to have that conversation on WhatsApp? Why can't we have it? If I see you click a link, why can I click in the link and immediately respond to you? And then we're within the chat having, within the app having this conversation. And you're starting to back into a social network in a certain respect, which is, by the way,

how they built Instagram, right? Instagram started out as you have these, these cameras that take really crappy pictures. So counterintuitively, let's make it look crappier by putting these filters on it, and then we'll have all these capabilities to share it to Twitter, to share it to Facebook, to share it to all your social network. It was a utility.

And then, oh, let's also make it, well, we can have a list of your friends, which, by the way, we scraped off of Twitter, and we can actually see them all in one place. And that's how Instagram didn't start as a social network. It started as a utility that backed into a social network. And it sure seems to me they're doing the same playbook now. And what is also interesting, it may be the case if my thesis about Twitter is correct,

which is it's not particularly good for monetization for lots of different reasons. Number one, they're already successful. They don't need like there's a bit where they can afford to go after a relatively smaller market opportunity because they've already made it big. Number two, starting out with the algorithmic basis from day one, that is, you're already in a better environment for advertising from a certain perspective, because that's the sort of expectation that's sort of built in and you're already building the capabilities of deeply understanding whoever you're interacting with and showing them stuff that might be relevant to them. Number three, to the extent you can create an environment that is more talking about stuff as opposed to your currency is dunking on other people and making people upset. you're just in a better mood when using it, which is also better for advertising. So I think it's a worthwhile bet. Now, my skepticism about just the fundamental nature is there, but it is countered by

the fact that I like it. I enjoy it, even with all the sort of problems that I talked about before. Yep. And the interesting parallel to Twitter is it's like, when I open Twitter, how frequently do I close it feeling better? And with an app like artifact, I've not tried Artifact. yet, but like a parallel there for me, you know,

this is embarrassing. I still love the Instapaper app. And when I open Insta paper, well, I don't know. It seems like it's sort of a vestige of an internet that doesn't exist anymore. But when I close Instapaper, I feel good. I feel better about my day

and everything else. And so there is sort of that aspect to it where I think if you could create a more positive way for people to consume news, that's one of the biggest problems I have is that I still get a lot of my news from Twitter, which is this app that is shitty in all sorts of different ways. And I have a real love, hate relationship with the dynamics on there. So if there's another way for people to get news in some sort of a social context,

there is definitely upside there. The question is whether people are really, I think the emailer is spot on and that like convincing people to actually adopt a new app at this point after the last 15 years is going to be a real challenge. Yeah, there is a bit where it would be nice if like artifact and substack were one company, right? Like, and substack had some sort of bundle plan, which again, due to lots of decisions they've made, it's to be hard for them to institute, but it is clearly where they need to get to. And so you had access to all this information in a way that writers could be consistently compensated

for it, but also you could find stuff that that's really meaningful and useful, right? Like all, I actually was in a text message with Gruber about the bank stuff and we were sharing some links back and forth. And there was one article that was very, very good. I think we put in the show notes last week that sort of summarized the crisis. And he observed to me, he's like every link within this article was to another substack. And the point was, was not that people were, he's being super pro substack. That's where all the good information was. Like that's, you know, like, yeah, what are you going to do?

Go read like, you know, some free site that has ads everywhere. and they're just like trying to turn out the lowest comment in the November content. No, if you can find stuff, like if there's a motivation to actually invest in writing good stuff, you're going to get better stuff. And boy, it'd be great if, you know, substack is trying to do better sort of monetization. But one of the, you know, the real, let's do a diversion on substack. The fundamental mistake I believe substack made early on is they were two pro writer. Now, that sounds kind of weird, right?

Because they want to get the writers on there. But the problem is that what's not in their contracts with writers is the, right to bundle, for example. Or like, I feel they should have leaned into their differentiation. It's not where the ultimate in writer-friendlyness, we're the ultimate in ease of use and customer acquisition. And in exchange, you're going to be, except some amount of lock-in to substack and some amount of,

as we develop this business model, you're going to have to trust us that we're going to deliver more readers to you over time by virtue of being able to put you in a bundle or to do sort of X, Y, Z. And instead, they're going to have to go back and negotiate with the Matthew glaces of the world. Why should he join a bundle when he's making a million dollars a year or probably more than that now? So they're going to have to like figure out how to get over that hump with their established writers if they want to get it all together into one bundle, which in my assumption is the way they sort of reach a size and scale that would be better

for everyone in the long run, but you have to actually get there. This ties into this bit about sort of recommendations. Because you can imagine an artifact where if artifact got really, really good and it actually consistently surfaced the best article about the Microsoft demo, that's tough for me, right? It's good, it's good for me to be the fault, oh, if I want to read about tech, I'm going to read Ben Thompson. But if we, like, we, you think about this idea of AI and what it might be capable of, you can definitely imagine a world where actually artifact, because there's access to, say,

all the substacks in the world can actually surface consistently the best article on any given topic where the author becomes less important. And that's arguably good from a user perspective. It's arguably good from the underlying platform perspective. And it's much tougher for the authors. And you can see this is a replay of the newspaper thing. You go back to the initial internet. It's like, oh, I want to read something.

I'm going to go to a publication and then read it. Google comes along. It's like we don't want to, you think about the index idea. They're indexing sites. Google discovered pages. Google changed the most important organizing piece on the internet from sites to pages. And we're going to actually find that specific page that is most relevant to you. Today, in a substack world, in a strategy world, the organizing principle for finding good content is authors.

And that's good if you're top of the heap. It's good for me. But is that the best most optimal outcome, you could see a world, particularly with AI, where it can actually find the best content on the topic as opposed to just the best author on the topic. Well, I am here for the re-bundling to take place. So let's get Artifact together with Substack. And it's amazing. I told you it might be bad for me.

How to be careful what you waste for. Yeah, that's true. I don't know. I will say this, though. You talk about the crap that shows up on Artifact and like the reminder of what the free internet looks like. I've become a big Substack user. I read the Wall Street.

and the New York Times, which has fewer ads. And you forget, like, if you go to Sports Illustrated's website, it's so hard to just read it. Like, the UI across the general internet is so piss poor. It's a death spiral because it's not just that you have a worst product in basically every aspect. You don't have the best writers because they are capable of these new business models. You don't have the best content.

And so you're like scrambling on how to monetize it. And your product just becomes worse and worse and crappier and crappier. And yeah, it's a death spiral. Between that and then you look at Twitter, another pretty rough UI these days. So there is room for some new products here. And we'll see what it looks like. Yeah. To your point, I love text.

So I am unabashedly cheering for both substack and artifact. I love the idea of this model is broadly available. I could do that well acknowledging that. ideal outcome for them is probably bad for me. But hey, that's why we're podcasting. Yeah, well, here we are. Here come AI voices for podcasts. That's what we really have to be worried about.

Episode 51 in the books. All right, Ben, we will come back later in the week. And people can email us at email at sharp tech.fm. This is fun. And I look forward to keeping it rolling. Sounds good. I'll talk to you later.