Interna Talks 3 - Privacy vs. Convenience: Exploring the Trade-Offs in the Age of AI

In this episode, of our monthly fireside chat, we discussed the role of tech leadership, specifically the chief compliance and chief information security officers, in dealing with compliance and security matters related to AI. The importance of the Chief Technology Officer (CTO) understanding how AI can benefit the business and being aware of the inherent risks involved was also highlighted.

The conversation delved into the challenges and opportunities presented by AI in various contexts. For instance, the adoption of generative AI and its impact on the environment was explored, as was the evolving role of the CTO in handling data engineering and AI system creation. Questions were raised about the necessity of having a Chief Data Officer and whether that role should be integrated with the CTO's responsibilities.

Compliance with regulations, such as the upcoming EU AI Act, similar to GDPR, was identified as a significant challenge for AI in social media and other high-risk sectors. The podcast also touched upon the importance of privacy and government oversight in the context of everyday apps that have control over users' information and lives.

Additionally, we shared their thoughts on topics like the best approach for layoffs in tech companies, the current hype cycle for AI, and the future of technology frameworks, databases, and mobile app development.

Transcript:

David Subar:

Hello everybody and welcome to Interna Talks, our monthly fireside chat. We talk about the intersection between technology, product management and business, things that we've seen working with our clients and general trends in the industry. Today we're talking, Mark Golding, Jeff Yoke, Eddie Schack and I are talking about a few things that have come up and a few evergreen topics. So welcome everybody.

David Subar:

So the first thing we're gonna talk today is about the Everything app, particularly Elon Musk's push to make X and Everything app. This has been done in Asia before, particularly in China, particularly with WeChat is one of the examples. So we know there's existence proofs of people wanna do this, have one app with a whole bunch of stuff in it. Don't really have it in the States yet. So my first question is, good idea, bad idea. What makes it a good idea? What makes it a bad idea? What do you think? I'm gonna start with you, Eddie. What are your thoughts?

Eddie Shek:

My thoughts, I think you used the word that there's proof that people want it. I'm not sure in the case of whether it's this whole geopolitical aspect of it that we're not going to get into, but I don't know if it's a matter of people wanting it or that they have to have it. But back to at least in the U S or U I'm surprised that no one has really tried it before obviously we constantly is trading off privacy. Obviously the concern is privacy. when a company, when a vendor has all of your information control aspect of your life in the context of a everyday app, right, you are giving up a lot of privacy, but we are constantly trading privacy for convenience, from letting Google Map track wherever you go to saving passwords. So we're constantly doing that. So I think from a... purely for a lot of people, that's something that they would want is just for the ultimate convenience in terms of online experience or just life in general. But for if there is, I'm sure there's a big part of the population is not going to be comfortable with that. And at what point does government oversight comes in? There'll be a thing probably the opposite kind of geopolitical. push and pull that happens in the US or EU compared to say in China. I think for a lot of people it would be a good thing, but for mankind, I'm not so sure.

David Subar:

Got it. So you think bad thing that people are going to use anyway. That's your vote.

Eddie Shek:

I think a lot of people will use it without having second thoughts.

David Subar:

Yeah.

Eddie Shek:

because

David Subar:

Jeff?

Eddie Shek:

it's, you know.

Jeff Yoak:

I agree with everything that Eddie said, but to address another aspect of it, we have a lot of good experiences with organizations doing what they're brilliant at and trying to do as little else as possible. And everything is a very hard thing to be brilliant at. Certain things drift that way. Apple controls its ecosystem in a very tight way. Facebook has expanded far beyond. what it was originally envisioned to be, so it's Google. But it was a gradual thing and not with the goal of covering the entire map. X is already failing to do some of the things that it's trying to do well and only grow and gradually towards them. And an attempt to make a big step into handling payments, handling other kinds of communication, handling connectedness of various kinds with other people. Um, in addition to questionable questions about privacy and, and other, uh, and control and, and other things, uh, that come from unification, even if it works well, um, it's also going to be a very hard mission to execute on. It's something I would expect there to be a lot of problems with. So in addition to the problems, uh, if they're successful, there's also going to be a tough transition as, uh, they try to build it.

David Subar:

So I buy that and I'm going to jump to Mark in a second about this, but I think that there's another player here, which is Apple. And Apple wants to control the App Store. regulations that you're trying to open the App Store. But if I have an everything app, but my everything app doesn't have everything, I'm going to want to add other apps into it. Those other apps might be ones that I built, or ones that a third party built that I want to add in. X, I'm just making this up. X might want to add in PayPal for financial transactions. Now this creates a business tension. between the owner of the everything app and the owner of the platform, the OS that lives on, this case, Apple. So I think that adds some complexity. I don't know, Mark, what your opinion's on the subject in general or this tension that might exist between data collection and the different platform owners.

Mark Goldin:

Yeah, it's a very interesting question. I've had a chance to think about it and of course listen to you guys talking. I think I'm more aligned with Jeff that the ability for anyone to pull this off, I think the bar is extremely high. This is gonna be tough. There are a lot of challenges and to be good at everything, it's a stretch. So you could separate the question of, is Elon Musk gonna be good at it from the question of, is this a good idea to have an everything app? I've learned not to bet against Elon Musk. He's very determined, very capable, but on the other hand, Nobody's infallible and it just doesn't seem to me feasible that he's going to be able to put everything together in a single, well-crafted, easy-to-use app. Besides, there's a trust factor. I think he's kind of jumped the shark a little bit and there are large numbers of people that wouldn't entrust him with all their data. Could somebody else take it on? Possibly. As to the question of the platform, if Musk is determined to make this work, I think he would find a way to bypass the platform. He'd create an app so compelling that you would buy a different phone. just to get out of the clutches of Apple. That would be my hypothesis if he wants to be successful. As was done in China, Chinese apps are really independent of the platforms that we are obsessed with here. They

Eddie Shek:

Yeah,

Mark Goldin:

have alternatives,

Eddie Shek:

I absolutely agree

Mark Goldin:

so

Eddie Shek:

with what

Mark Goldin:

they

Eddie Shek:

Mark

Mark Goldin:

might

Eddie Shek:

is

Mark Goldin:

happen

Eddie Shek:

saying.

Mark Goldin:

here.

Eddie Shek:

Yeah, I absolutely agree with what Mark was saying. It's, it's, I wouldn't be surprised that part of the, uh, Elon Musk playbook for taking over the world with this, uh, everything app includes the delivery platform, a new phone, he already has Starlink, which is the networking part of it. Uh, wouldn't surprise me part of the playbook includes the device.

Jeff Yoak:

There's a great parody out on the internet

David Subar:

I for one welcome

Jeff Yoak:

of that.

David Subar:

our ro- What's that, what's that, Jeff? Say it one more time, Jeff.

Jeff Yoak:

There's a great parody out there about the... Can you hear me? Okay. There's a great parody of that out on the internet that announced the... I think they even called it the X phone, even though it was before Twitter changed its name to X. And it was going to have solar panels and self-rela... I mean, it was kind of all the kinds of things that Musk did, but all on a phone. And it was very, very funny. It took a little bit of a way into reading it to even realize it. parody until you realize things like you're not going to charge your phone with solar panels on the phone. You know, you run into something like, okay, this is a joke that even a crazy person would suggest if it's quite

Mark Goldin:

Let's

Jeff Yoak:

funny.

Mark Goldin:

not forget his other initiative. What's it called? The one that is a

David Subar:

Neuralink?

Mark Goldin:

Neuralink. Exactly. You could have the whole thing embedded in your brain.

Eddie Shek:

Peace.

David Subar:

So I'm reminded of what Ken Jennings said on Jeopardy once when he lost to Watson. I, for one, welcome the robot overlord. So Elon continues to take over. That's our prediction. Let's move on. Another kind of hype factor, gen AI. Some use, some not use. I mentioned the hype factor because it seems as if every company, every startup, everybody's talking about GEN.AI everywhere. Now, it can't be true that everything's going to be replaced by GEN.AI. It's certainly true that there's places where GEN.AI has value that we haven't created yet. So my question is, how much of this is real? How much of this is hype? And what will that do to our future, both our future investing and future of products? I'm going to start with you, Jeff. What are your thoughts here?

Jeff Yoak:

I think that in general, and I think this will sound intuitive and natural except when you apply it to the exciting hotness of the day, we have problems and we look for the technology tools, platforms that will help us to solve those problems. And sometimes as now, there's something and right now what it is AI and large language models in general, AI in particular, where you'll find people asking people like us, What can we use that for? How can we bring that into our organization? And that's always a big red flag for me. Um, if there's no problem that's presenting itself or opportunity, it's presenting itself that this is the solution for why not leave it alone. Um, I think that, um, we actually are, we're not the beginning of this process for 25 years, we've been building intelligent. Well, let me avoid that word. That's speculative. We've been building predictive and algorithmic systems that do things like optimize advertising, things like that. We are seeing some of a non-linearity with the emergence of things like large language models, but I think that there are places where the application is obvious, for instance, in trying to set up automated customer service systems or generate text or copy of some sort which requires human editing. where I think experimenting with it is natural. And I think where we are in terms of the stage is those things still need a lot of human oversight and are very risky to have be complete replacement of human beings. I think they can augment humans and allow the workflow to be tremendously magnified more than replacing them in most cases. Coding is a good example.

David Subar:

Yeah,

Jeff Yoak:

You still

David Subar:

so

Jeff Yoak:

need

David Subar:

I'm...

Jeff Yoak:

to check the coding.

David Subar:

Yeah. Well, maybe not, right? Maybe if you have inputs and you validate the outputs, you don't care about what it did in between so long as your inputs and your outputs match in a way that you like. Maybe you don't need to check the cutting.

Jeff Yoak:

Well, you do it, maybe you can do it in an automated way, but you have to do things like test edge cases, things like that. And what you end up with is introducing a, you know, a much more evolved QA system. Um, to check

David Subar:

Yes.

Jeff Yoak:

those mappings of inputs to outputs and, you know, if it works in your particular area, I think going with that is fine.

David Subar:

Yeah. Well, I'm reminded, beginning of my career, I think you all know, audience may or may not know, I did research and development in AI and ML at a military owned think tank in DC. And we were doing, just imagine there was big cameras in the sky looking down at things and you want to identify what those things were. Just imagine you might be in that kind of situation. And we were doing some, some really interesting things. And then we didn't get to. In the AI world in general back

Eddie Shek:

I think

David Subar:

then,

Eddie Shek:

you're all done with today.

David Subar:

we didn't get to the hype that was pitched. And so we entered this AI winter. And this AI winter that we just started coming out of, call it a year or two ago, and we're going back on the hype scale. So to Jeff's point, I think it's true that you need to. use this technology like every other technology to create value. But what I'm concerned about is, what I'm concerned about is, are we over hyping this? And we're going to get to another AI winter and another venture capital winter and another Nvidia stock drops because suddenly people aren't using it for AI. I'm sure that by the way, I'll show you there's, we'll find something else to use with GPUs. It's very handy. But Are we going to another hype cycle? Are we entering danger? Mark, what do you think?

Mark Goldin:

There is a hype cycle for all of those things. You could certainly apply that to this. There have been winters, AI goes back to the 40s. I mean, it's an incredibly, in a way, an incredibly mature technology that has come along and leads some bands. I think it's very hyped right now. I think it's going to settle into the steady growth phase where you'll see early adopters and later adopters coming aboard. I don't think it, yes, it's overhyped, but I don't think you're going to see it go back into a winter situation. I give you a couple of proof points. I find that, you know, as a consultant, I talk to a lot of people. And of course, people in the sort of, in the technology sphere are very anxious to embrace the technology and not be left behind. But I'm finding people in say nonprofits. There's one small nonprofit that I consult with where the executive director is just, she's exploding with ideas. What can I do with AI? And she's finding many really good applications for the technology. And she's pressing me to find her ways to incorporate that. into her stack. And I think that's very, very encouraging. They make sense and I think they're real. So yeah, no, I'm a big fan. I think it's happening. And chat GPT sparked this latest wave of enthusiasm, but I don't think the fire is going out anytime soon, if ever.

David Subar:

Okay. Great.

Jeff Yoak:

If I could add something to that, I think there are real winters and perceived winters. I think the first AI winter was a real one. People really stopped, well, dramatically reduced the amount to which they were looking at it. But we go through hype cycles where companies, there was a time when if you were a company, you were doing what you were doing plus you were a blogging platform. And another time when you were doing what you were doing plus you were a social network. Well... We didn't stop building blog, new blogging platforms or improving them or social networks. Those continued. But if you were one of the big groups that jumped on the bandwagon, it probably felt like a winter. Because all of a sudden just saying the equivalent of I'm a prompt engineer didn't mean you all of a sudden had job opportunities and everybody was really excited because people were properly rejecting

Eddie Shek:

I'm

Jeff Yoak:

the

Eddie Shek:

in.

Jeff Yoak:

hype. But real work is still, was still going on in areas like social networking. I think AI may be the same. At some point, the hype may die down. And so the weaker players, the more or less technical, the more casual people may feel like the environment's disappearing, but I think we'll see real advances in the underlying technology and its utilization.

Eddie Shek:

My concern is more, it's specifically that the hype is all about generative AI. Generative AI is just one branch of AI technology and research and activities in general. Basically, all air has been sucked out of everything in AI other than generative AI. That's actually more of my concern because there are generally just one technique to solve a particular set of problems. There are all kinds of other supervised learning, similar regression, you name it. Other kind of techniques for other kinds of problems that we can use AI for. It's just, they're just all quiet.

Jeff Yoak:

You know, this was beyond a redone, I think that I didn't mean to cut you off.

Eddie Shek:

No.

Jeff Yoak:

Um, it wasn't part of our original question, but it just occurred to me, uh, one of the things we should, might want to look at is the, some of the negative aspects that are growing out of it. Paul Graham, for instance, has been talking a lot about, um, spearfishing and we all know what fishing attacks are, but the AI is now getting, you know, probably this year to the point where it can analyze your organization, understand who your boss is and then attack you by impersonating your boss. We start to see some of those things. I'd be interested to know what other people think, you know, the impact of that, if they think that kind of thing is going to happen, where this technology is going to start to be brought to bear, where attacks can be made much more personal because they don't have to be done by individual humans and what, what that, what the level of impact will be over that kind of thing.

David Subar:

Actually, so I think there's a really interesting play here. there are systems where people are creating AI to create fakes and creating another system to detect fakes. And they're self-trained, they train each other. So for instance, one, so basically the systems are in competition. One will create a fake that will feed those fakes and real images into the detector. The detector will try to detect what's real or what's fake. and then be trained on how well it predicted what's real and what's fake. So let's say it got 50 fake and 50 real, it predicted them inaccurately, it'll train on that. And then that training, what it predicted was real and fake, will then go back into the one that generated fake stuff for training. So they're aggressive systems to each other, training, in this case, spear phishing attacks and protection against spear phishing attacks. I think this is an arms battle.

Mark Goldin:

It's an arms race. Yeah, exactly. We're going to see that again. I mean, look how bad the, uh, the, the fishing emails that one gets that you can immediately tell them, especially if you have some experience, you can tell them apart right away. The language is so poor. The phraseology is incredibly weak. It's obviously written by somebody whose English is not first language. Imagine using chat GPT to generate better emails, right? To do some spear phishing attacks. There's

Jeff Yoak:

So

Mark Goldin:

so many possibilities here.

Jeff Yoak:

an interesting thing about those, those are actually made transparent on purpose because the next step involves human time and they want to isolate the most naive audience. The ability to have something like one of these AI things do the next step as well means now they can actually try to fool you. They can actually change the incentives.

David Subar:

So I want to move on to a related topic. Okay, so we have this gen AI and some's hype and some isn't. Some is negative for the environment, some is positive for the environment, but we know there's going to be a role for generative AI. We deal with CTOs and Chief Product Officers. How does this change the role of the CTO? How does this change the role of the CTO in worlds where there's a chief data officer, someone who might be responsible for data engineering, pulling in the data, data science, building an AI system to create some kind of output, some kind of data AI to look at the output the way we were talking about before, seeing on retraining? whether the output is drifting, understanding whether the output is drifting because the training was drifting or the data input's drifting. These are all new things that we never had to do at scale before. Some companies, I said, are having a chief data officer responsible for this. For some companies, it's getting folded under the CTO. How does the CTO roles change, chief product officers role change? Should there be a chief data officer or not? Eddie, what do you think?

Eddie Shek:

I don't think it changes the job per se. To me, I may be in minority, like generative AI or AI technologies in general, I think of it as just new tools, new system, new tools. So just like any CTO, we'll have to keep up to date on the latest, greatest tools that allow the company, the team to build whatever they need to build. It's just continuation of that spectrum. Right? Does it get infinitely more complicated? Yeah. And changing infinitely faster than ever? Absolutely. But maybe more specialized skill set to deal with that. Does it create, turn into a new role for that? Maybe if you can slice and dice it in different ways. But to me, that's the other thing. It fundamentally changes the job. But one thing that's kind of interesting, I don't think that, I don't get, I don't see a lot of news or word on. eggs on social media is the whole compliance aspect of it. The EU AI Act is basically the equivalent, the GDPI equivalent for AI and that is moving slowly but surely forward. And they're looking at my finger on the glass I heard that they're shooting, hoping for end of year for approval. That is going to be crazy. Imagine that difficulty a lot of people ask, a lot of us dealing with GDPR, if we do business in Europe, that is going to be the year 2000 problem for people who use AI and do business in Europe. Like for example, chat GBT or GND AI will fall into the category of high risk and now you only have to provide explanation, right, to provide transparency. That is something, it's something that a lot of tools are ready for.

David Subar:

I want to put a pin in there. I want to come back to that just for a second. But I want to go to... the first point about whether the CTO's job change. And I'm going to argue with you for a moment. Because the skills that you needed for managing teams of developers and org designs for developers, you do still have developers for the data engineering. But maybe you manage the data scientists differently. Maybe you manage QA differently because you're not inspecting code. You can't inspect code. But you are. much more that automation that Jeff was talking about earlier, getting the inputs and outputs, maybe that's much more important. And then, then going back to your second point about the GDPR equivalent for AI, maybe you have to understand that differently. Is your argument at E that those are 10 degree shifts, not 90 degree shifts in the CTO's job? I'm glad with that argument if that's what

Eddie Shek:

I

David Subar:

your

Eddie Shek:

see

David Subar:

call

Eddie Shek:

it as a

David Subar:

is.

Eddie Shek:

continuum, just that the horizon has just gotten much, much bigger. I don't see it as a directional change.

David Subar:

Any other thoughts about the GDPR kind of thing? Mark, I saw you smile. Yes.

Mark Goldin:

I actually am more aligned with Eddie in this one. I think, yeah, there are new challenges. David, I would agree with that. But I work with the tech leadership and I would include the chief compliance or the chief information security officer, the guys that deal with these compliance and security related matters and for the CTO, understanding how AI is going to be beneficial to the business, understanding the risks. Those are skills that are going to be very valuable to have. As we were saying earlier, people are confused. They're seeing all the hype. They don't know what to do. the CTO has to make a lot of sense of that and help people understand where the opportunities lie and how it can be applied to the business. And for the CISO and the related roles, CISO privacy offers to the guys that really need to understand how to sell or how to operate successfully. And Europe, fantastic opportunity. Again, it's a challenge, but to make sense of that, explain how to deal with these issues is just a wonderful thing. Again, like a lot of the things we've been discussing here, nothing remarkably new here. We were a previous company, we were selling recommendation systems into European customers, and they wanted to know what's inside the black box? How are you making the decision? Those questions only came out of Europe, it's for some reason, they're much more compliance oriented. And here too, they're gonna wanna know what was the decision making process of the AI? How did it arrive at that conclusion? Open up the black box for me, that's very challenging, but it's also an opportunity.

David Subar:

Yeah, essentially I worked at a fintech company in the States and when you turn someone down for credit, you have to explain to them by regulation. impossible to do

Mark Goldin:

Very difficult.

David Subar:

with the neural network. You can just say here are the inputs and here are the outputs. So

Mark Goldin:

Yeah,

David Subar:

we're going to

Mark Goldin:

I was on thought on that. I've heard some experts debating this point. Some say like you said, oh, you can't do it. It's impossible. And others say, well, you know what? We need to instrument it that way. We need to build it that way. This is a solvable problem, but nobody's working on solving it. I don't know the answer for sure, honestly, but I would tend to believe there is a way of solving that problem if we consider it important enough.

David Subar:

But let's jump from there to just general startup tech. What are people using? Has it changed? Has it not changed? Not just for Gen.ai, but just in general. Gen.ai, there's a whole giant set of tools, which I think are out of scope for us to talk about now, Hugging Face, things like that. But let's talk about the standard procedural code that we've been using since effectively Turing. What's the new hotness? Is there new hotness? Mark?

Mark Goldin:

Seeing a couple of things which are interesting, I think Django has really taken hold in the startup community. Python framework, obviously, I think a lot of people are gravitating towards that. I've been shocked to find that Ruby on Rails is still alive and well. I keep thinking it's gonna die. I guess it's just my prejudice, but it appears to still be out there. It's also this debate about whether to build native on mobile or not is still raging. I think native React, React Native, I should say, has taken hold to some extent. If you have a small team, it's obviously a lot cheaper to build in React Native than to have to build purpose, build apps for every platform. Just love to hear what other people are saying. And of course, when it comes to databases, NoSQL remains popular, but then again, so does Postgres. Postgres and AWS is often a go-to choice as well. That hasn't really changed in a big way. I'm curious to hear what other people have to say on this.

David Subar:

Jeff, Eddie, you have any opinions here?

Eddie Shek:

Um, I don't see anything new because there's just every, all the air been sucked out of like innovation just by the gen AI. Um, so a lot of what Mark is saying, I agree with, but they're not new either. It's all the, some of the trade-offs and arguments kind of continues with data versus a hybrid app. Right. So I don't have a whole lot to add there.

Jeff Yoak:

I have seen... Go ahead, Mark.

Mark Goldin:

A lot of people working on Python backends, there's an initiative now to improve the speed of Python quite dramatically that people are pretty excited about. So there's one that I might have thought would have kind of faded out, but in fact the opposite has happened. It is not. It has really taken hold.

David Subar:

Our

Mark Goldin:

It's getting...

David Subar:

argument is there's no new hotness, there's just a bunch of old hotness.

Mark Goldin:

Hotness improved maybe, yeah.

David Subar:

Yeah.

Jeff Yoak:

out.

David Subar:

Make your hot even hot. Make your hotness even hotter.

Jeff Yoak:

Following that theme, the thing that I've seen the biggest spike in recently is actually Golang. And the last big spike in Golang was 10 years ago, when it kind of reached broad market awareness. So it's an older technology that's as a percent, I actually saw some statistics on this. It was the biggest jump in the last year for implementation of platforms and start-up tech, which surprised me because where did that come from all of a sudden? And I don't know Go, so I don't know.

Mark Goldin:

At

Jeff Yoak:

Sorry,

Mark Goldin:

the same

Jeff Yoak:

Mark.

Mark Goldin:

time, some of the old battles have been decided. For example, for a number of years it was Angular versus React, and each one had its adherence. I think you could see it was going towards React. I think you can now finally say React 1. I don't know if you guys agree or disagree, but I think

Jeff Yoak:

I agree.

Mark Goldin:

that one is pretty apparent.

David Subar:

It's still funny for Liz.

Jeff Yoak:

Hey,

Mark Goldin:

I love it.

Jeff Yoak:

a new version of Pearl just came out. I'm just

David Subar:

Yeah,

Jeff Yoak:

saying.

David Subar:

I'm sorry for everybody.

Jeff Yoak:

Hahaha

David Subar:

Let's move on then. So

Jeff Yoak:

CLEARS

David Subar:

companies

Jeff Yoak:

THROAT

David Subar:

are still doing layoffs. The JNI companies keep, for some reason I keep talking about JNI in every question, the JNI companies are hiring, but the folks that haven't gotten on the hype train or the hype train is not inflating them, bunch of companies are still doing layoffs. So you're a CTO, you're a Chief Product Officer. What do you advocate for? Do you go with the layoffs? Do you maximize the layoffs? Do you just try to get one big cut? Do you try to do the minimum? Obviously every company is different, but how do you think about the problem as a CTO, as a chief product officer, advocating for change? Big or little, and when do you advocate for one or the other? Mark, what about you? Let's start with you.

Mark Goldin:

So I always think first about the remaining team. Think about those that are left behind and the impact on that team. Of course you have to be compassionate. That almost goes without saying when you're laying people off. To the extent you can afford to be compassionate, be compassionate. But you've got to think hard about the remaining team and the impact on morale. Successive waves of layoffs will destroy that bond of trust if there ever was one between the employer and the employee. And so what do you do about that? One is minimize the number of layoffs. If you're gonna lay people off, get a good sense of how much you're gonna have to do and do it all at once as opposed to doing it in phases. Because then you can say, you know, we've done it, we don't have any more planned and you can move on. And people will breathe a sigh of relief and maybe not worry about it so much. The second thing you need to do is treat people as well as you can afford to. Number one, it's the right thing to do. Number two, that also has an impact on existing team. If they see that you've communicated well, you've been transparent, you set expectations and you've been generous, then they'll feel relieved that if the same thing has to happen to them, that you'll act towards them in the same way. So that's my message. Think about the remaining team.

David Subar:

Jeff, you have any thoughts on this?

Jeff Yoak:

Yes, and a big part overlaps with what Mark just said, and I agree with that completely. I lean strongly towards preferring doing it in as few rounds as possible, you know, a big one up front if that's what it needs to be and hopefully not more. And I would regard the worst possible approach as say planned quarterly layoffs. And that's because I believe that The worst thing that any tech company can do is create an environment where the people who can leave do, and the people who can't leave don't. And if people feel that this company's in trouble or my job's in trouble, the people who can most easily leave, which are your best people, and you may or may not know who they are, find other places where they're less afraid of that problem. And you don't want to be left with the 80, 90% that are left. So, you know, there's more that you can do that I would have first emphasized the exact piece Mark did about treating people compassionately, having everyone see that you treat people as well as possible. That kind of falls all on both what Mark said when I said already fall under the category of, you know, managing with your peers in the organization, how the organization approaches layoffs. Another thing that I think is important is kind of down, managing down. applies to their layoffs or not, which is people should really understand where they are in the organization. A very extreme, which I've never done by the way, but very extreme version of that would be like a Jack Welch style differentiation, where not just the people know, but everybody in the organization knows who the top 10% of people are. Not necessarily the most senior, but the people who are performing best at their job and everybody knows who the worst 10, 20%. And when layoffs happen, it's going to be those people at the bottom. And you actually don't care so much if those people, you know, you're hoping, helping, hoping to nurture them and help them improve, but, um, they're the least costly if they do leave. Now I've never gone that far. It may not be real appropriate for our industry, but at least through review processes and informal communication, um, people ought to know how safe they are from that and you really think need to step up that kind of attention when layoffs are on the table. so that the people who can leave that you don't because they know that they're the people who you value most strongly. And vice versa, the other people may select themselves out or at least they know they need to pull their performance out and it matters.

David Subar:

Well, we see, we see, maybe unfair to Jack Welsh to say this, but GE is performing so much gangbusters. It's the number one, it's S&P 500 for the last 30 years. So there you go. Thank you, Jack Welsh. Hope you didn't have GE's, Doc. Well.

Jeff Yoak:

But it did better when he was there.

David Subar:

It did, but it immediately collapsed. We can have our email at jackwelshlater and ge-later. Thank you, everybody. Thank you, everybody listening. We'll be coming out with another few podcasts in the coming weeks, one that I've been on that we're sending out piece by piece over this couple of weeks, talking with the CTO quickly or soon about more about. the CTO's role in a heavy gen AI environment. It's a CTO who is in a very heavy gen AI environment doing it at large massive scale on some of the problems that they're addressing and how they're addressing that should be a very interesting talk. I had a conversation recently with Peter Bell talking about CTO communities and CTO learning opportunities. And then there's some podcasts that we have with private equity firms in relationship between. private equity investors and CTOs and Chief Product Officers. So those are all coming out in the coming weeks. So stay tuned for that. In the meantime, Mark, Jeff, Eddie, thank you very much. And I'll talk to you all soon.

Eddie Shek:

Thank you. It's been fun.

Jeff Yoak:

Thanks, David.

Mark Goldin:

Yeah.