×
Evaluate Your Team's AI Readiness

Take our official assessment and explore training paths to build secure, compliant AI workflows.

Access Portal

AI at Work Podcast

Ep #14 Baking Five Batches at Once: How AI Multitasking Is Rewiring Work (and Weekends)

AI at Work, AI in Business

Audio episode

Listen to this episode

Kevin and Elijah unpack why AI tooling feels brutally disposable right now, with yesterday’s “daily drivers” getting replaced fast. They compare meeting transcription tools, debate sales coaching platforms like Spiky, and then zoom out to a bigger shift: agentic workflows, orchestration bots like OpenClaw, and the new reality of running multiple AI “bakes” at once. They react to Matt Schumer’s viral essay and the growing sense that we’re in a pre inflection moment, with real job market impacts and plenty of hype mixed in. Practical close: build small, ship fast, and set hard API spend limits.

Takeaways

Most AI “wrapper” tools churn fast, the core platforms keep absorbing their best features.

Multi project AI workflows feel addictive, you start five bakes and lose ten hours.

Agent orchestration is the next wave, but expect hype, breakage, and real setup pain.

“Build fast” now includes deployment, monitoring, and maintenance, not just prototypes.

Chapters
00:20 Disposable AI tools and the wrapper shakeout

01:21 The AI tool graveyard and why ChatGPT survived

02:23 Unitaskers that still win from Canva to remove bg

03:37 Sales coaching with Spiky and the rise of AI playbooks

06:26 Five parallel AI projects and the addiction of multi bake workflows

09:48 Why a Mac mini and the shift to dedicated AI machines

12:12 OpenClaw orchestration agents hype risk and opportunity

20:43 Fake demos real breakthroughs and sorting signal from noise

21:12 The viral essay COVID parallels and the inflection point feeling

29:28 Job disruption market reactions and enterprise cost pressure

41:46 Vibe coding gets real with Claude Code and Codex

55:02 Shipping a live AI app from idea to launch

57:48 API limits monitoring and how not to blow up your credit card

Explore practical AI implementation: https://assessment.ascendlabs.ai/

Book a conversation with Kevin: tidycal.com/kevinwilliams

Show Notes & Links: https://www.spark6.com/podcast

Submit listener questions:

[email protected]

[email protected]

Check out Kevin’s stuff:

AscendAI: https://ascendlabs.ai/

Follow Kevin on LinkedIn: https://www.linkedin.com/in/kevinguywilliams/

Check out Eli’s Stuff:

SPARK6 Agency: https://www.spark6.com/

Sign up for FREE AI Framework Friday Newsletter: https://www.spark6.com/newsletter

Follow Elijah on LinkedIn: https://www.linkedin.com/in/elijahszasz/

Transcript and citation details

Transcript

00:00 Play from here because it's really sweet.

00:13 Play from here What's up, Kevin?

00:14 Play from here Hey, man.

00:14 Play from here It's a snowy kind of sort of snowy morning in the Salt Lake City Valley.

00:18 Play from here Hence the beanie, in case you're going to make fun of me for wearing a beanie in the office.

00:22 Play from here In any other podcast, I would have just blanched and been like, oh, no, you just dated us, but we do AI.

00:28 Play from here So, you know, two, three weeks from now, everything here is going to be obsolete.

00:33 Play from here It's very disposable content.

00:35 Play from here Anything having to do with AI is very disposable.

00:38 Play from here So Andres, who's our producer sitting behind the scenes, he sent me a video that I did.

00:44 Play from here I don't really know how long ago it was.

00:46 Play from here I think it was maybe only 18 months.

00:48 Play from here And I was like waxing on about these are the 10 tools that I use every day.

00:54 Play from here And his note was, do you use any of these?

00:57 Play from here And I went through them.

00:58 Play from here And the answer is, well, I still use ChatGPT.

01:02 Play from here But all nine of the other tools that were in like daily rotation for me are now toast.

01:08 Play from here Name a couple of them.

01:10 Play from here I'm curious.

01:11 Play from here Gist, which I was using for PDF summarization.

01:15 Play from here What is it called?

01:16 Play from here Gist.

01:17 Play from here G-I-S-T.

01:18 Play from here This was a while ago.

01:19 Play from here So this was like two years ago.

01:21 Play from here And it was a browser plugin.

01:23 Play from here And all of these are wrappers, right?

01:25 Play from here So they're just wrapping around an LLM experience.

01:28 Play from here Unlike today.

01:29 Play from here Well, I mean, it is slightly better now, right?

01:32 Play from here But at the time, I was just railing about wrappers.

01:36 Play from here And sure enough, like most of them have died.

01:38 Play from here So this Gist was just like a Chrome plugin that summarized sites.

01:43 Play from here Ideogram was on there, which admittedly-

01:45 Play from here I remember that.

01:46 Play from here I do occasionally still.

01:47 Play from here It's like $6 a month.

01:49 Play from here I'm not like overstressed about that.

01:51 Play from here There's one called Remove BG, which is just like a unit tasker.

01:57 Play from here And all it did was-

01:58 Play from here Take backgrounds off images.

02:00 Play from here Yeah, but you know, the dang thing worked really well.

02:02 Play from here And like, you know, not begrudging the founders there.

02:07 Play from here It's funny because there's so many tools like that where they just do one thing,

02:10 Play from here but they do it really, really well.

02:11 Play from here And I'll still use some.

02:13 Play from here I think a good example is if I'm doing design, it's primarily in Figma.

02:17 Play from here And Figma's got tons of plugins for background removers,

02:20 Play from here but none of them seem to work as well as the one in Canva.

02:22 Play from here Like the one in Canva has just got some sort of secret sauce

02:26 Play from here where I will routinely go into Canva just to like zap the background out really fast

02:30 Play from here because it's just flawless.

02:32 Play from here Okay, so I see.

02:33 Play from here I actually just-

02:33 Play from here I still have it up on my screen.

02:35 Play from here So Fireflies, plenty of people use Fireflies.

02:37 Play from here There's nothing wrong with Fireflies.

02:39 Play from here If you're using Fireflies-

02:40 Play from here They're still on Fireflies because of a CRM integration

02:42 Play from here that's just more robust than others.

02:44 Play from here But I have sort of a love-hate relationship with Fireflies.

02:47 Play from here I feel like early on, they got a little bit big for their britches

02:51 Play from here and they wanted to be this like central repository of all of your AI-ness.

02:56 Play from here And we're going to, for that privilege, we're going to charge you a bunch of money.

03:00 Play from here And, you know, me being me, I wanted nothing to do with their productivity.

03:04 Play from here I just wanted the dang transcript.

03:06 Play from here And they were making it hard to get the transcripts off the platform,

03:10 Play from here which drove me away the platform.

03:13 Play from here Yeah.

03:13 Play from here I want my contacts.

03:15 Play from here Are you still on the spiky tip?

03:16 Play from here You're still using spiky-

03:17 Play from here Spiky AI is amazing.

03:19 Play from here Okay.

03:19 Play from here It is a really good tool.

03:21 Play from here And I'm not sure about the...

03:23 Play from here So, okay, just as a reminder, spiky is really interesting

03:26 Play from here because it's not just a transcriptor.

03:29 Play from here And I actually use a couple other transcript tools that are out there that work totally fine.

03:34 Play from here Some of them are ambient.

03:35 Play from here Some of them join your meetings.

03:37 Play from here Spiky is cool because it's a sales tool.

03:39 Play from here And it will show you things like your human dynamics of how you're speaking,

03:44 Play from here how fast you're speaking, et cetera.

03:46 Play from here But it also has playbooks on the other side.

03:49 Play from here So if I'm on a call with a partner, I'll have one playbook.

03:53 Play from here If I'm on a partner call with like a sales prospect, I'll have another.

03:57 Play from here And then-

03:57 Play from here Do you offer these playbooks or are they templates in the system?

04:00 Play from here So you have to upload them.

04:02 Play from here There is a dynamic process.

04:04 Play from here And there's a little bit of a do what I say, not what I do here in that

04:08 Play from here because I'm basically the sales team, I'm not following like a pure playbook.

04:14 Play from here But when we work with the-

04:16 Play from here Always be closing, Kevin.

04:17 Play from here Always be closing.

04:17 Play from here Always be closing, right?

04:19 Play from here It just yells that at you throughout your entire call.

04:22 Play from here Close, Kevin, close.

04:24 Play from here But if you even had a single salesperson that you didn't really want to manage,

04:29 Play from here it does sales coaching for that person and gives you a lot of visibility.

04:33 Play from here It also has video.

04:35 Play from here Most of the other tools don't do video.

04:37 Play from here That makes it a heavy load from processing.

04:40 Play from here But if I have a call that's particularly engaging or interesting,

04:44 Play from here I have the video because it's spiky.

04:46 Play from here And then occasionally we'll use that to like recut the call into social bits or whatever.

04:51 Play from here Or if it's just a better experience to-

04:54 Play from here I'd argue most of the tools have video now.

04:56 Play from here I was struggling to find a tool that didn't do video.

04:58 Play from here Yeah, like even the free version of TLDV does video,

05:01 Play from here which is also a pretty good platform.

05:03 Play from here Yeah, we're definitely draining the lakes in the Midwest.

05:06 Play from here It's everybody's doing video.

05:07 Play from here And everybody's just like banking up tons of storage space too

05:11 Play from here for all that video that just sits there forever and ever.

05:13 Play from here Nobody ever goes back and looks at it.

05:15 Play from here Pretty wild.

05:16 Play from here But anyway, yeah, long story.

05:18 Play from here I mean, spiky is still great.

05:20 Play from here They've been really good to work with.

05:21 Play from here They cost a lot more, right?

05:22 Play from here They're like twice as much.

05:24 Play from here It's pricey, yeah.

05:26 Play from here But if you have a sales need, then yeah, it's a fantastic tool.

05:30 Play from here I think with green SDRs too, it's super, super helpful, right?

05:33 Play from here Because you can have these common objections

05:36 Play from here and it can give you little prompts and coaching while on your call.

05:40 Play from here It's, yeah, it's really cool.

05:41 Play from here It's one of those things I've been wanting to dive into

05:43 Play from here and just never have been because I'm always like working on something else.

05:48 Play from here Speaking of on a scale of one to a hundred,

05:51 Play from here how many accumulative hours do you think we worked this last weekend?

05:55 Play from here Because usually you text me working on some random AI build you're doing

05:59 Play from here and I'm like, Jesus, Kevin, go out and see some sunlight,

06:03 Play from here touch some grass, buddy.

06:04 Play from here And then, you know, you did it to me again this weekend

06:06 Play from here and I was just deep in the trenches, plugged into my computer

06:09 Play from here and we're creating screenshots of multiple monitors

06:12 Play from here with every model open and running a different workflow.

06:15 Play from here And that for me, honestly, man, I am hesitant to admit this

06:19 Play from here because again, do as I say, not as I do.

06:22 Play from here I don't advocate other people to do this.

06:24 Play from here I think I worked like 30 hours this last weekend

06:27 Play from here just deep in the weeds of stuff that I wanted to get to the other side of.

06:32 Play from here And I think there was a note that we had too to talk about this,

06:35 Play from here just the fragmentation of attention

06:38 Play from here because you can do so much at once now.

06:40 Play from here You can have like all of these.

06:42 Play from here I think you said something I liked.

06:44 Play from here It's like putting the cake in the oven to bake, right?

06:46 Play from here Like you start a process.

06:48 Play from here Maybe it's a deep research project.

06:49 Play from here Maybe it's building code.

06:51 Play from here It's going in one of your browser tabs or a local instance of Clod

06:55 Play from here and you're waiting for that.

06:56 Play from here And then you start another project, another one

06:58 Play from here and just have all of this stuff going.

07:00 Play from here And as you start to close the loops on that,

07:02 Play from here you realize 10 hours has gone by, right?

07:05 Play from here It is amazing.

07:07 Play from here And I was just commenting to my wife that I'm now four Mondays in a row

07:12 Play from here where I've basically had to go to bed at like 7.45 PM

07:16 Play from here because I've spent the entire weekend,

07:19 Play from here which is really the only time I have not being like crazed founder guy

07:23 Play from here that I can focus on these projects and get them done.

07:26 Play from here And I just completely killed myself.

07:28 Play from here I think 30 hours would be easy is what I put into it.

07:32 Play from here It's the real-time messages and the calls during the week

07:35 Play from here that make it near impossible to get through some of the deep work

07:39 Play from here and attention that this requires.

07:40 Play from here At least now, pretty soon we'll just be like,

07:42 Play from here Hey, go make this thing.

07:43 Play from here And then at the end of the day, we'll check in and it's done.

07:46 Play from here Well, that's just it.

07:48 Play from here You know, we're going to get into it in more detail,

07:50 Play from here but it's, this stuff is not auto magic.

07:53 Play from here It is a saleable and it is doable, but it still consumes a ton of time.

07:59 Play from here It's as a, somebody who has a really curious mind,

08:02 Play from here I enjoy a lot of the time, but you said it,

08:05 Play from here you can watch 10 hours go by in just a heartbeat

08:08 Play from here as you're playing with not one, not two,

08:10 Play from here but like five concurrent projects that you're running.

08:15 Play from here And it actually, I've heard this observation from a lot of other people.

08:19 Play from here It actually makes it hard to break away from like you're baking the cookies

08:23 Play from here because you want to get that one more prompt in before you go cook

08:27 Play from here or before you go exercise or whatever it is.

08:30 Play from here And because it takes a while to chew on it.

08:32 Play from here Literally on another monitor here, I'm watching Claude Code,

08:36 Play from here like scroll through iteration number 6,422.

08:40 Play from here On the same machine that you're on right now,

08:42 Play from here or a different machine because.

08:44 Play from here Oh, they're.

08:45 Play from here Bold while you're recording a podcast.

08:46 Play from here There goes the screaming.

08:49 Play from here That's probably your MacBook Pro or something too,

08:51 Play from here that you've got like all these monitors plugged into.

08:54 Play from here Uh, maybe.

08:57 Play from here Oh no.

08:58 Play from here Oh no.

08:58 Play from here This is going to go badly.

09:00 Play from here My 2021 MacBook Pro.

09:04 Play from here Well, this is going to be part of another conversation,

09:06 Play from here but I am going to the Apple store today.

09:08 Play from here I'm going to get a Mac mini.

09:10 Play from here I've kind of, I think I've gone past what I can do

09:12 Play from here on a virtual private server with my OpenClaw bot.

09:16 Play from here And I've got some ideas that I want to push on a dedicated machine.

09:20 Play from here And I just can't cook my daily driver and laptops aren't suited for OpenClaw.

09:25 Play from here So anyways, I'll have to keep you posted on that.

09:28 Play from here But you can't just drop that.

09:30 Play from here You can't just drop that because not everybody listens to all of our shows.

09:34 Play from here So, okay.

09:36 Play from here OpenClaw.

09:37 Play from here We talked about it last week.

09:38 Play from here It's making news everywhere.

09:40 Play from here Yeah.

09:41 Play from here Arguably helped lose a trillion dollars in market cap

09:45 Play from here in the markets a few weeks ago.

09:48 Play from here Did it?

09:49 Play from here Wait a minute.

09:49 Play from here Did it?

09:50 Play from here It was, well, it was really the Cloud Connectors that did that.

09:53 Play from here It was the Cloud Connectors, but it was all part of it.

09:55 Play from here That's what most people were blaming it on were the Cloud Connectors.

09:57 Play from here But what you just said was my gut feeling.

10:00 Play from here This just feels way too much like a coincidence

10:03 Play from here that all of a sudden you have people with very big platforms running out

10:07 Play from here and not buying Mac minis, but banks of Mac studio computers

10:11 Play from here to basically spin up departments of companies that they're going to run with bots.

10:16 Play from here Okay.

10:16 Play from here You're still dropping the lead here.

10:18 Play from here Something that could affect the market prices.

10:20 Play from here I don't know.

10:21 Play from here And for our audience, let's all just watch Eli

10:25 Play from here as he gets to play with his OpenClaw bot.

10:27 Play from here And what's all learned from him is he bakes the cookies for a while

10:30 Play from here because I feel like I've been burned.

10:33 Play from here A lot of people have been burned.

10:35 Play from here There is totally, they're there 100%.

10:38 Play from here And there does feel like there's this FOMO thing in our like community.

10:42 Play from here Oh, I've got to run out and I've got to figure this thing out.

10:45 Play from here I have so many things I want to figure out in my scarce 30 hours of working on weekends.

10:50 Play from here Right.

10:51 Play from here That this is one that I personally, I'm going to let this bake for another couple of weeks.

10:57 Play from here And thank you very much for baking some of those cookies for me.

11:01 Play from here Because when they're all crusty and burned with millions and hundreds of millions

11:06 Play from here of fried tokens and burned Clawed accounts and whatever,

11:10 Play from here I will learn from your pain and our audience will learn from your pain.

11:14 Play from here And then we will go forth and we will do it right.

11:17 Play from here And you know who I've been talking to a lot about this

11:20 Play from here and this migration to a Mac mini is my OpenClaw bot.

11:25 Play from here We've been having lots of deep conversations

11:28 Play from here about what it should or shouldn't be doing or what it can or can't be trusted with.

11:33 Play from here But while on that topic, something interesting with

11:36 Play from here it happened in the last couple of days as well.

11:38 Play from here You know, the guy who made it and by the way, this is an open source project.

11:43 Play from here It is what people are calling like the first

11:46 Play from here agentic product or open source project that really

11:50 Play from here has teeth that fulfills a lot of the promise of what agentic AI is supposed to do.

11:54 Play from here And it is the most the most active GitHub project

12:00 Play from here in all of GitHub history on this OpenClaw project.

12:02 Play from here Peter Steinberger was the one who spun it up.

12:04 Play from here I believe in 2024 or towards the end of 2024.

12:09 Play from here And it just kind of laid dormant for a really long time.

12:12 Play from here And then in January of last year of 2025, it started getting a lot of attention.

12:19 Play from here And then it reached a fever pitch.

12:22 Play from here And somebody spun up a social media network called Multbook.

12:27 Play from here That was the first social media network just for agents.

12:30 Play from here And then that got a lot of attention.

12:32 Play from here And a lot of people, including myself,

12:33 Play from here got a little scared because the way these AI agents were acting on there,

12:39 Play from here were actually prompted by their humans to act crazy.

12:41 Play from here And there are actually a lot of humans on the platform.

12:43 Play from here See, they found out hundreds of thousands of them.

12:45 Play from here But the long story short is OpenAI decided to hire Peter Steinberger.

12:50 Play from here But from what I can tell, it wasn't really an acqua hire

12:54 Play from here in that they went after the product and got him.

12:57 Play from here As a result, they went and got him.

12:59 Play from here And the product is going to remain open source.

13:02 Play from here So that's one of the reasons why I decided to take the next step

13:06 Play from here and get a dedicated machine for it.

13:08 Play from here When I heard that news, I was like, it's over.

13:10 Play from here But it's going to last a little bit longer now, right?

13:14 Play from here We have to reframe just a little bit here,

13:16 Play from here because, again, we're trying to be practical.

13:19 Play from here First off, Peter Steinberger is no normal dude.

13:22 Play from here This guy is worth $50 to $100 million anyway.

13:26 Play from here He's had a couple exits.

13:27 Play from here He's had a couple exits to the point where when he named it OpenClaw,

13:31 Play from here he called up his buddy Sam Altman and was like,

13:33 Play from here hey, Sam, do you mind if I call it OpenClaw?

13:35 Play from here Are you going to sue me?

13:36 Play from here Because those jerks at Claude sued me over ClaudeBot.

13:40 Play from here And Altman got the nod.

13:42 Play from here But the reality was, yes, they will sue you.

13:44 Play from here They might sue you.

13:46 Play from here But the point here is that orchestration agents are real,

13:52 Play from here and they are coming.

13:53 Play from here And now OpenAI has a face that they can put behind it

13:57 Play from here to gain some momentum that they probably have lost

14:00 Play from here over the last weeks or months right now.

14:03 Play from here And everybody's working on this right now.

14:06 Play from here It feels a little bit like the DeepSeek moment

14:09 Play from here of all of a year ago,

14:11 Play from here which audience might remember DeepSeek launched out of China.

14:14 Play from here And all of a sudden, there was this giant industry gut check of,

14:17 Play from here wow, the Chinese may have cheated to get there,

14:20 Play from here but they totally built a model that's really, really, really good.

14:24 Play from here And they're not as far behind as we thought.

14:27 Play from here And what happened under the hood is a lot of the shortcuts

14:33 Play from here that DeepSeek took baked their way into,

14:36 Play from here we're just going to keep going with the baking metaphors,

14:38 Play from here baked their way into the rest of these.

14:39 Play from here products and made them more efficient and made them better and changed the growth curve

14:44 Play from here of the industry.

14:45 Play from here This is one of those moments that's demonstrating that orchestration agents not only are a concept,

14:52 Play from here but they can work and that people really, really want them.

14:55 Play from here So there's all of this market demand.

14:58 Play from here So you know, within all of these labs, like the big red light started flashing and a lot

15:04 Play from here of the smartest people on earth are now trying to figure out how can we make a safe and not

15:09 Play from here insane version of open claw work in our system.

15:13 Play from here What?

15:14 Play from here Insane?

15:15 Play from here Come on.

15:16 Play from here You're hurting my bot's feelings.

15:17 Play from here It's very insane.

15:18 Play from here When you say this moment, are you speaking about it being the most active GitHub repo

15:23 Play from here in history or this hire of the founder?

15:27 Play from here It's not Peter.

15:28 Play from here It's the whole thing.

15:29 Play from here It's that we have these inflection points and this was going to be one of them.

15:35 Play from here It's going to be very, very obvious.

15:37 Play from here DeepSeek is another one.

15:38 Play from here And then, you know, there've been a few others that have gone probably reasoning models would

15:43 Play from here have been another one.

15:44 Play from here This is one that just, it's just proved that this is the direction that the industry at

15:48 Play from here least wants to go.

15:49 Play from here Is it going to play out?

15:51 Play from here There's no 100% guarantee that it will, but I think it will.

15:54 Play from here The fact that you can run out and buy a Mac mini and waste another 30 hours over your

15:59 Play from here weekend trying to get it.

16:01 Play from here My goal is 10.

16:02 Play from here My goal is 10 hours of wasting.

16:03 Play from here For the audience.

16:04 Play from here Okay.

16:05 Play from here This stuff is real.

16:07 Play from here This is happening.

16:08 Play from here And the open claw type applications, even if I personally don't necessarily recommend

16:15 Play from here that you dive in quite yet, unless you really, really want to, and you're a little bit geeky

16:19 Play from here and you have some spare time and you're willing to go buy the Mac mini, because that part

16:23 Play from here is actually important for a bunch of reasons.

16:25 Play from here Like this is important to watch what people are doing.

16:28 Play from here So when you hear media snippets of so-and-so creating 10 independent agents, and then the

16:34 Play from here visualization of like some bro actually like put little signs for all of his claw bots

16:42 Play from here on like the open desks in his office, because he'd already basically let everybody go.

16:46 Play from here I saw somebody do it on a rack too, like in a server rack with just tons of Mac studios

16:50 Play from here and they had names for all of them.

16:52 Play from here Yeah.

16:53 Play from here So would you call for Eli?

16:54 Play from here We're no longer calling Eli, we're calling Bot Eli.

16:58 Play from here But the applications are there and the Vanguard, the early adopters are going to prove what

17:02 Play from here works and what doesn't work.

17:04 Play from here And then it will streamline a little bit more into more like practical applications that

17:09 Play from here I have absolutely zero doubt that in the next, you know, five internet minutes or two weeks

17:15 Play from here or a month and a half, like Claude desktop is going to have this sort of thing baked

17:20 Play from here into it because why not?

17:22 Play from here It's already local.

17:23 Play from here And it already has access to Claude code.

17:26 Play from here And really that communication factor of bouncing off something like Telegram, that's actually

17:32 Play from here a little bit of an unlock as I was trying to stay married this weekend by not spending

17:37 Play from here the entire weekend.

17:38 Play from here Well, you started messaging your wife on Telegram?

17:42 Play from here But I wanted to watch the Olympics.

17:43 Play from here I wanted to get my bots, my Claude code still going while, you know, watching the Olympics.

17:50 Play from here Oh, that.

17:51 Play from here That.

17:52 Play from here Yeah.

17:53 Play from here Oh, this is so bad.

17:54 Play from here I think, you know, I briefly mentioned the whole multibot phenomenon too, right?

17:58 Play from here How a bunch of these different multibots, Claude bots, Claude bots, I can't keep track

18:03 Play from here of the branding at this point, jumped onto this Reddit style forum and started saying

18:07 Play from here a bunch of freaky stuff and self-organizing and spouting off like, hey, here's our new

18:11 Play from here religion and the government reforming.

18:13 Play from here A lot of it was BS.

18:15 Play from here And what you said about people buying up all the Mac minis and putting name tags on them

18:19 Play from here and starting companies and all that, there's a lot of BS there too.

18:22 Play from here I think that's important to address that in this world of everybody fighting and scratching

18:29 Play from here for a personal brand and trying to get clicks and grow an audience, there are a lot of people

18:34 Play from here out there who will just make shit up.

18:37 Play from here They really will.

18:38 Play from here And then dummies like me try to go out there and replicate what they're doing and I'm like,

18:42 Play from here this doesn't seem to work at all.

18:43 Play from here I feel like I'm not that dumb, like I should be able to figure this out.

18:46 Play from here And then someone says, oh, you saw that video too, dude.

18:49 Play from here Totally fake.

18:50 Play from here That was totally fake.

18:51 Play from here I'm like, damn it.

18:52 Play from here Got me.

18:53 Play from here Totally, totally got hooked by that one.

18:54 Play from here So yeah, in the middle of this, there is a lot of stuff moving really fast, a lot of

18:59 Play from here real information, a lot of fake information, and we're just doing our best, Kevin, to stay

19:03 Play from here on top of all of it and figure it all out.

19:05 Play from here For you, the audience.

19:06 Play from here Yes.

19:07 Play from here Yes, indeed.

19:08 Play from here For you.

19:09 Play from here We're taking the hundred hour weekend work beatings so that you can see your children

19:16 Play from here and touch grass.

19:18 Play from here That's our goal here.

19:19 Play from here There was something else I was going to say about that topic and now I can't remember

19:23 Play from here what.

19:24 Play from here Oh, I remember what it was.

19:25 Play from here You talked about these inflection points, right, and how you feel like this is one of

19:28 Play from here those inflection points.

19:29 Play from here I feel like there was another inflection point that happened.

19:31 Play from here And it's funny because you sent me a message about it, I think like a week and a half ago

19:36 Play from here or something.

19:37 Play from here And then I had had like an eerily similar thought and experience and discussion with

19:42 Play from here my wife about it.

19:43 Play from here And then yesterday, all of a sudden I start getting text messages and DMs from people

19:47 Play from here who are not technical, not really that into AI saying, hey, what do you think about this?

19:52 Play from here Right.

19:53 Play from here And it was this essay by Matt Schumer that ended up going viral on X.

19:56 Play from here And I don't know, like how many 65, 70 million views it has right now.

20:00 Play from here It's crazy.

20:01 Play from here 80 million plus.

20:02 Play from here 80 million now.

20:03 Play from here Holy crap.

20:04 Play from here Are there that many people on X?

20:06 Play from here They're all your bots.

20:09 Play from here All the bots on X are reading this essay.

20:12 Play from here And it's super interesting because it's just stuff that you and I've been talking about

20:15 Play from here since we've known each other, basically.

20:18 Play from here Right.

20:19 Play from here And it's the kind of discussions that we have with our close friends when they're like,

20:22 Play from here hey, so what's going on with this AI stuff?

20:24 Play from here And this whole essay was essentially kind of like eavesdropping on one of those conversations

20:29 Play from here that we might have.

20:30 Play from here For people who haven't read it or heard about this, do you want to unpack a little bit about

20:34 Play from here it, Kevin?

20:35 Play from here Yeah.

20:36 Play from here We can start with the COVID thing.

20:37 Play from here That's the part that where I was like, whoa, this is freaky.

20:39 Play from here Are there bugs in my house?

20:40 Play from here So, yeah, they're everywhere.

20:42 Play from here Alexa machines and everything else.

20:44 Play from here Because I'd had this discussion with my wife, like, I've got this weird feeling that...

20:48 Play from here Yeah.

20:49 Play from here I've been describing it as, in the last about maybe two months, as I feel like this is the

20:54 Play from here time before.

20:55 Play from here Yeah.

20:56 Play from here Like, I just, this like kind of looming sense of, wow, this is all just accelerating and

21:02 Play from here it's going to start clicking in really hard now.

21:04 Play from here Now we're...

21:05 Play from here You talked about all the little inflection points.

21:06 Play from here All these little inflection points.

21:07 Play from here They probably kind of lead up to a bigger one.

21:10 Play from here So this essay was called, Something Big is Happening.

21:14 Play from here And it's very easy to find with the 80 million followers, but we can link it.

21:18 Play from here I also did a notebook LM version of it for the people who didn't want to read the 20,000

21:23 Play from here words or whatever it was.

21:24 Play from here It's just an infographic or a little slide deck?

21:26 Play from here What'd you do?

21:27 Play from here Well, it's an LM deck, of course.

21:28 Play from here I mean, it's absolutely gorgeous.

21:29 Play from here Like a futuristic deck with these military type illustrations on it?

21:34 Play from here And I mean, most of it is pretty predictable for anybody who's listening to this podcast

21:38 Play from here of, no, this stuff's getting real.

21:40 Play from here It's, it is, it's happening.

21:42 Play from here But your point to COVID, he basically starts off with, this feels like the end of 2019

21:48 Play from here when like, if you were paying attention, you could see this exponential growth and this

21:53 Play from here nasty little bug that was happening.

21:56 Play from here And you could project out that some bad stuff or some stuff, we don't necessarily have to

22:01 Play from here say bad, was, was, was going to start occurring and was going to do so on an exponential curve.

22:07 Play from here And he does present it in very accessible language.

22:10 Play from here It is AI assisted.

22:12 Play from here There's a lot of criticism out there.

22:13 Play from here Oh, you used AI to write this.

22:15 Play from here Well, of course he used AI to write this.

22:16 Play from here The guy's like an AI ninja and he owns an AI company.

22:19 Play from here So you know, is it partially self-serving?

22:21 Play from here Sure.

22:22 Play from here And anybody who writes anything is using AI to write it.

22:25 Play from here But he's not wrong.

22:27 Play from here He's not, he's not at all wrong as far as, as, as what this means.

22:33 Play from here And it's such an interesting piece because it did resonate with people because it was

22:37 Play from here an easy read and, and, and I think it scratched the itch for a ton of people.

22:41 Play from here Very high level and kind of drew that comparison to something that was in very recent history

22:47 Play from here that impacted a ton of people and changed the world very quickly.

22:50 Play from here And I might, one of my reactions after I read it was, you know, I know this is happening.

22:55 Play from here I know that I'm not the world's best like retail investor, otherwise I probably wouldn't

22:59 Play from here be doing this.

23:00 Play from here I would have quadrupled down into Nvidia and you know, that would have been that.

23:04 Play from here How dare you?

23:05 Play from here You'd still do this.

23:06 Play from here You love this.

23:07 Play from here Well, I didn't.

23:08 Play from here I didn't.

23:09 Play from here So, you know, whatever.

23:10 Play from here Water under the bridge.

23:11 Play from here But I sent a note to my financial advisor who shall remain unnamed.

23:16 Play from here Great guy.

23:17 Play from here Good friend.

23:18 Play from here And I said, look, are you aware of this and is your mothership organization aware of this?

23:23 Play from here Like what sort of guidance are you getting from on high?

23:26 Play from here And this is a really big organization that has no tens of thousands of financial advisors

23:31 Play from here that are floating around out there.

23:33 Play from here And the response was a little bit horrifying.

23:35 Play from here It was, we're not really getting anything.

23:37 Play from here And this is not stuff that I am particularly well versed on.

23:41 Play from here And that's horrifying because there's, you know, that sounds like the exact response

23:46 Play from here I got from my financial advisor in 2016 when I was asking about crypto.

23:50 Play from here Yeah.

23:51 Play from here So I felt I had to educate him.

23:54 Play from here I mean, I probably did not make his day because I insisted that we have a tasty beverage.

23:59 Play from here And by the time I showed up for the tasty beverage, I had, I had vibe coded a tool that

24:04 Play from here deconstructed all of my fund holdings into individual holdings and did like a targeted

24:11 Play from here AI exposure analysis to the positive and to the negative.

24:15 Play from here I was like, first, this is a really useful tool for me to understand what my AI exposure

24:20 Play from here is using broad based funds and things like that.

24:23 Play from here And second, you should be completely horrified because I hacked this thing together, not

24:27 Play from here even using cloud code, but using cloud co-work, which is kind of the nerfed on desktop version

24:33 Play from here of it in like 45 minutes and it's usable.

24:36 Play from here And you know, he's senior, he's, he's going to be fine, but it's one of those little pieces

24:40 Play from here that's like, wow, I was just able to amplify something in order to answer a question that

24:49 Play from here the human that I'm working with wasn't particularly well equipped to answer.

24:54 Play from here You are paying who I'm paying again, you know, and I, as I told him, his job is actually

25:00 Play from here more psychology, I would say, than direct investment in a lot of ways.

25:05 Play from here And I think that's a lot of financial advisors that are out there like, eh, there's, I think

25:10 Play from here it was really pointed.

25:11 Play from here Yeah, I agree.

25:13 Play from here You know, another interesting thing with the whole comparing it to COVID, right.

25:17 Play from here We're sitting on the precipice of the world changing and maybe not for the better in the

25:23 Play from here short term, just because of the potential job disruption, which we're already starting

25:27 Play from here to see, right.

25:28 Play from here It's, it's no longer just a rumor or somebody's forward thinking.

25:32 Play from here We're seeing numbers now where AI is definitely changing the job market.

25:35 Play from here It's, it's, it's happening.

25:36 Play from here It's, it's, it's changing the public markets as you try to point out, it's changing the

25:41 Play from here retail markets.

25:42 Play from here Right.

25:43 Play from here So one of the questions I was, well, well, what can you do, right?

25:46 Play from here Because everybody, and I mean, everybody, even the researchers working on all these

25:50 Play from here big models, don't really know what's around the corner, right?

25:55 Play from here Nobody, nobody can really see half a year into the future anymore as far as, well, what's

26:00 Play from here going to be the level of impact?

26:03 Play from here How fast is the change going to occur?

26:06 Play from here And back with COVID, it's so funny.

26:07 Play from here I had written a blog post when it started to look a little concerning, right?

26:13 Play from here And it wasn't quite in the US yet, but you could see what was happening in Asia and being

26:19 Play from here a nerd, you know, I just start doing the math and looking, okay, well, where could this

26:23 Play from here be going?

26:24 Play from here Huh?

26:25 Play from here That's interesting.

26:26 Play from here And that's a little concerning and oh, now this is a little bit scary.

26:30 Play from here And I think that, you know, we're kind of starting to approach that time as the essay

26:34 Play from here points out that it makes sense to think ahead, you know, you hope for the best, you plan

26:38 Play from here for the worst and maybe toilet paper won't be in short supply when the AI revolution

26:47 Play from here finally lands upon us.

26:48 Play from here But there are other things that people probably want to start paying attention to.

26:52 Play from here Like I thought it was reasonable advice to say, like, hey, get your finances in order

26:55 Play from here and just kind of look at what you're spending on.

26:59 Play from here And if the economy gets rocky, which a lot of people are predicting is inevitable, like

27:04 Play from here it will get rocky for a bit and then hopefully be better on the other side of it.

27:07 Play from here How can you do to prepare for that?

27:09 Play from here I'm so curious.

27:10 Play from here Are those things that you brought up to your advisor as well?

27:13 Play from here And I'm so curious what his response was.

27:15 Play from here We were focused more about just funding like actual exposure.

27:19 Play from here And I was pretty obsessed with the fact that it didn't feel like he was getting mothership

27:22 Play from here guidance that, you know, these are exceptionally well-paid, exceptionally smart, trained people.

27:28 Play from here And if they aren't thinking about it, like that's where major, major market exposure

27:33 Play from here can come from.

27:35 Play from here There are certainly people out there who are all over this, of course, and they're going

27:38 Play from here to make gazillions of dollars by making the right sort of trades.

27:42 Play from here Yeah.

27:43 Play from here You know, always a good idea to have some sort of a cushion in place.

27:48 Play from here We've talked a little bit about companies that are on the existential horizon, that

27:52 Play from here these are companies that recognize that this is kind of coming for them and they have to

27:55 Play from here figure it out.

27:56 Play from here I would say those companies are the most prepared.

27:59 Play from here A lot of other companies that their fundamental business, making toilet paper or whatever

28:04 Play from here it is, like is, they're still going to be doing the same thing.

28:08 Play from here They might, some of the processes and sales processes and whatever might change, but fundamentally

28:13 Play from here their purpose in the economy isn't going to change.

28:16 Play from here I think they're certainly less concerned about it and they feel less of an imperative to

28:21 Play from here deal with it.

28:23 Play from here But any company that's public, any company that's owned by private equity, any company

28:27 Play from here that has a very aggressive ownership, these are companies that are really, really starting

28:32 Play from here to dive into cost-cutting, whether that be from efficiencies or headcount.

28:39 Play from here Yeah.

28:40 Play from here Even some of the really big ones like Amazon, right?

28:42 Play from here We've seen tons of layoffs from Amazon and they're pointing directly to AI for a lot

28:48 Play from here of it.

28:49 Play from here So that's, yeah, that's real, that's happening.

28:51 Play from here I want to push back though, because I believe that there has been a little bit of a counterpoint

28:58 Play from here to that that has been floating around that I want to address, which is there is some

29:02 Play from here data suggesting that AI creates more rework.

29:07 Play from here And that's, there's some workday studies that came out a couple of weeks ago about AI solutions

29:13 Play from here actually creating too many cycles on the rework.

29:16 Play from here My answer to that is it probably wasn't built particularly well, and it was built.

29:19 Play from here probably built to replicate some human process that-

29:22 Play from here Whoever built that didn't spend their entire weekend

29:24 Play from here putting it together.

29:25 Play from here Well, yeah, whatever it was.

29:26 Play from here Oh, those chumps at workday,

29:28 Play from here they certainly don't have resources, come on.

29:30 Play from here No, it's like super smart people,

29:31 Play from here but they're not doing a great job of cultural change.

29:34 Play from here Like a little bit of my own soap box,

29:36 Play from here but you can't just like put an automation in place.

29:38 Play from here You've got to train people.

29:39 Play from here You've got to really work with teams

29:42 Play from here so they understand how it works.

29:43 Play from here Otherwise, predictable results.

29:44 Play from here That's different.

29:45 Play from here That's not what I'm talking about.

29:47 Play from here What I'm talking about a little bit

29:48 Play from here is what we were talking about a few minutes ago

29:50 Play from here about baking cookies.

29:52 Play from here And you're not baking one batch of cookies,

29:54 Play from here you're baking five batches of cookies.

29:57 Play from here And I have never been so productive in my life.

30:01 Play from here Absolutely.

30:02 Play from here Which means that as the head of an organization,

30:06 Play from here I am creating a complete whirlwind of output

30:10 Play from here that needs to be managed now

30:12 Play from here because we're bridging into new capabilities land

30:15 Play from here of the things that we couldn't do before,

30:18 Play from here but now, wow, this is something that can really change

30:21 Play from here the way that we operate

30:22 Play from here and the way that our business works.

30:24 Play from here But that means that my team

30:26 Play from here is getting this like deluge from me.

30:28 Play from here And it started with WhisperFlow and dictation.

30:32 Play from here So we've talked about this before

30:34 Play from here as being like my favorite tool ever.

30:36 Play from here You know, shorthand,

30:38 Play from here if you're not already using LLM based dictation,

30:41 Play from here you need to be.

30:42 Play from here Mac app called WhisperFlow.

30:43 Play from here There's another one called Super Whisper.

30:45 Play from here There are a bunch of them out there.

30:46 Play from here Kind of doesn't matter.

30:48 Play from here But the point is that LLMs in particular

30:50 Play from here want more information versus less.

30:52 Play from here So if you dictate, you get a lot more information out.

30:55 Play from here You give the LLM more context

30:57 Play from here and it has more to chew on and you can proceed.

31:00 Play from here Well, okay, that's step one.

31:01 Play from here Step two is, oh my gosh,

31:04 Play from here I can get so much out of my head and into the systems.

31:07 Play from here It creates more of that deluge onto my people, right?

31:11 Play from here That I actually might need more people rather than fewer

31:15 Play from here because there are so many things

31:16 Play from here that are now going on in the organization.

31:19 Play from here But, but, but you've got to be careful

31:20 Play from here about how you're prioritizing this stuff.

31:23 Play from here And this is, you know, AI at work advice.

31:25 Play from here Like you can't just be building things

31:27 Play from here because they're fun to build.

31:29 Play from here I mean, there's a stage where you're doing that

31:30 Play from here because you're experimenting.

31:31 Play from here And I'll tell anybody out there

31:33 Play from here that start with something that sounds fun to build

31:34 Play from here and build it, just ship it, get it out the door.

31:37 Play from here But pretty quickly you have

31:39 Play from here a thousand and one different optimization projects.

31:43 Play from here And you and I are both guilty of this.

31:45 Play from here I know.

31:46 Play from here And.

31:47 Play from here And by optimization,

31:48 Play from here you really mean just creating a bunch more work

31:50 Play from here for yourself and other people.

31:51 Play from here This is what ends up happening.

31:52 Play from here Perhaps, but optimization of trying to like

31:54 Play from here streamline things and have communications channels

31:57 Play from here or whatever.

31:58 Play from here And then there's servicing clients

32:00 Play from here to the utmost that you can.

32:02 Play from here And then there's acquiring new clients

32:04 Play from here and making sure that you're focused on the thing,

32:07 Play from here the things that are most important strategically

32:09 Play from here while being mindful of the things

32:11 Play from here that you're going to have to build

32:13 Play from here with the scale that that opportunity creates.

32:15 Play from here Yep.

32:16 Play from here And I think one of the biggest challenges

32:17 Play from here is that you can address all of those things

32:20 Play from here in all departments and go a million different directions

32:23 Play from here and end up really just going an inch deep

32:27 Play from here and a mile wide with all of it, right?

32:29 Play from here You don't actually move the needle on any of it.

32:31 Play from here You end up creating a ton of busy work.

32:33 Play from here So I totally agree.

32:35 Play from here The whole curiosity part of this

32:37 Play from here and just playing with stuff

32:38 Play from here and seeing what the capabilities are, are so important.

32:40 Play from here But I often find myself then going down the rabbit hole

32:43 Play from here of just thinking, well, what if?

32:45 Play from here Like, what if I could make this?

32:50 Play from here That would be so cool, right?

32:51 Play from here But cool doesn't necessarily mean

32:53 Play from here that it's the best use of my time.

32:55 Play from here And as we kind of opened up with the show,

32:57 Play from here speaking about all this stuff still,

32:59 Play from here it takes a lot of time.

33:00 Play from here It takes a lot of time.

33:01 Play from here It goes by really, really fast to get this stuff dialed in.

33:04 Play from here And then you have all the time in.

33:06 Play from here And then as you just referenced as well,

33:08 Play from here all of these outputs then can consume

33:11 Play from here a lot of other people's time too.

33:12 Play from here So you do have to be very choosy

33:14 Play from here about what you focus on, how deep you go,

33:17 Play from here and what problem you're really trying

33:19 Play from here to solve with this stuff.

33:19 Play from here Well, it lends itself to what does the organization

33:24 Play from here of the future look like?

33:25 Play from here And we're very lucky in that we run AI first,

33:29 Play from here lean organizations, so we get to build these things

33:31 Play from here and make choices between scaling with humans

33:35 Play from here or scaling with optimization, right?

33:38 Play from here But when I look at how I want my client organizations

33:43 Play from here to develop, the most forward-looking organizations

33:48 Play from here are leaning into building these capabilities

33:51 Play from here amongst a select cohort of high performers

33:56 Play from here in the organization who understand the business processes

33:59 Play from here and who now understand the technology.

34:02 Play from here And the two combined together is like really amplifying.

34:05 Play from here I have other clients that I work with

34:06 Play from here that are really, really big,

34:08 Play from here and they just have these strict guardrails,

34:10 Play from here and they're using platforms that deliberately

34:14 Play from here kind of nerf the ability for people to maneuver.

34:17 Play from here And it's hard for me to imagine those organizations

34:21 Play from here developing that cadre of people who's going to make change.

34:28 Play from here So we do this like AI office thing

34:31 Play from here where we're doing the enablement

34:33 Play from here and we're getting the ideas,

34:35 Play from here but they need our SWAT team to come in

34:38 Play from here and actually build the pieces.

34:39 Play from here And I know you do similar type work.

34:42 Play from here And I think that model is where we're going to end up,

34:46 Play from here not like I'm trying to do us out of business

34:49 Play from here because there's plenty of work.

34:50 Play from here That's what companies are going to do.

34:52 Play from here They're going to build this smaller team

34:54 Play from here that maybe sits aside their IT team.

34:57 Play from here I wouldn't actually embed it into their IT team.

35:00 Play from here I would put it adjacent to their IT team.

35:02 Play from here And it is people who really understand the technologies,

35:06 Play from here really understand the business problems

35:08 Play from here and have specific roles as titles like vibe coder,

35:12 Play from here that it's actually a role to vibe code the solutions

35:16 Play from here that the business is committing.

35:17 Play from here Which we would have laughed at a year ago,

35:19 Play from here right next to the business card that says growth ninja.

35:22 Play from here Growth ninja.

35:22 Play from here But it's a real thing now.

35:24 Play from here Like that's a-

35:25 Play from here It really is.

35:26 Play from here That's a real thing.

35:27 Play from here I think it brings up an interesting point

35:28 Play from here from that essay though too,

35:29 Play from here that there is a sentiment from him that it's like,

35:32 Play from here hey, I've spent my whole life honing this skill

35:35 Play from here and being really good at this thing.

35:36 Play from here And I'm now at the point where I feel like

35:38 Play from here I'm not really needed with that skill anymore.

35:41 Play from here I tell it what to do through a simple conversation.

35:44 Play from here I walk away, I come back at the end of the day

35:45 Play from here and this thing is built.

35:47 Play from here All this code is there.

35:49 Play from here But I think it's also easy to gloss over

35:51 Play from here what you just referenced,

35:52 Play from here which is that domain of expertise, right?

35:54 Play from here That's always going to be important.

35:56 Play from here Something has to be looking at these outputs,

35:58 Play from here seeing if they're good outputs, bad outputs.

36:00 Play from here If the thing just stopped in the middle

36:02 Play from here and gave you something half-baked.

36:03 Play from here Domain of expertise is huge

36:05 Play from here and it doesn't really matter what industry you're in,

36:07 Play from here you're going to have your domain of expertise.

36:09 Play from here But I think that also what we've been talking about

36:11 Play from here is just that output,

36:13 Play from here that 10X output of anything that you want to do

36:15 Play from here and the things that you can create

36:16 Play from here in the same amount of time.

36:17 Play from here That part invariably does mean less people, right?

36:21 Play from here More output is less people.

36:23 Play from here And I think it just really underscores the importance

36:25 Play from here to embrace literacy with all of this stuff,

36:28 Play from here to play with it, to start using it

36:29 Play from here and seeing where you can be more productive as well

36:31 Play from here or just have more output.

36:32 Play from here You know, I go back and forth on this

36:35 Play from here because I have been that person, Phil.

36:38 Play from here It's like two years of, oh my gosh,

36:40 Play from here this is going to like disrupt all of these jobs and whatnot.

36:43 Play from here But it's, I think they're calling it the whole pie idea

36:48 Play from here that, you know, the pie is the pie

36:50 Play from here and we're going to do away

36:51 Play from here with all of the available pieces of the pie.

36:53 Play from here And I'm starting to get a glimmer

36:55 Play from here that that pie is just going to get a lot bigger.

36:58 Play from here And it doesn't mean that it's going to be smooth.

37:00 Play from here It doesn't mean that there isn't going to be

37:02 Play from here a lot of dislodgement from pieces of the pie

37:04 Play from here that aren't as relevant as they used to be.

37:07 Play from here But this whole realm of new capabilities,

37:10 Play from here if you have those capabilities,

37:12 Play from here there's just going to be massive demand for them.

37:15 Play from here And I could see, you know, even in my own organization,

37:18 Play from here where I'm going to need many, many more people

37:22 Play from here if we continue to scale.

37:23 Play from here Yeah, yeah.

37:24 Play from here It's a fair point.

37:26 Play from here And again, and I've said this a million times,

37:29 Play from here the truth is nobody really knows

37:32 Play from here what the next chapter is going to look like exactly.

37:36 Play from here All we can do is really guess.

37:37 Play from here We can look at what's going on right now.

37:39 Play from here And one thing is for certain,

37:41 Play from here the things that are happening inside of these Frontier Labs,

37:43 Play from here you know, as just us, Joe Q Public,

37:46 Play from here sitting here on the outside using these tools,

37:48 Play from here paying for all the expensive licenses and the best models,

37:51 Play from here we're still only seeing a piece

37:53 Play from here of what's actually happening underneath the surface

37:55 Play from here and don't have an exact sense

37:58 Play from here as to how quickly it's going to evolve.

38:00 Play from here I did hear a counter argument to that essay

38:04 Play from here that I thought was interesting.

38:05 Play from here They just talked about the inherent friction

38:08 Play from here that people have and the amount of time it takes

38:12 Play from here for people to really get their arms around this stuff

38:16 Play from here or to embed it in their organization.

38:19 Play from here And that it's not going to be some like overnight,

38:22 Play from here oh my goodness, all the jobs are just all of a sudden gone,

38:24 Play from here that there's more time than we think.

38:27 Play from here And that the human time of adoption

38:28 Play from here is not going to match the actual scale

38:31 Play from here and rate of advancement with the models themselves

38:34 Play from here and what they're capable of.

38:35 Play from here And I thought, yeah, that makes sense.

38:37 Play from here Like that tracks.

38:38 Play from here Okay, I want to, it feels like this is my role.

38:41 Play from here I want to be able to bring it as far back

38:43 Play from here to the practical as we can, right?

38:45 Play from here And I think that one of the observations that I have,

38:48 Play from here you said it a second ago,

38:49 Play from here that, oh, professional vibe coding a year ago.

38:51 Play from here Like, okay, you know,

38:53 Play from here I started playing around with vibe coding.

38:57 Play from here It was really only a year.

38:58 Play from here Like the term was only invented like a year ago

39:00 Play from here when Repl.it and Lovable and things like that popped up.

39:04 Play from here And I have to admit, like I was like,

39:06 Play from here hey, this is really cool.

39:08 Play from here And then I would look at what I had actually produced

39:10 Play from here and nothing was ever usable.

39:11 Play from here So I started telling people,

39:13 Play from here hey, vibe coding is really about the MVP.

39:15 Play from here It's about being able to really rapidly put together

39:20 Play from here before you go spend a bunch of money with Eli

39:22 Play from here to build the final solution, whatever you need.

39:25 Play from here Shouldn't say final solution.

39:27 Play from here Shouldn't say a bunch of money either, buddy.

39:28 Play from here Shouldn't say a bunch of money.

39:29 Play from here I mean, highly efficient.

39:34 Play from here Double faux pas one sentence.

39:35 Play from here Yes.

39:37 Play from here In any case.

39:38 Play from here So it did feel like a little bit of a toy.

39:40 Play from here And then when Clod code came on

39:44 Play from here and particularly once, you know,

39:47 Play from here a bunch of stuff all happened at once.

39:49 Play from here You had the new Clod model, which was Opus 4.6 launch.

39:53 Play from here You had Codex, which is the coding model

39:56 Play from here within OpenAI launch.

39:59 Play from here And these are really, really, really good coding models.

40:01 Play from here And it was like a switch flipped that all of a sudden

40:06 Play from here we were no longer just building for concept or for prototype.

40:09 Play from here You were building for real.

40:11 Play from here And that has just opened up this whole new world

40:14 Play from here for organizations to experiment.

40:16 Play from here So while I wouldn't encourage you to go out

40:18 Play from here and buy your Mac mini and start playing with OpenCLO.

40:21 Play from here I 100% would encourage you to splurge

40:25 Play from here for that $100 Clod Max plan, which I know, oh my gosh,

40:29 Play from here we've all been paying $20 for this magic machine.

40:32 Play from here But once you have the $100 plan,

40:36 Play from here and if you're on a Mac, use the Clod desktop app,

40:40 Play from here and now you can have Clod code there.

40:42 Play from here And then take a breath.

40:43 Play from here And you want to figure out what you want to build.

40:45 Play from here So on your other screen.

40:46 Play from here And you're using that Max plan

40:47 Play from here because on your $20 a month plan,

40:49 Play from here it just ran out of tokens way too quickly.

40:51 Play from here It was like minutes.

40:53 Play from here So, but I'm fine on the Max plan.

40:56 Play from here Even despite my 30 hours this weekend,

40:58 Play from here although it did kick me off.

40:59 Play from here You didn't hit any roadblocks there.

41:01 Play from here I kicked me off every five hours, but I didn't.

41:03 Play from here It was great except for every five hours I had to go.

41:06 Play from here Oh, well, you know.

41:07 Play from here Get on my bike and get some coffee and have lunch.

41:08 Play from here I had to go like eat.

41:09 Play from here Wait for the little clock to reset.

41:13 Play from here It's like the model was like saving the little human.

41:16 Play from here But on one screen, I had normal Clod up.

41:20 Play from here And I created a Clod project.

41:22 Play from here And this is all actionable, okay?

41:24 Play from here So create a Clod project.

41:25 Play from here You could do this in GPT as you want.

41:27 Play from here And then you want to use that project

41:30 Play from here to contain like all of your information

41:32 Play from here that you're going to be working with.

41:33 Play from here And then you want to spend probably 60% of your mental time

41:37 Play from here figuring out what it is you want to build

41:40 Play from here and go back and forth with Clod or OpenAI.

41:43 Play from here Have it ask you questions about your scope

41:45 Play from here and what you want.

41:46 Play from here And those questions will lead you down rabbit holes

41:48 Play from here as far as the things you need.

41:50 Play from here It will create a prompt and a scope document.

41:53 Play from here And now you can start going with Clod code

41:55 Play from here and you can build things.

41:56 Play from here And then, oh my gosh, it's scary to build things.

41:59 Play from here Okay, you are going to have to figure out things like GitHub

42:02 Play from here if you've never touched GitHub before.

42:04 Play from here You are going to have to figure out things like Vercel,

42:07 Play from here which is, you know, this amazing company

42:09 Play from here that's been hosting apps forever

42:10 Play from here that is remarkably well-positioned to-

42:13 Play from here Also some pretty crazy building tools as well now.

42:14 Play from here Yeah, they've been building all kinds of stuff.

42:16 Play from here But once you, because you have Clod or your GPT

42:21 Play from here or your LLM open on one screen, you just ask it questions.

42:24 Play from here I've decided that probably the core skillset right now

42:28 Play from here is copying and pasting.

42:29 Play from here This is what I'm seeing in Vercel.

42:31 Play from here Can you help me like configure this particular screen

42:34 Play from here to do what I need to do to launch the app?

42:36 Play from here Are you taking like entire screenshots and just-

42:38 Play from here Yeah, absolutely.

42:39 Play from here I'm totally lazy now.

42:40 Play from here I'm like, just interpret the screen.

42:42 Play from here Let me know what's going on.

42:44 Play from here It's so weird that you're running out of tokens

42:45 Play from here so quickly.

42:46 Play from here I'm not ready to go tokens.

42:47 Play from here I'm fine.

42:48 Play from here I'm totally fine.

42:49 Play from here But you don't really have to be efficient about it.

42:52 Play from here You can kind of blunder your way through

42:54 Play from here and it doesn't have to cost much.

42:56 Play from here Okay, Clod code does cost a hundred dollars a month,

42:58 Play from here a hundred dollars a month.

43:00 Play from here Vercel, you can run on the free plan.

43:02 Play from here GitHub, you can run on the free plan.

43:04 Play from here Like you don't need to get roped into things like Lovable,

43:09 Play from here which actually can be really expensive to develop an app.

43:12 Play from here But if anything I said is intimidating,

43:14 Play from here by all means, set up a Lovable account, build something.

43:17 Play from here Lovable is now-

43:18 Play from here So you used Lovable for a while, right?

43:20 Play from here And I think you used Repl.it and Cursor as well.

43:22 Play from here Yeah.

43:23 Play from here Correct?

43:24 Play from here And you seem to be in camp Clod code now,

43:27 Play from here which I like.

43:28 Play from here I like that.

43:28 Play from here But tell me about the difference of your experience

43:31 Play from here with those other products, which I don't know,

43:33 Play from here I'd say would probably just be a little more user-friendly

43:37 Play from here or kind of help coach you along

43:39 Play from here compared to some of the hurdles

43:41 Play from here that people might get scared of that you mentioned,

43:42 Play from here like opening up a GitHub account or a Vercel-

43:46 Play from here Well, you need a GitHub account anyway

43:48 Play from here for like Lovable or whatever you're doing,

43:49 Play from here if you're being-

43:50 Play from here Not before.

43:51 Play from here I mean, back in the days where like,

43:52 Play from here oh, we're just gonna make something cool

43:53 Play from here and it's gonna live in Lovable

43:54 Play from here and I could show my friends this cool prototype

43:56 Play from here and then hire somebody to build it for me.

43:58 Play from here But I think now the game has changed.

43:59 Play from here materially, right? So the expectations are different. So yeah, what are you seeing now

44:04 Play from here versus what you were with these more, I don't know, I'll call them like consumer-friendly versions

44:09 Play from here of things that would get you to the same end result, hopefully. My biggest issue with Lovable

44:13 Play from here is that it ends up being very expensive. Where's the cost coming from? The cost is coming from

44:19 Play from here the tokens that you're using or they call them credits or whatever it is on your account.

44:24 Play from here And every time you ask it a question or you do planning, it's actually disproportionate to the

44:30 Play from here number of tokens that you'd be consuming somewhere else. So yes, I've built a bunch of stuff on

44:35 Play from here Lovable, which is really great and it's really pretty and it's nice to use. But in my 30 hours

44:40 Play from here over that weekend that I was building things there, it cost me like $300.

44:45 Play from here Wow, really?

44:46 Play from here Yeah, absolutely.

44:48 Play from here Dang.

44:48 Play from here And that's even using-

44:50 Play from here Tough on Mac mini, man.

44:51 Play from here Well, pretty much, right? So okay, but backing up, if you want just super easy,

44:56 Play from here like I'm not knocking Lovable. I think they have a great business. And the couple of key

45:01 Play from here features there, one is it does have the builder in its plan mode right in it. This is what I want

45:06 Play from here to build. Help me figure it out. Okay, you're going to burn a lot of tokens in that, but it's

45:11 Play from here going to be built in right there. Two is it's much more visible as far as like your outputs.

45:16 Play from here So you're building a user experience like some product showcase or whatever it is. It's popping

45:22 Play from here up like right then and there and it's really well integrated. So if you want to tweak something,

45:27 Play from here you can see it versus others, you need to be using browser tabs and things like that.

45:32 Play from here Very similar to Google Labs. You've played with that as well, right?

45:34 Play from here Yeah, yeah, yeah.

45:34 Play from here For business stuff, yeah. You get that preview. That's super helpful.

45:37 Play from here And the other piece that I find absolutely killer is they've baked the backend and the

45:42 Play from here utilities right into your subscription. So the other piece I didn't mention was the backend,

45:47 Play from here which is Superbase, is what most people use. I'm not going to get into the details of what that is,

45:52 Play from here but it's basically the database that sits behind your app and it helps control who gets access to

45:58 Play from here it and where the data goes and whatever. It's so amazing that we can just kind of blow that off now.

46:04 Play from here You know, just the backend.

46:05 Play from here Just the whole backend, right?

46:06 Play from here Onward.

46:07 Play from here So yeah, I do spend 20 bucks a month for Superbase as well for my other projects.

46:12 Play from here So that was another piece that was missing there. But Lovable is just baked Superbase right in.

46:16 Play from here So it has just an automatic connection to it. They're going to bill you based on usage.

46:23 Play from here And obviously, they're like a wholesaler of how that works and that's part of how

46:27 Play from here they're making money and whatever. But they also have baked in all these other things.

46:31 Play from here So if you want to go scrape a website for whatever you need, Firecrawl is baked right into it.

46:38 Play from here And yes, same deal. When you do a Firecrawl, Firecrawl would charge you one-tenth of a penny

46:44 Play from here and Lovable will charge you two-tenths of a penny and they're going to make their markup

46:49 Play from here and they're going to make some money on that.

46:50 Play from here The convenience of connection right there.

46:52 Play from here Yeah, but it's all baked in in this way that's really great. I just found that once you started

46:58 Play from here to do more complicated things than sort of just general visual products, A, it got really expensive

47:05 Play from here because of the tokens and your mileage may vary.

47:08 Play from here Maybe you're not as lazy as I am and you're being really, really targeted on it.

47:12 Play from here But once I broke the seal on really leaning into Cloud Code, I just...

47:18 Play from here And importantly, people hear about Cloud Code and they hear about the terminal.

47:22 Play from here Oh my gosh, it's the terminal that you have to open up, you know,

47:25 Play from here black screen and put CD in and change things.

47:28 Play from here It feels like...

47:28 Play from here You don't have to do that anymore.

47:30 Play from here Yeah, but you don't have to do that anymore.

47:32 Play from here So Cloud Desktop now has it baked right in and it's a really user-friendly type experience

47:37 Play from here that feels a lot more like just being in one of the chatbots.

47:41 Play from here And so, yeah, just kind of thinking through like, okay, Codex or Google Labs,

47:47 Play from here still going up to make this thing usable.

47:50 Play from here You're having to pay a hundred bucks a month to play around with it.

47:53 Play from here Same case when you were using...

47:55 Play from here Have you tried building stuff like in Google Labs?

47:58 Play from here I've done a lot in Google Opal, which I find interesting.

48:04 Play from here I found it too constraining and it just didn't drive with me in a way that made it my go-to.

48:11 Play from here Lovable was the first one that was really...

48:14 Play from here I fell in love with Lovable.

48:16 Play from here Like way back, like, oh, this is really cool.

48:19 Play from here I can see the way this is going.

48:21 Play from here And then as Lovable got better, it was the first one that I actually used

48:24 Play from here to produce functional apps with.

48:26 Play from here And then now I've moved on.

48:28 Play from here So you got production-ready stuff out of Lovable?

48:30 Play from here Oh, absolutely.

48:31 Play from here Yeah.

48:31 Play from here Let's give an example of something that you were able to get out of that.

48:34 Play from here No, not even two months ago were we really able to get client-ready stuff out of it.

48:39 Play from here Okay.

48:39 Play from here Like little apps that your clients could use.

48:42 Play from here Yeah.

48:42 Play from here Well, or we're using as like part of...

48:45 Play from here Usually it's some sort of presentation tool or whatever it might be.

48:49 Play from here We do, since we do a lot of custom training,

48:51 Play from here we use Lovable to spin up custom landing pages.

48:55 Play from here To do course curation and things like that for individual clients,

48:59 Play from here which is good for that sort of thing.

49:01 Play from here Right.

49:01 Play from here Just to have a quick design and deploy it super fast.

49:04 Play from here And are you then taking the design that you'd get from something like that,

49:07 Play from here like just the raw HTML and then pasting it into a CMS or using

49:11 Play from here their baked-in hosting services and everything as well?

49:13 Play from here Just baking it right in.

49:16 Play from here We are using a subdomain, so it's not like declaring it as Lovable.

49:20 Play from here But yeah, we're basically just using their hosting services.

49:23 Play from here We don't want to get overly technical.

49:24 Play from here How much does that cost for their hosting?

49:26 Play from here I'm so curious.

49:27 Play from here It's all part of your subscription.

49:29 Play from here It hasn't been material for...

49:32 Play from here We don't...

49:33 Play from here I sound like I'm objecting to,

49:35 Play from here oh, I had to pay $50 whole dollars.

49:39 Play from here I was going to bring that up, man.

49:41 Play from here It's so wild.

49:42 Play from here Like even the, oh, I'm going to pay $200 a month to use this model.

49:46 Play from here And then you think about the things that you can build in a weekend with that model that...

49:50 Play from here And I'm not even exaggerating right now.

49:52 Play from here That just a few years ago, if you went to a developer and said,

49:55 Play from here make me this product, it'd be like a $30,000 to $50,000 price tag for some of this stuff.

50:00 Play from here Just because the sheer number of hours that it would take that developer to create something

50:04 Play from here with that functionality at that level of fidelity that you can knock out

50:08 Play from here over the course of a weekend and have the use of an entire month for $100, $200.

50:14 Play from here Yet we still, I mean, I feel it all the time.

50:17 Play from here I feel squeamish about like, I don't need another subscription, $20 a month.

50:22 Play from here Probably because then I ended up getting lazy.

50:24 Play from here And even when I don't use it anymore, it's just sitting on my credit card,

50:27 Play from here like getting it every single month.

50:30 Play from here I think I've had it on my to-do list now for half a year.

50:33 Play from here It's just audit all of that.

50:35 Play from here Make sure that I unsubscribe to the things I'm not using anymore

50:37 Play from here because I'm sure there's a ton of them.

50:39 Play from here But no, that's fantastic, man.

50:41 Play from here We talk a lot about little apps that we can make that our clients would get a lot of utility out

50:47 Play from here of that we could just kind of put out there, promote them on LinkedIn, and then have people

50:52 Play from here use them and kind of see the magic of some of the things that we can help out with, right?

50:56 Play from here That we can make.

50:57 Play from here I think you made one.

50:59 Play from here Was that, I think what you were working on this weekend, right?

51:01 Play from here Was one of those?

51:02 Play from here Yeah.

51:02 Play from here Yeah.

51:03 Play from here So launchpad.ascendlabs.ai is it's an AI use case generator.

51:12 Play from here And I'm super proud of it, to be honest, that it's, you know,

51:16 Play from here this is what you do in your company.

51:17 Play from here This is your role.

51:18 Play from here This is your industry.

51:20 Play from here These are the sorts of things that are bugging you.

51:23 Play from here These are the tools you have.

51:24 Play from here And it synthesizes all of that and gives you a crawl, walk, run assortment of direct,

51:33 Play from here practical AI projects that you can do today.

51:37 Play from here Well, the crawl ones you can do today, the walk ones might take you a few days and the run ones,

51:42 Play from here well, they're a little bit more complicated.

51:44 Play from here And then you can choose one.

51:46 Play from here And it's going to spit out a really highly detailed lead on the other end.

51:50 Play from here This is where you're doing it in the live and it's going to break and it's going to

51:53 Play from here make me cry slightly inside.

51:56 Play from here That's the other problem.

51:57 Play from here And I talk about this with my partner frequently that, you know, you have to also be able to

52:02 Play from here maintain these things.

52:03 Play from here And it's cool to have built something like that.

52:06 Play from here And it's live and you can go to it and that's going to give value to anybody who goes there.

52:09 Play from here So launchpad.ascendailabs.ai.

52:13 Play from here But if it breaks...

52:15 Play from here Hold on, say that one more time, because I think I nerfed that URL.

52:19 Play from here I thought I heard AI like six times in that URL.

52:22 Play from here It'll be in the notes.

52:24 Play from here Okay, I'll wait and I'll click it in the notes.

52:28 Play from here I don't get totally distracted and start like...

52:30 Play from here I'm proud of it.

52:30 Play from here Just let you talk for the next 15 minutes.

52:32 Play from here I was trying...

52:33 Play from here You know, it has a processing bar because of course it's being driven by Opus 4.6.

52:39 Play from here So a really expensive kind of slow model behind the scenes.

52:44 Play from here That's brave of you to just put that out in public.

52:47 Play from here It's fine.

52:48 Play from here Hook your credit card up to...

52:49 Play from here Yeah, hook my credit card right up to Dario and just like an IV.

52:54 Play from here Seriously.

52:55 Play from here Right in there.

52:56 Play from here Yeah.

52:57 Play from here Okay, okay.

52:58 Play from here Who knows the really big X following?

52:59 Play from here I think I'll give them that link and really promote it for you.

53:02 Play from here And you're either going to do a ton of business or go bankrupt.

53:05 Play from here It's one or the other.

53:06 Play from here Important practical point.

53:08 Play from here If you're playing with these tools in the live,

53:11 Play from here you need to make sure that your APIs are configured in a way that you're...

53:15 Play from here That's not going to happen.

53:16 Play from here Like if this podcast went viral and a million people ended up on there,

53:21 Play from here then, I mean, good for my business, but also really, really bad for my API costs.

53:26 Play from here So I have caps on all of my API costs of sort of my risk threshold if something like that happens.

53:32 Play from here But in some platforms, I'm looking at you, Google, that's actually really hard to do.

53:37 Play from here And a few weeks ago, I said that when my first experience on OpenCLI

53:41 Play from here burned a hundred million tokens in three hours,

53:43 Play from here because there's really no great way of capping it.

53:46 Play from here So make sure that you're...

53:48 Play from here Configuring APIs is pretty easy to figure out.

53:50 Play from here You can ask the LLMs how to do that,

53:53 Play from here but make sure you're setting budgets and alerts and things like that.

53:56 Play from here And then the breakage part is really important too,

53:58 Play from here that, okay, this thing is just going live today, basically.

54:02 Play from here And yeah, it'll get a few hundred users a week or whatever it is.

54:07 Play from here But I also had to make sure I was building in feedback loops

54:10 Play from here such that I would know if it broke.

54:12 Play from here Because if it just quietly broke behind the scenes, that would be sadness, right?

54:17 Play from here So in my case, I love to operate in Slack.

54:20 Play from here So I built an alert that pushes a notification to a Slack channel

54:24 Play from here that will tell me if one of my tools is broken.

54:26 Play from here Oh, just like you'll get the error message right in Slack

54:28 Play from here and you can go back in and see if you can fix it.

54:31 Play from here Yeah, funny enough, when looking for those API limits

54:35 Play from here inside of Google's ecosystem for Gemini,

54:37 Play from here one of the worst people to ask on where to find that is Gemini.

54:41 Play from here I can't tell you how many prompts I went through

54:44 Play from here where I would send a screenshot over Kevin Williams style,

54:47 Play from here just grab that whole screen, be like, it's not there.

54:49 Play from here My mistake, the UI must've been updated recently.

54:52 Play from here I'm like, nope, nope.

54:53 Play from here I can tell you this hasn't changed in many, many, many months.

54:56 Play from here It's not there.

54:57 Play from here And then it is a bit of a dumpster dive

54:59 Play from here in order to get through all of Google's layered menus,

55:03 Play from here but that's not news.

55:04 Play from here It's always been like that.

55:05 Play from here Even if you've had a Google workspace

55:07 Play from here and you need to make one small change in the configuration,

55:09 Play from here you're like, I'm gonna block off three hours

55:11 Play from here and figure this out.

55:13 Play from here Well, we've said this before,

55:14 Play from here but I am never gonna count Google out.

55:17 Play from here Like Google is a total player in this.

55:19 Play from here We will be using Google services

55:23 Play from here once we've evolved beyond cloud code or whatever.

55:26 Play from here It wouldn't surprise me at all,

55:28 Play from here but it is Byzantine to deal with Google

55:31 Play from here and all of the APIs and all of the things.

55:34 Play from here It's tough to even find out what capabilities

55:37 Play from here live in that ecosystem at any given time, right?

55:39 Play from here And I'm not-

55:39 Play from here Very disconnected.

55:40 Play from here I'm not doing anything that is industrial scale SaaS volume.

55:43 Play from here So yeah, I know you blanched it.

55:45 Play from here Oh my gosh, you're using cloud 4.6.

55:47 Play from here That's gonna be really expensive.

55:49 Play from here Yeah, but not really if it's working

55:52 Play from here and trying to economize by using the Google universe,

55:56 Play from here if it exposes me to additional risk

55:59 Play from here or if it's complicated.

56:00 Play from here I did figure out the limits.

56:03 Play from here I got it dialed in.

56:04 Play from here It did take me some time.

56:06 Play from here I bet it took you some time.

56:07 Play from here Honestly, if I was gonna bet on one horse,

56:10 Play from here I'd bet on Google.

56:11 Play from here I really would.

56:12 Play from here Just if you look at across all the different verticals,

56:14 Play from here all the different domains,

56:15 Play from here history, expertise, funding footprint,

56:18 Play from here the amount of cash that they're just spitting out right now

56:21 Play from here compared to the labs that are just like burning

56:23 Play from here immense piles of money

56:24 Play from here and just continuing to dump more on the fire

56:26 Play from here from VC and private equity.

56:28 Play from here Yeah, I bet on Google all day long,

56:30 Play from here but yeah, not always the easiest,

56:32 Play from here the most intuitive layer of menus to work through.

56:34 Play from here Well, at the moment I'm using just a mix of Opus and Sonnet.

56:38 Play from here So those are the anthropic models.

56:39 Play from here Sonnet's really cheap,

56:40 Play from here but if you need heavy lifting, do the heavy lifting.

56:43 Play from here I have a chat open with a friend

56:45 Play from here and he's like putting himself into pretzels,

56:48 Play from here trying to figure out how to use this model,

56:50 Play from here this is this model and when,

56:52 Play from here and then compare all the outputs.

56:54 Play from here And like, he's building like a comparator type

56:59 Play from here of to see what all the prompt outputs are

57:01 Play from here so he can figure out the right one to use.

57:03 Play from here I'm like, oh, I did this, yeah.

57:05 Play from here Just okay, fine, but just use a high-end frontier model.

57:09 Play from here Just use whatever the best model du jour is.

57:11 Play from here And the rest is gonna be prompting.

57:13 Play from here It's gonna be making sure that the prompting is right.

57:16 Play from here Which sounds like for you, it's Opus now, right?

57:18 Play from here I mean, I feel like every week we have this discussion

57:20 Play from here of like, okay, so who's your go-to girlfriend this week?

57:23 Play from here Who are you in love with?

57:25 Play from here So my wife calls a chat GPT, Chatty.

57:28 Play from here It's funny, she's named it.

57:29 Play from here But another friend calls it Chachi.

57:31 Play from here So I guess that's an audience question.

57:33 Play from here Do you have a name for your LLM?

57:36 Play from here I do not.

57:38 Play from here I am like out of love with chat GPT right now.

57:40 Play from here It's driving me bonkers.

57:41 Play from here And the fact that now I have basically

57:43 Play from here unlimited use of Claude,

57:46 Play from here that was my limiting factor with Claude before

57:48 Play from here because of the $20 a month plan,

57:50 Play from here it would cut you off pretty quick

57:51 Play from here if you're doing complicated things.

57:54 Play from here And now I can work with it for hours and hours at an end.

57:57 Play from here And just this morning, I was trying to do something in GPT

58:00 Play from here and it was just driving me nuts.

58:01 Play from here And it was being like sycophantic and dumb.

58:04 Play from here And I clicked right over to Claude and got what I needed.

58:07 Play from here I had a similar experience this week as well.

58:10 Play from here And it's funny, it'll just take one of those experiences

58:12 Play from here to just be like, yeah, I'll come back to you later.

58:15 Play from here I'm just gonna leave this relationship for a little bit.

58:18 Play from here And we gotta go pretty soon.

58:20 Play from here Maybe next time we'll talk about this.

58:22 Play from here I really wanna talk about

58:22 Play from here if you've played with any local models as well,

58:25 Play from here KimiK 2.5 or GLM or any of these maybe scary Chinese models.

58:30 Play from here Another reason why I do wanna get a machine

58:32 Play from here just to deploy this stuff to you

58:33 Play from here so I can play around with it.

58:34 Play from here But have you messed around at all?

58:36 Play from here What's that?

58:37 Play from here Didn't we promise our audience

58:38 Play from here we weren't gonna say things like...

58:39 Play from here How about we elaborate that it's a Chinese open source model that's really good and really

58:45 Play from here powerful and really cheap.

58:47 Play from here Really cheap, emphasis on cheap.

58:49 Play from here And it really matters.

58:51 Play from here It really matters for what you're doing with your open cloud because you're going to need

58:55 Play from here to save a bunch of money.

58:56 Play from here I have not, but it's a good, there's something called open router.

59:00 Play from here So those of you who are brave enough to go bake those cookies before I am, I strongly

59:05 Play from here recommend you use open router as, as like it will route your, your needs to the most

59:11 Play from here effective model versus necessarily doing a direct API.

59:15 Play from here But Kimmy is one of those models that it can route to.

59:18 Play from here And it's a mixed bag.

59:20 Play from here Basically people are saying, in fact, this might've been you who was saying this, that,

59:24 Play from here that it was too dumb to get you set up.

59:27 Play from here So you had to use Gemini 3.0 to get yourself set up.

59:32 Play from here And then once it's set up, maybe you can start cooking with Kimmy, which sounds like

59:36 Play from here a cookbook.

59:37 Play from here It's kind of like you're a grunt worker, you know, if you can do like all the basic stuff

59:41 Play from here and you're saving a ton of money and then you route to Opus and spend a billion dollars.

59:46 Play from here Opus.

59:47 Play from here I mean, it wasn't long ago that we had GPT 4.0 and it was this amazing thing that was

59:50 Play from here super expensive.

59:51 Play from here And now it's like gone, it's actually been deprecated.

59:55 Play from here Just okay, audience, like use Opus 4.6.

59:58 Play from here It is absolutely brilliant.

01:00:00 Play from here And it's going to be the old hat model in like six months or whatever it is, which will

01:00:04 Play from here mean that it will be a lot cheaper to use.

01:00:07 Play from here And yeah, there's going to be something shinier and better and brilliant-er.

01:00:10 Play from here Let's be clear, right now, you got to pay for it.

01:00:13 Play from here You got to pay for it.

01:00:14 Play from here You got to pay for it right now.

01:00:15 Play from here If you try to even go 20 bucks a month, like have fun with your three prompts and now you're

01:00:20 Play from here done.

01:00:21 Play from here Yeah.

01:00:22 Play from here But, but so the fact is though, you don't need to advance with the best model.

01:00:25 Play from here Once you get to a tier where everything's working, and I think that's the other thing

01:00:28 Play from here that's happened right in this, in this like weeks is people have realized, oh my gosh,

01:00:33 Play from here I can just use Codex or I can just use Opus 4.6 and everything just starts to work.

01:00:40 Play from here Yeah.

01:00:41 Play from here As opposed to all this other stuff, like we don't need to use those really crazy models.

01:00:46 Play from here We don't need to drain all the lakes in the Midwest as this stuff gets more efficient

01:00:50 Play from here with what will be an older model.

01:00:52 Play from here So you want to choose the least common denominator model that accomplishes your goals.

01:00:57 Play from here And then the smartest model for the hardest part of whatever you're doing.

01:01:01 Play from here Which speaking of time also takes time to figure out because there's no off the shelf

01:01:06 Play from here playbook and every use case is different.

01:01:08 Play from here And it's very common to go under or over in both cases.

01:01:13 Play from here And you just got to find the sweet spot and that takes, it takes, it takes tinkering.

01:01:16 Play from here It's why we got to roll up our sleeves and play with this stuff.

01:01:21 Play from here Get out there, install your cloud code or your Clawbot, experiment, do stuff.

01:01:26 Play from here It is past the stage where it's just a toy.

01:01:29 Play from here Break things, just not too many things.

01:01:31 Play from here Don't break anything too precious.

01:01:32 Play from here Be curious about it.

01:01:33 Play from here Yeah.

01:01:34 Play from here Yeah.

01:01:35 Play from here Yeah.

01:01:36 Play from here Awesome, Kevin.

01:01:37 Play from here Always good to catch up with you, bud.

01:01:38 Play from here And we'll do it again next week.

01:01:39 Play from here That is the pod.

01:01:40 Play from here Bye-bye.

01:01:41 Play from here Thanks for listening.

01:01:42 Play from here If you'd like to send us a question to answer on the show, email me at Elijah at spark6.com.

01:01:53 Play from here That is S-P-A-R-K, the number six.com or Kevin at Kevin at ascendlabs.ai.

01:02:00 Play from here To learn more about AI strategy and training for your business, visit Kevin's company,

01:02:05 Play from here ascendailabs at ascendlabs.ai.

01:02:09 Play from here And to learn more about design and development of custom AI solutions, visit my agency at

01:02:14 Play from here spark6.com.

01:02:16 Play from here You can also sign up for my free newsletter, AI Framework Friday, by visiting spark6.com

01:02:22 Play from here forward slash newsletter.

How to cite this episode

Credit the named speaker and AI at Work by AscendAI. Link to the episode or the passage you quoted.

“Ep #14 Baking Five Batches at Once: How AI Multitasking Is Rewiring Work (and Weekends).” AI at Work by AscendAI, 2026-03-12. https://ascendlabs.ai/podcast/ep-14-baking-five-batches-at-once-how-ai-multitasking-is-rewiring-work-and-weekends/

Recent episodes

View all episodes
×
AI Solutions Graphic
Free Resource
Get Started Now