Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Designing Human-AI collaboration to make people heroes not cogs
49 Plays6 hours ago
No one gets fired for doing what McKinsey says. That's not a joke, it's the actual product being sold in half of enterprise consulting: credibility a board can point to when a decision goes wrong. This week, Tim Lidman, founder of Clyde and formerly of a collaboration platform absorbed into Accenture in 2021, argues that AI won't replace that function — and says so without flinching, even though it means some prospective customers will balk.
Tim also shares why he thinks “human in the loop” is such a dehumanizing turn of phrase.
The conversation moves through the politics that quietly derail group decisions, the trap of building for experts at the expense of everyone else, and a real case where a district court judge canceled a trial because both sides had submitted AI-generated arguments neither side had bothered to check.
And this episode ends somewhere unexpected: a founder building frontier AI for a living who gives his own children zero screen time, and why he thinks boredom is the skill nobody is protecting anymore.
Mentioned:
Recommended![Deep learning vs Intuition: AI models for VC investing [Replay] image](https://media.zencastr.com/cdn-cgi/image/width=112,quality=85/image-files/6322f03255d97a700020586e/4d3f7df1-655b-4735-a002-7e17193218c2.jpg)
![The Everything Machine and the Trillion-Dollar Bet [Replay] image](https://media.zencastr.com/cdn-cgi/image/width=112,quality=85/image-files/6322f03255d97a700020586e/f0ee1901-e926-4895-9598-4c323bd806cb.jpg)
Transcript
Human-Centric UX Design
00:00:00
Speaker
But I've also made sure that we've built UX to very intentionally keep not just the humans in the loop, which to me is a very dehumanizing way of talking about how humans are part of the workflow. Just human cog plug in here.
00:00:12
Speaker
But actually keeping the humans kind of at the center of it um and making sure that we are letting humans become beings. kind of superheroes where needed, but then still making sure that those human parts of the workflow are intact.
00:00:28
Speaker
ah And then making sure that we're also to some extent democratizing expertise.
Introduction to Podcast and Guest
00:00:42
Speaker
Yo, this is Bare Knuckles and Brass Tacks, the tech podcast about humans. I'm George K. And I'm George a And today, our guest is Tim Lidman, who is the co-founder and CEO at Clyde. You can check them out at meetclyde.com. And this was more interesting than not. So full disclosure, crew, we get guest pitches all the time.
00:01:06
Speaker
And Tim came in through, i think it was an agency, i We often just don't ah answer them because they're garbage. um And then I pushed back.
AI-Native Collaboration and Clyde's Vision
00:01:19
Speaker
I said, here are the guidelines. If your guest wants a platform to be a commercial, we are not that show. But if they want some pushback and we're going to ask these sorts of questions.
00:01:30
Speaker
And to his credit, Tim was still game. And so he came on and he's here to talk about What he's building, which is he calls an AI native collaboration platform, and we necessarily get highly skeptical, really pressure test the hypothesis. And ah Tim surprised us, I think. One, not because we were dicks, but we were definitely a little adversarial and skeptical of the idea and and really tried to push him on it.
00:01:56
Speaker
Yeah, I think that was the cool part about it. And he took it on the chin and he was a good sport. um And I got like, you know, I kind of told him there at the end, like, look, man, you're a founder who actually have already built and exited from a successful product and you knew the game and you still, you know, you built something where you have a practitioner's level of focus. And, um,
00:02:17
Speaker
I have such a deep appreciation for someone who actually understands the problem statement that their solution is going to market with. And Tim really kind of demonstrates that. So, George, kudos to you. Really good. Fine. Thank you for following that instincts and being a prick to people.
00:02:32
Speaker
You're welcome. Anytime. that produced gold for us. But yeah, Tim is a worth is a
Implementing AI in Business Strategies
00:02:38
Speaker
worthwhile listen. I think he... um approaches how people should look at how they implement AI in their businesses and how they actually use it. And ah Clyde, is his platform is definitely worth taking a look at.
00:02:48
Speaker
Yeah, and listen to the whole episode because towards the end, we do get into the contradiction of the CEO co-founder who also highly limits his kid's screen time. And we have a whole dig into that. So let's turn it over to Tim Lidman.
Clyde's Evolution and Purpose
00:03:08
Speaker
Tim Lidman, welcome to the show. Yeah, thanks so much for having me. Yeah, so we're excited to talk about what you're building at Clyde. um Let me give you ah the platform for just a few minutes. Why don't just tell us quickly what Clyde does and then then we'll get into it.
00:03:27
Speaker
Yeah, absolutely. So to best explain what Clyde does, probably worth explaining why Clyde exists. um so Clyde is is very much off the back of another startup that I was running that got acquired by Accenture in 2021.
00:03:44
Speaker
um So that startup was was Think Tank. And that was a collaboration platform that sort of evolved from... So I kind of grew up in the collaboration space And Think Tank was very much a structured collaboration platform. So what it allowed particularly consultants to do was instead of showing up with a myriad of, you know, PowerPoint, Excel, Miro, Mural, Whiteboards, it allowed you to collaborate with clients in sort of one place.
00:04:16
Speaker
It had decades of behavioral science and sort of group decision support research built into it. And so it was all about aligning people on outcomes and it had sort of all the features that you needed to be able to do that.
00:04:32
Speaker
um And the reason that's important to Clyde is that when Think Tank was acquired, i then operated as a partner at Accenture for about four years.
Automation and AI in Consulting Workflows
00:04:43
Speaker
And this is right when sort of Gen AI hit right in the middle of that. So that was sort 2023, 2022. And what I realized was that, um you know, one of my challenges with with Think Tank was even though it allowed you to get to a much better aligned output, it still required a lot of manual sort of heavy lifting, a lot of manual synthesis, you know, poor analysts kind of staying up till two in the morning, you know, fixing decks, things like that.
00:05:15
Speaker
And so my hypothesis was what, what if you could, basically train an AI on the, I'll just call it the workflow of consulting. It's a super simplified way looking at it. But what if you could train an AI on that and really start to take out a lot of that really, really manual effort that just burns, burns hours and productivity And now the AI has reached a level of capability where you truly can do that.
00:05:47
Speaker
And you can take, you know, what an analyst, what a manager, senior manager does as a consultant and really automate large, large portions of that.
00:05:59
Speaker
um What you can't do, though, is there are parts of that workflow that are still very, very much human, right? And so there are parts of that workflow where you still need to, you know, extract tribal knowledge and institutional knowledge from people's heads.
00:06:14
Speaker
Until they automate CEOs and leadership, you still need to make human decisions. You still need to drive change out and implement things into the organization.
AI's Role in Collaboration and Decision-Making
00:06:24
Speaker
And so what I've built with Clyde is I've i basically built something that where it makes sense, uses ai for what AI is really, really good at as far as synthesizing really large amounts of information, um you know, sourcing very, very large data sets, things like that.
00:06:44
Speaker
But I've also made sure that we've built UX to very intentionally keep not just the humans in the loop, which to me is a very dehumanizing way of talking about how humans are of the workflow. human cog, plug in here.
00:06:56
Speaker
But actually keeping the humans kind of at the center of it um and making sure that we are letting humans become kind of superheroes where needed, but then still making sure that those human parts of the workflow are are intact.
00:07:12
Speaker
ah And then making sure that we're also to some extent democratizing expertise. And so we've built into Clyde in addition to, you know, just AI productivity and human collaboration.
00:07:27
Speaker
We've also built this concept of sort of your AI advisor team where you can bring in expertise from a number of subject matter areas and domains so that a consultant could use this to supercharge what they're doing, but also a non-consultant.
00:07:44
Speaker
We can get into this if you'd like later, but I have third grade teachers and healthcare professionals and people that aren't consultants using this to also be able to get outcomes that they've never had access to sort of in the pre-Clyde, pre-AI era. So...
00:08:02
Speaker
Yeah, that's rough roughly what Clyde does without getting into the the features and stuff like that. No, that's helpful, but we we are going to dig into some of that and collaboration. All of that stuff is kind of my my favorite soapbox to stand on with respect to knowledge work these days. So why don't we start in an obvious place here, which is can you define, and you did a little bit of this, but I'm welcome to step into the mechanics of it or the design decisions.
00:08:33
Speaker
can you define ai native collaboration in a way that isn't just, yeah you know, LLM built onto what we already have? Make it a real thing because it it sounds like hype speak when you say the the words like AI native collaboration. Like I,
00:08:49
Speaker
Honestly, as someone like trying to study for the episode, I was like, this just sounds like hype speak and I'm angry, but you seem like a pleasant person. So please. Let's let's bridge anger to and the pleasantry into something that we can agree on. no so No, so AI Native Collaboration, if if you kind of, the way I define it is if you look at the history of collaboration, or at least the way I would look at it.
Evolution of Collaboration Tools
00:09:14
Speaker
I started my career, you know, trying to convince people to, ah you know, use audio conferencing instead of in-person meetings, right? So that was all about, hey, you could probably just talk about this instead of traveling for a million hours. um I then saw that evolve into, as broadband became more you know available I then saw web conferencing and video conferencing take off. And now you were able to, you know to an even larger extent, do things virtually a hybrid way that you know couldn't do before.
00:09:50
Speaker
um Then social networking hit the scene and it sort of evolved from this unified communications aspect into, okay, now we're, you know, layering enterprise social networking into that, right?
00:10:04
Speaker
So we've got all this unstructured collaboration and, you know, Yammer, got acquired Microsoft, and you had that sort of thing. And then that, and then we'll get to what AI Native Collaboration is. That has sort of morphed now into, know, i called it structured collaboration with Think Tank, ah Mural and Miro.
00:10:22
Speaker
You know, they very much are about the infinite canvas and being able to not just connect people remotely, but actually start to bring in content and workflows, you know, in a more meaningful way into collaboration.
00:10:37
Speaker
And so if I then think of how is that going to evolve in the age of of AI, right? I then think of this category, which, yeah, I'm calling AI Native Collaboration, um which is where one one of the challenges Miro and Miro have right now, and and some of those really popular, I'll call it pre-gen AI,
00:11:01
Speaker
platforms is they're trying to just bolt AI onto their existing stack, right? So they're putting an AI assistant on top of, you know, a bunch of mural boards and post-it notes things like that.
00:11:14
Speaker
And they're asking it to, you know, perform certain tasks. The challenge is it's not AI native. So AI isn't a first-class citizen in the actual architecture.
00:11:25
Speaker
And so as a result, you're you're constantly playing catch up on trying to get the real power of AI into ah into those platforms. It's like Gmail asking me to if it wants me to summarize a two sentence email. it's like, no, Gemini, I do not need you to do that. Yeah, exactly. And so, so it's kind of a, yeah, hammer looking for nail. at Adversely, if you look at the LLMs, right, the LLMs are, and if you look at sort of all the AI tools which where they are first class citizens,
00:12:00
Speaker
they are not designing with collaboration first in mind,
Limitations of LLMs and Clyde's Contextual AI
00:12:05
Speaker
right? And they're not thinking about how do we actually connect people and AI to get to aligned outcomes.
00:12:14
Speaker
And the way that surfaces beyond the hype speak is, that there was, you know, LLMs had this promise that, you know, SaaS is dead and ah SaaS used to force you to align to their user paradigms and LLMs aren't going to do that. You can just talk to an LLM like you would to your friend and your LLM is going to magically do everything for you.
00:12:35
Speaker
The challenge is that LLMs are still requiring you to adapt to them. You still have to be really good at context engineering. You still have to be really, really good at prompts. And if you're not, you know the LLM is going to take you down horrible dark roads that don't get you anywhere.
00:12:52
Speaker
And so on that part of the AI Native collaboration side, what we've built in Clyde is something that one doesn't put the onus on the user to truly understand what you're trying to achieve and the goal that you have. And so Clyde actively extracts that content from you and that context from you as a user.
00:13:12
Speaker
The other thing is knowing when you need actual human beings to be part of your process and when you need to collaborate with them, when you need them for certain decisions, when you need them aligned.
00:13:24
Speaker
And again, Clyde has built features that actually use AI to help you determine when you need to collaborate and how you need to collaborate. And then lastly, we actually have features that aren't just chatbots.
Enhancing Collaborative Decision-Making
00:13:40
Speaker
We have collaborative features that allow you to brainstorm ideas with people, organize those ideas, prioritize them, vote on them together, you know, actually click buttons to vote.
00:13:52
Speaker
ah do decision trade-off analyses together um to actually reach the outcome. And so back to the hype speak comment, I think we're going beyond the hype speak because we have actual features that allow you to collaborate, align and make decisions, like very intentionally built into the platform in a way that I just, I haven't seen another product out there that enables human collaboration and AI in the same platform. So yeah.
00:14:20
Speaker
I find that fascinating though, because what you're saying to me, like because I have to go through this ah at my work for for both of our jobs. Like, how do we use AI properly and how do we actually attain ROI?
00:14:34
Speaker
um And so I think like, instead of just organizations that are just like blasting AI investment everywhere, just forcing it down people's throat, like use AI, and then you get that AI resistance and everything. and And people will they'll see this and they'll see the ambiguity orders that come up from up top. And then they see the headlines of mass layoffs, which are attributed to AI and said people form an AI resistance.
00:14:57
Speaker
right And I think a big part of that, um which is a natural human correlation, um it has to come down to like what is the process you're trying to automate? and And I really like that you're getting into the actual mechanics of like how to build a prompt.
00:15:12
Speaker
um Because i've I've found that in my experiments with it, George and I both had different but but similar deeply um intellectual journeys with it And I've had to look at it from a technical implementation standpoint.
00:15:25
Speaker
How you're actually using it and what are the outputs are kind of like the critical thing, because that because that will determine return on investment. So I really appreciate that you're building a solution based on how do we attain business value and how do you have to think about the problem for you to do that? Because what I'm hearing from you is you've built a product that guides your customers into how to think about solutioning correctly to get the most value of their AI. Is that generally correct, sir?
00:15:53
Speaker
Yeah, because I mean, the first part of Clyde's workflow is when you give Clyde problem, right? You could tell Clyde, ah you know, my sales process is broken. Or you could say, ah my boss told me to implement AI. I don't know what to do.
00:16:09
Speaker
um You could give it a very loose problem. And the first thing Clyde does is helps you break down that problem into a challenge statement, similar to how a consultant would. Like if you hired a McKinsey consultant, right? The first thing they're going to do is they're going to spend most of their time in the problem space and they're going to make sure it's really, really well defined.
00:16:30
Speaker
And then only once that's been defined, are they going to move on the building a hypothesis with you and then validating that into ah solution. And so Clyde does exactly that, but it doesn't do it with words that are going to completely alienate, you know, 99% of the user base, right? it It just simply asks you like, okay, So your sales is broken. um You know, what, what does that mean? What are your problems there? Like, how is that hurting you? but Let me, but let me ask this to follow on to that then. So why would you say in general, why does adding AI to the room solve a problem that perhaps like better bandwidth or better user experience or better software never touched?
00:17:14
Speaker
Like, why do you think AI would be the appropriate solution in these situations based on, based on what you built, of course? Yeah, I mean, i I think the reason AI is really appropriate and, yeah, in in this whole context is one...
00:17:29
Speaker
one thing I think both as individuals and as groups of people, we are pretty bad at is we let our, I'll just call it, uh, you know, human and social biases really get in the way to get to kind of common and shared understandings. Right. So think about how many times politics, egos, um,
00:17:55
Speaker
hierarchies, ah confirmation bias, like just gets in the way of truly trying to understand like what's, what's even the problem. Do, are we even aligned with the problem is? Right. Clyde's really focused on,
00:18:10
Speaker
getting to that shared and common understanding either within yourself or within a group of people. And I think that if you, if you give the right guardrails to AI in the right way, it can actually be a phenomenal thought partner in getting you there.
00:18:26
Speaker
what What's terrifying is if you don't give it the right guardrails and the right context, well, then it it creates the worst version of groupthink and confirmation bias and sycophancy and all the things that you see everywhere, right? So it's, I i think is,
00:18:46
Speaker
if you want to look at it even societally, right. I think it's the biggest opportunity and the biggest threat to like humanity a whole. And, and I think leadership and rain meter if you don't want to on this tangent, but I think, I think leadership has done a pretty terrible job across all aspects. And I've, have a front row seat to it because i talked to a lot of them um doing a pretty terrible job in,
00:19:13
Speaker
enabling people with both the right tools, but more importantly, the right philosophy training on sort of how to navigate all this so it doesn't become your worst nightmare. and And I think that's a real problem.
00:19:28
Speaker
Yeah. I, particularly for our younger listeners, um you may have heard the word sycophancy as it relates to LLM, but I want to assure you that it was a problem before it. I can't tell you how many rooms I've been in where I'm like, oh, we're not going to get to the real problem statement because no one wants to like offend the group leader because it was originally their idea like or or more to the point their edict. And then it went sideways.
00:19:55
Speaker
um So let's sit a little bit in this contradiction, right? On the on the website, ah you write, we've spent our careers helping teams work through hard problems and we learned something important. The tools weren't the problems, the process was. I'm a big believer in this statement. How do you reconcile that with essentially building a tool?
00:20:16
Speaker
You know, it's like the tool wasn't the problem, but here's a better tool to help process. I know it's sort of like becomes like a Mobius strip. um And ah so just help me with that reconciliation.
00:20:28
Speaker
Yeah. And I think the the way I kind of reconcile that is, you know, the the way we've always looked at sort of call it enterprise or processes or don't how the world works.
00:20:42
Speaker
I think people really like to split things into, you know, people, process, tools, right? And I think people are really comfortable with those little swim lanes. I've seen thousands, i was about to say millions, probably not, but thousands of projects where those are swim lanes and you're figuring out what do we do across tools? What do we do across process, people, leadership?
00:21:04
Speaker
I think all of that is massively converging to a point where the the process lives in the tool. The tool becomes the process But more importantly, people and leadership need to sit in the middle of it. So rather than them being swim lanes, I would almost re-architect it as, you know, maybe maybe even take like Simon Sinek's why, what, how famous little circle thing, and then put that in there instead, right?
00:21:35
Speaker
ah And just completely change how we even think about that. So short answer is I i don't think just... myopically thinking about tools and then process and how they interrelate is valid anymore. And honestly, take AI aside, like put that ah to the side.
00:21:53
Speaker
were Were they ever supposed to be separate, right? The fact that people treated them as these separate things. And then think about how that plays out in the economy, right? So tools equals SAP, Oracle, they become the biggest ah ERP systems in the world. yes Process becomes Accenture, the big four consulting trillion dollar industry, right? And people and leaders become the buyers.
00:22:17
Speaker
And then suddenly you've got this Very, very sort of siloed way of looking at it, which could probably account for why 70% of large initiatives fail because we're not treating them oh holistically.
00:22:34
Speaker
You bring those up too and like... In Canada, we had, I think it was EY and Deloitte both got nailed separately on two separate incidents for submitting million dollar process projects that turned out to be mostly AI slop that they charged top billing rates for that some junior analysts threw together before deadline.
00:22:55
Speaker
Church governments. these are These are sovereign governments paying for this work. Yeah. So i I very much empathize with the example you're talking about.
00:23:08
Speaker
um I think it is it is always a people process technology thing, especially in a digitized economy. I do think, again, it goes back to how the solution is implemented and what exactly you're trying to solve for, which I think is, again, exactly what Clyde is essentially trying to help lead customers to, is understanding. like From my understanding of it, mean it helps you understand better how to use AI to solve your problem specifically.
00:23:35
Speaker
Because people just, they think like AI is a magic black box that you just wrote at a thing and it produces results. And that's not- To be fair, they have been sold that bill of goods. They have been told that it will solve all their problems. No.
00:24:07
Speaker
Can you give us like one real specific example where an AI advisor changed the outcome versus, you know, what a human team would have landed on alone?
00:24:17
Speaker
Like, is there a tangible trackable, you know, one for one comparison where you can show the value on a decision making basis? Yeah, I mean, let's take ah let's let's take, I'll take a corporate example because um think it's most apt here.
00:24:33
Speaker
But you you had mentioned the AI slop and submitting that to governments and how that didn't end well. Well, there was also ah a case recently where a a New York judge had to throw out a case because both the prosecution and the defense had used Gen AI to create their arguments.
00:24:53
Speaker
and hadn't fact-checked them. And then the judge caught it and literally was like, both of you suck. Like like we can't oh god we can't progress anymore, right? um And so if if you think about a use case that's really hot right now with Clyde, it's pressure testing proposals and and points of view.
00:25:14
Speaker
Right. so let's say you've got a large consulting firm or boutique consulting firm and they have, you know, you know, an RFP for a big transformation job or strategy project.
00:25:27
Speaker
um Really popular use case for Clyde is step one to use Clyde and advisors to figure out, like, should we even bid on this?
00:25:38
Speaker
RFP, right? So bringing in a host of advisors, so you know ah a CFO, a risk manager, a strategist, um and really sort of work with them to just even define, like, should we...
00:25:55
Speaker
go after this. Is it profitable? Is it worth our time? You know, what's the benefit? Do we have the right win themes? Can we compete? You know, those types of things. um And then once you've decided to do that, back to my judge example, um we have an expert advisor in Clyde that's that's called the fact checker.
00:26:13
Speaker
And what that advisor does is joins your little workspace in Clyde. And as you're building out your proposal and you're building out your arguments and your assumptions, um this fact checker will will go in and it won't just do a random search and say, this is right, this is wrong.
00:26:32
Speaker
It's actually going to, together with you, interrogate all of your arguments and make sure that One, there's a a source for them if it's required and that that source is accurate and can be defended and wasn't generated by ah AI.
00:26:47
Speaker
um Two, it's going to actually help you sort of almost categorize your assumptions and arguments in that, okay, these are rock solid.
00:26:58
Speaker
These are iffy. You can still make them, but you should call out to your client that these are a little iffy they don't have the right sources. And then at the end, when you're done and you sort of get your output from Clyde,
00:27:10
Speaker
um we've seen a lot of examples now where people are even providing that output to the client almost as a show your work to to say, hey, this is how we got to this proposal. Like we we pressure tested this idea. Here are the breadcrumbs, like here are the breadcrumbs, here's the pressure testing, here are the advisors that we used, right? And so here you go kind of thing. But how's like how was that any different from what you did before? Like if I produce a report, especially a research a research-based report,
00:27:39
Speaker
whether professionally or academically, i have to cite my work, right? have to cite my research. Isn't just, ah isn't this just another form of citation of properly referencing something that you use to research what your outputs were?
00:27:53
Speaker
So yes, if you take that specific advisor and carve him out of Clyde and just say, you know, is this the same as me having to go just site cite my work? i mean, you could argue it's saving you, you know, some time because it's synthesizing across large data sets. But the power of it is really that that's just one part of that advisory team and that you also have,
00:28:17
Speaker
you're your sales advisor, your CFO, your risk manager, all looking at this together and that you're bringing in, say, three people from your sales team, your bid manager, um and these are human beings, not AI advisors, all into that same process as well and that it's all happening in one in one platform.
00:28:37
Speaker
um one One of the complaints I get from you know from users or things they used to complain about no longer do is that you know, they've been told, go use AI. um And then when they do it, they end up having to, you know, piecemeal themselves together, you know, three people from here, LLMs from there. Now I have to go use a legacy collaboration tool here.
00:29:01
Speaker
And then I have to spend all this time getting it into something that suits my workflow. Whereas Clyde does all that in sort of one platform. And so, If you try to take one specific feature out of Clyde and then say, there's a thousand things that do this for free, you know, you're largely going to be able to do that. It's it's the combined experience in the UX that's built for that workflow, I think, that that differentiates it.
00:29:29
Speaker
Nice. um One thing that you touched on earlier was related to like office politics and some other stuff which can hamstring processes. But also, to be fair, like sometimes consultants aren't primarily hired for analytical labor, but they're kind of purchased as political cover, like an external credibility for decisions that leadership want to make. Like, oh, look, McKinsey said we should do X, y and Z. And then it's, you know, they get to offload that responsibility. I might say abdicate, but they get to say to offload it.
00:30:06
Speaker
Does, do you think that Clyde is replacing that function or does it just get, faster at the, at the part of consulting that is aside from the credibility question.
00:30:18
Speaker
So it's interesting if you're looking for the rubber stamp, like no one gets fired because McKinsey said that was okay. Right. Like Clyde's not going to do that. Right. Clyde's not going to become the new rubber stamp for the board.
00:30:30
Speaker
But what Clyde will do is, you know, you'd mentioned the office politics stuff. Right. I'd mentioned that there's a bunch of research in group decision support systems and behavioral dynamics.
00:30:44
Speaker
One thing that Clyde does is Clyde will, Clyde has features that allows you to you know sort of almost take away the issue of the highest paid person in the room is going to make statement and everyone's going to have to agree.
00:31:00
Speaker
The introvert is going to get drowned out by the extroverts, right? The politics are going to make sure we don't get to the right outcome. You know, Clyde has features. A really basic one is when you're brainstorming ah ideas or problems inside of Clyde,
00:31:17
Speaker
um Clyde can enable an anonymous mode that means that, you know, say you've got 10 stakeholders, 20 stakeholders, all brainstorming ah what's the biggest risk with X or whatever it might be.
00:31:32
Speaker
They're all brainstorming these ideas with the boss there and this anonymity gives it sort of, it it separates the the idea from from the person to some extent, and people say what they really think and mean.
00:31:46
Speaker
I think that forward thinking companies that don't just want the rubber stamp actually want to be able to help people, that we've been through a process that takes away the highest paid person in the room syndrome, takes away the extrovert, drowning it, the introvert.
00:32:02
Speaker
And here's how, and here's the process we went to to get there. um And that creates something much better than a rubber stamp. That creates that this has been through the w ringer and we've gotten what the real people on the ground think and believe, which is going to create better business value because it'll actually get implemented.
00:32:22
Speaker
um So the short answer to your question is i I think we're actively fighting the rubber stamp syndrome with a better way. and And not everyone's ready for that. And they they probably won't be our customers, at least for the first batch.
00:32:37
Speaker
I, I, ah one, I appreciate you being willing to say like they might self select out because I think sometimes SAS founders are like, we're everything to everyone. Um, and then also the skeptic, I have some system prompts that I've put in my instances Claude, both work and personal that basically turn it into like a hard ass skeptic. Yeah.
00:32:59
Speaker
It will like rate its recommendations as to like certain likely and like unsure. And it will just push back on me, which was annoying at first because it's like, I just want you to do this one thing. um But then I began to appreciate that I had put in that setting because I didn't want it to just like. say like I'm brilliant at all things.
00:33:17
Speaker
i I actually over, I over corrected for that because I did the same, but, but mine just became so mean that I couldn't, like I couldn't handle it. So I had to dial it back because I was like, I don't think I'm strong enough to handle like this level of just. You, you, you created the anti-Claude, like the bully.
00:33:37
Speaker
ah George, are you? Yeah. So, I mean, there's also something flagged, which, which I find interesting, which is like the, the product builders trap, right? So, you know, on that point of like highest papers in the room, like going through and designing for experts and then you kind of shrink your own market, right? Because you don't make something consumable enough. You don't expand your TAM, if you will. Right.
00:34:01
Speaker
And so collaboration tools have historically had kind of the opposite failure where they're designed for mass adoption and then they end up very shallow. um And there's a balance that to be achieved there. So how do you avoid landing in the second trap, and the second trap which is obviously shallow trap, well without or while still dodging the first one?
00:34:22
Speaker
For folks listening, my apologies. It's like 9 in the morning. i speak well. i'm I'm a functional executive. I cut that sentence out. Tim and I understood me. But I do think it's a legit problem because I complain about all the time where...
00:34:35
Speaker
You know, when i evaluate tools, I'll be like, okay, well, this is too basic. This is just a dashboard tool just to make a board happy, but it's useless to a proper team, technical team. Or this tool is way too complex.
00:34:47
Speaker
you know, for example, George and I have some side projects going and something that we're building, there's no UI for a certain component. And it was a big hullabaloo for like two weeks of like, well, we can't demo SQL prompts to customers. There has to be some kind of flashy, nice...
00:35:04
Speaker
thing. Um, so you know, that it's annoying, but it's asked to be balanced. So how do you, how do you strike that balance and how do you avoid the product trap? Yeah, this is this is a constant, constant topic ah with myself and my co-founder and CPL. And we're we're so aware of this problem and we we literally design for it every week. And so this also gets into a much broader topic of investor dynamics and all all sorts of things happening in the AI industry. So...
00:35:38
Speaker
the way ai AI industry talks about that is horizontal versus vertical AI, right? They talk about if, you know, all investors are saying, if you're not vertical AI from day one, you're dead, right? And horizontal AI has no future.
00:35:55
Speaker
um The challenge is, Back to your sort of broad versus deep question. I think a lot of products try to go deep too quickly. um And then because they're told they have to in in the market and users say they have to, investors say they have to, but they do it without having enough data. So they arbitrarily choose that, okay, we're going to get really, really good at this specific use case and we're going to go super deep on UX and complexity.
00:36:25
Speaker
But they haven't really listened to their users, and they don't have a large enough sample size to know that that's where they need to sort of drill for oil or whatever whatever like what analogy you want to use.
00:36:37
Speaker
um So the way we approach that is we start very broad and shallow across sort of call it a number of of use cases and ICPs and stakeholders.
00:36:51
Speaker
And then what we've done is we've put a very sort of advanced analytics system on top of our user data user base and our user data. And we listen to that thing several times a week. And we look at where are people getting value? Where do they wish they were getting value?
00:37:09
Speaker
And then we start building out kind of the universe of technology. user and use case. And then we're dynamically deciding where we need to go deeper and where we need to verticalize.
00:37:21
Speaker
And the true answer to your questions, we don't, we don't know yet. We're too early in sort of our, our stage of company right now to say that, okay, for proposal creation, you know, we're going to go all in, right?
00:37:35
Speaker
Right now we're seeing that as a really strong use case, but At the same time, we're also seeing a really strong use case where people want to use this just to know, you know, how to use AI because my boss told me I had to.
00:37:48
Speaker
So AI road prioritization, AI roadmap prioritization is also a huge one. If you go down deep on any of those two, you're going to rabbit hole your product really, really quickly into being that.
00:38:01
Speaker
And so right now we're on top of that and we have not yet, I think, fully fully decided just because we don't know yet where we're going to verticalize.
00:38:13
Speaker
Which, by the way, is is a kiss of death if any investor's listening to this because they're going to tell me we don't have completeness of vision. But I think in this industry and where we're at, it's actually the right thing to do and it's the thing you're not allowed to say.
00:38:27
Speaker
I really appreciate that. I mean, when I do a lot of go to market advisory, it's really around, especially for early stages, like how much like essentially listening are you doing right with your design partners and your early customers? Like you should be sort of rabidly focused on them because they can help guide the product. But I think to your point, there's either rapid verticalization or just like hunt as many logos as possible to get the next round of funding. But it's like you haven't really decided what you want to be when you grow up, you know, and so then you get a lot like,
00:38:59
Speaker
Bolt on Klugey features, roadmaps get delayed, whatever, because you can't decide because you haven't been listening, right? You've only been meditating. VCs also throw money at dog shit. That's another problem, too. But let's just not go there. Correct. Like to hedge their bets. I want to close out here, Tim, taking a detour away from technology. Before we started recording, you and I were talking about a post that you'd put out when you said...
00:39:23
Speaker
um You know, I build frontier AI for a living and my kids get zero screen time. so I definitely want to sort of dig into a little bit of that contradiction there. We also know ah that you play in a dad rock band, which is kind of cool. So the guess my question is outside of the company, outside of this and and how you look to raise your kids, what are you thinking of or what is top of mind here?
00:39:51
Speaker
about the most human aspects of your life, whether, you know, I hate to use the word recharge because we're not iPhones, but like, is it a source of energy? Is it a source of inspiration? Is the lack of screen time a reflection on giving your kids kind of free reign to imagine? Talk to us a little bit more about that personal side of philosophy.
00:40:15
Speaker
Yeah, no, absolutely. And i I definitely feel strongly on this topic. So I think that, you know, start with the screen time. um i think one of the big challenges of this, I'll call it generation, has been, I'll call it the last 20 years.
00:40:32
Speaker
is that we unleashed all this technology on people, right? let's Starting with social media, that's the big one, right? And now AI sort of part of that discussion as well.
00:40:43
Speaker
And we had no idea what the effects were going to be. Well, we now know. There's been a lot of research around it. um Australia banned social media for all kids under 16. The UK just did it as well a couple of weeks ago.
00:40:56
Speaker
um because they know it's really, really hurting the development of children. And so for me, what's really important is before my kids start using ai and start benefiting from all this technology, because I think they have to, I mean, i and they will, I need them to really develop, you know, their critical thinking skills, their creative thinking skills, their social skills and dynamics, and And until they've built a really, really solid foundation in those areas, i think it's extremely dangerous to give them these tools they're just simply not ready for. um You know, I think a lot about how when I was a kid, i was bored a lot, right? Right.
00:41:40
Speaker
And kids today are just not bored. And so they don't develop critical and creative thinking skills. um I want my kids to have that, right? i Daydream for the love of God, people. dayd yeahream Daydream. Yeah, daydream. My daughter says every now and then else she looks just lost. And I'll be like, what are you doing? She's like, oh, i was just daydreaming. And then she'll spend like 15 minutes telling me what she was daydreaming about.
00:42:01
Speaker
And it's it's this incredibly creative, imaginative storytelling that's happening. And I think if she was glued to an iPad, I don't think that'd be happening. Right. i I have so many things to say and not enough time to say it, but yes. So we know neuroscientifically, psychologically, human creativity is combinatorial, right? It's going to pull from deep memory, short memory, things you've read. I was just reading the other day about how Nikola Tesla basically came up with the the theory of magnetic fields that led to the alternating current because he was reciting poetry by Goethe, you know,
00:42:39
Speaker
and Like, if you are in a world where you don't have that and essentially you've kind of outsourced cognition to these other things, ah I have increasingly come to believe, and I think a few social critics have said the same, Henry Oliver, David Brooks, a handful of others, I feel like...
00:43:01
Speaker
There's like this dangerous emerging underclass. There's kind of like a huge distracted portion of the population that's just like glued to screens and they'll make you. I was sat next to somebody on the plane and it was like just scrolling TikTok videos for five hours.
00:43:19
Speaker
on a transcontinental site. And i was like, what is that doing to your brain versus people who can hold conversations? Like I told you, like you're coming on here as a CEO founder, we're going to press up against you. Are you comfortable with that? I think you have done, you you probably have that comfort because you can hold contradictory ideas in your head. And I don't know, it's it's a huge soapbox, but i I love that you pointed out. And I, anyway. I think there's just way too much technology in school. Like, look, I was a daydreamer or two in school. My grade seven teacher or my grade eight teacher thought I wasn't going to pass high school because it's not that I didn't care. It's that I understood it too fast and I then didn't care. Right. And I think that creativity, like I know from my elementary school class or whatever, like
00:44:03
Speaker
You know, they're all still the same thing back home. And the kids that are kind of dreamers end up doing the big idea stuff. So i I really love where you're coming at this from, Tim. And to echo George's sentiment, thank you for just, you threw, like every curveball we threw at you, you nailed it. And I really appreciate that, particularly as someone who's a founder who understands their technology. um We need more practitioner founders. I really appreciate you what you're doing, Tim.
00:44:28
Speaker
No, I appreciate it. Thank you. Yeah. All right. Well, ah thanks again for the time, Tim. And I hope to follow up maybe in a year or so, see where Clyde is at. And um we'll we' look forward to that conversation.
00:44:42
Speaker
Great. All right. Well, thank you, George. Thank you, Georges. All right. Thank you. we'll talk to you soon. Cheers.
00:44:54
Speaker
All right, crew, questions to leave you with and think about. i am still thinking about human process. I think it's where we give the least amount of attention. We just want to slap techno solutions onto everything.
00:45:09
Speaker
And i guess ah my question to you is to think about your process, whether that is at work, or how you come up with your best ideas under what conditions you come up with your best ideas how you cultivate creativity and just give a lot how are you giving time and attention to thinking about that process whether it needs to change whether it needs to iterate um because that's where the juice is it's actually not in the uis and the dashboards and the widgets yeah i think i think the biggest thing ah is looking at you know
00:45:42
Speaker
What is value? Like, what exactly are you trying to create business value out of? And is AI the right answer? Because at the end of the day, i think where people are failing to kind of differentiate where AI's value is, you know, there's process automation, there's collation of a ton of data sources, and then there's creativity and original thinking.
00:46:03
Speaker
And people should really begin to ask themselves, is this is this a problem that requires original thinking? Or is this a problem that requires taking a bunch of data I already have and coming up with a collated answer?
00:46:15
Speaker
And I think that that question alone will help you bring more success to your use of AI. Nice. All right. We'll leave you with those and we will see you next week.
00:46:29
Speaker
If you like this conversation, share it with friends and subscribe wherever you get your podcasts for a weekly ballistic payload of snark, insights, and laughs. New episodes of Bare Knuckles and Brass Tacks drop every Monday.
00:46:42
Speaker
If you're already subscribed, thank you for your support and your swagger. Please consider leaving a rating or a review. it helps others find the show. We'll catch you next week, but until then, stay real.
00:46:57
Speaker
No, like the Air France food, right? Like I got this meal and and I was just, I mean, it was just good old coach, right? and And I got this meal and it was, i think it was like one of the better meals I had that week.
00:47:11
Speaker
It's just, yes. I do the business class. your but the They got the Michelin star meal on a flight with the lay down bed pod thing. Stop bragging. revolutionized every time i'm sorry. I'm sorry. We should do the show. but once once Once you taste the champagne, you can't go back to the tap water. that was it Actually, that that was impressive. That wasn't even humble bragging. That that was just bragging.
00:47:33
Speaker
It was like good is awesome.