WEBVTT
00:00:02.399 --> 00:00:12.689
This episode of The Catalyst is brought to you by Veeam Data Cloud backup, fully managed secure cloud storage for your backups with no surprise bills and zero configuration headaches.
00:00:12.960 --> 00:00:17.670
Click the link in the description to sign up for their AWS demo series and see it in action.
00:00:20.519 --> 00:00:23.789
Here's a story about building something from nothing.
00:00:25.679 --> 00:00:27.929
When I first heard the term, I'm like, I'm gonna test this out.
00:00:27.929 --> 00:00:30.399
And I, have, I run a baseball org, as you've heard me talk about.
00:00:30.638 --> 00:00:35.500
I was like, you know what would be cool is if I had an online tool that I could get my guys to use, what might that look like?
00:00:35.500 --> 00:00:38.918
So I actually went in, I started asking Gemini, how might I do this?
00:00:39.399 --> 00:00:43.088
And I gave it some scenarios and different things I wanted to learn about, and it just.
00:00:43.134 --> 00:00:44.695
Started popping out this app for me.
00:00:44.695 --> 00:00:45.564
Do you mean like this?
00:00:45.564 --> 00:00:46.585
And I was like, oh, wow.
00:00:46.823 --> 00:00:47.723
I was blown outta my seat.
00:00:47.723 --> 00:00:48.323
I'm like, okay.
00:00:48.924 --> 00:00:50.344
That's Ron Espinsa.
00:00:50.814 --> 00:00:53.155
He leads Softchoice Google category.
00:00:53.545 --> 00:00:55.973
He's not a developer, he's not a coder.
00:00:56.454 --> 00:01:02.814
But in a single afternoon he built a custom baseball analytics app player profiles.
00:01:03.250 --> 00:01:11.379
Speed ratings, defensive positioning, game scenarios, all by describing what he wanted in plain English.
00:01:11.799 --> 00:01:12.760
I don't like these colors.
00:01:12.760 --> 00:01:13.269
Change that.
00:01:13.299 --> 00:01:13.719
Okay?
00:01:14.019 --> 00:01:15.849
And it went and changed all the color schemes.
00:01:16.179 --> 00:01:20.079
What if I could factor in instead of just offense and defense?
00:01:20.290 --> 00:01:24.040
What if I could put in player profile and it recreated the app with the player profile?
00:01:24.219 --> 00:01:26.500
Next thing I know, I had the ability to put in a player's.
00:01:27.125 --> 00:01:34.444
Name, size in terms of height and weight, their speed ratings, how they perform offensively, how they perform defensively.
00:01:34.745 --> 00:01:36.334
And then I could say, where's the play?
00:01:36.575 --> 00:01:38.885
And there was no code that I generated.
00:01:39.694 --> 00:01:42.275
Ron's baseball app is a weekend project.
00:01:42.635 --> 00:01:55.055
It's fun, it's harmless, but that same technology, AI that writes code from conversation is now inside enterprise software teams at companies of every size.
00:01:55.810 --> 00:02:02.680
And in December of 2025, it potentially caused a problem that was anything but harmless.
00:02:04.120 --> 00:02:14.020
According to the Financial Times, Amazon's own AI coding tool, an agent called Kiro was given production level access by an engineer.
00:02:14.379 --> 00:02:21.400
The AI autonomously decided to fix a bug by deleting and recreating an entire live environment.
00:02:21.939 --> 00:02:35.205
The result, a 13 hour AWS outage by March, 2026, more AI assisted deployment failures reportedly cost Amazon millions of lost orders.
00:02:36.145 --> 00:02:45.835
The same executive who pushed engineers to use AI coding tools was forced to mandate senior engineers sign off on every AI assisted code change.
00:02:46.254 --> 00:02:48.775
Amazon called it controlled friction.
00:02:55.044 --> 00:02:58.284
From Softchoice, a worldwide technology company.
00:02:58.645 --> 00:03:00.384
This is the catalyst.
00:03:00.895 --> 00:03:01.594
I am Katey Teekasingh.
00:03:02.664 --> 00:03:05.335
This season, we're doing things a bit differently.
00:03:05.694 --> 00:03:14.305
We're making audio documentaries, real stories from the front lines of it, exploring the challenges of small teams chasing big dreams.
00:03:14.514 --> 00:03:17.224
Today's episode a technology that lets.
00:03:17.460 --> 00:03:26.159
Anyone builds software by describing it in plain English, has gone from a weekend hobby to the dominant force in how code gets written.
00:03:26.340 --> 00:03:28.620
It's already inside your organization.
00:03:29.009 --> 00:03:32.490
The question is whether anyone's watching what it's building.
00:03:32.879 --> 00:03:40.199
We are calling it the Vibe Coding episode, act one, the vibe.
00:03:42.580 --> 00:03:52.030
In February, 2025, a former open AI researcher named Andre Cari posted about a new way he'd been building software.
00:03:52.569 --> 00:04:03.639
He described it as fully giving into the vibes, telling an AI what he wanted, accepting whatever code it produced, and barely looking at the results.
00:04:04.099 --> 00:04:07.579
He called it vibe coating within months.
00:04:08.090 --> 00:04:17.389
Collin's Dictionary named it their 2025 Word of the Year when our producers sat down with Sean Larkin, AI principal architect at Softchoice.
00:04:19.120 --> 00:04:20.019
He put it simply,
00:04:20.439 --> 00:04:28.810
the way I understand vibe coating, it was, there's a term that was coined by Andre Carte who said, I developed this application over the weekend.
00:04:28.810 --> 00:04:34.000
And so he coined the phrase vibe coating to mean something that's a lower level effort.
00:04:34.180 --> 00:04:35.860
Typically done over the weekend.
00:04:36.218 --> 00:04:38.980
Something you would definitely not put into production.
00:04:39.129 --> 00:04:43.990
It's the sort of experimentation, tooling, rapid prototypes you might build on a weekend.
00:04:44.500 --> 00:04:46.694
Weekend projects, experiments.
00:04:47.678 --> 00:04:49.718
Things you'd never put into production.
00:04:50.468 --> 00:04:51.728
That was the original idea.
00:04:52.629 --> 00:04:54.939
But vibe coding didn't stay on the weekend.
00:04:56.348 --> 00:05:03.879
According to industry surveys, over 90% of US developers now use AI coding tools daily.
00:05:04.209 --> 00:05:08.858
Roughly 40% of all codes written globally are now AI generated.
00:05:09.399 --> 00:05:17.048
When our producers asked Greg Whalen, CTO of prove ai, a company building observability tools.
00:05:17.288 --> 00:05:22.629
Specifically for AI systems, whether this felt like a real shift, he didn't hedge
00:05:23.079 --> 00:05:23.649
I'll say that.
00:05:23.649 --> 00:05:26.829
I'm, at this point, I think it's a real thing.
00:05:27.069 --> 00:05:32.918
The amount of time it saves a professional is immense and I'm, I'm naturally skeptic, right?
00:05:33.098 --> 00:05:35.649
You know, I come with decades of experience, right?
00:05:35.649 --> 00:05:38.619
Generally, I'm resistant to these types of things.
00:05:38.889 --> 00:05:41.019
I'm definitely not calling this a fad, right?
00:05:41.019 --> 00:05:41.379
I'm calling.
00:05:41.408 --> 00:05:47.079
It's just that the time saved to bring applications to life is, immense.
00:05:47.468 --> 00:05:48.339
Not a fad.
00:05:49.000 --> 00:05:59.199
That's notable coming from Greg, who's been working on generative AI systems since the 1990s, back when the term meant something completely different.
00:05:59.860 --> 00:06:01.538
But Greg draws a sharp line.
00:06:02.228 --> 00:06:17.139
There's a difference between a professional who uses AI to code faster, someone who could write the code themselves, and a vibe coder, someone who has no engineering background, no ability to debug, no.
00:06:17.139 --> 00:06:19.598
Plan to support what they've built.
00:06:20.029 --> 00:06:29.569
Coding it explicitly means that the individual who's asking for the app is somebody who has not experienced software development, right?
00:06:29.569 --> 00:06:37.309
Somebody who's not prepared to operationalize, somebody who's not prepared to necessarily run or support the application that they're building.
00:06:37.999 --> 00:06:44.538
Sean Larkin back at Softchoice sees this distinction reshaping what work actually looks like.
00:06:45.139 --> 00:06:46.668
The shift isn't about writing.
00:06:46.788 --> 00:06:47.238
Code.
00:06:47.658 --> 00:06:50.029
It's about what kind of worker you need to be.
00:06:50.538 --> 00:06:58.968
The whole thinking from valuing what a knowledge worker does and having specific knowledge 'cause that's out there in the LLM to creatives.
00:06:59.089 --> 00:07:01.999
How am I going to now use all of these tools?
00:07:02.064 --> 00:07:06.113
To create something new and different, be a category creator and create brand new things.
00:07:06.204 --> 00:07:10.103
This is really me because this is where human creativity comes into the picture.
00:07:10.464 --> 00:07:20.814
Knowledge workers become creatives and for Ron Espinsa at Softchoice, all that creative speed changes the math on how companies test ideas.
00:07:21.113 --> 00:07:23.423
The old expression, the king is dead long live.
00:07:23.423 --> 00:07:24.983
The king applies here.
00:07:24.983 --> 00:07:28.858
The pilot is dead long live the pilot, just the notion of how we think about it is gone.
00:07:29.449 --> 00:07:38.689
We can roll out 20 pilots, 50 pilots, a hundred pilots in the time that it took to do one in the past, and we can adjust on the fly really quickly, which makes it no longer a pilot.
00:07:38.689 --> 00:07:39.920
It makes it production ready.
00:07:42.769 --> 00:07:50.839
The pilot is dead long, live the pilot, but even Ron, the non coder who built an app in the afternoon.
00:07:51.425 --> 00:07:52.415
Catches himself,
00:07:52.714 --> 00:08:02.915
and I don't wanna suggest that vibe coding takes the place of, of real production ready, but it allows you to do that rapid prototype in such a way that you can now say, okay, I'm ready to go code.
00:08:03.245 --> 00:08:04.444
Ready to go code.
00:08:05.074 --> 00:08:05.855
That's the pivot.
00:08:06.214 --> 00:08:14.194
Because what happens after the prototype, after the demo, after somebody says, this works, let's just ship it.
00:08:14.735 --> 00:08:17.134
That's where the story gets complicated.
00:08:21.274 --> 00:08:21.964
Act two.
00:08:22.593 --> 00:08:23.404
The blind spot.
00:08:25.173 --> 00:08:27.875
Here's what happens when the vibe meets reality.
00:08:28.384 --> 00:08:39.693
According to industry estimates, roughly half of all proofs of concepts never make it past the prototype stage, and the ones that do, Greg Whalen sees a pattern.
00:08:40.144 --> 00:08:49.024
I think there's a lot of usage of it, which means a majority, you can assume a majority are in fact using it or experimenting with it, but are they using it?
00:08:49.125 --> 00:08:54.254
For production, critical types of stuff, and that's where we find a lot of breakdown.
00:08:54.553 --> 00:08:58.154
A majority experimenting a minority in production.
00:08:58.724 --> 00:09:03.914
And the teams that do make it to production, Greg says they share one thing in common.
00:09:04.365 --> 00:09:08.984
The individuals who are getting to production and able to productionize, right?
00:09:08.990 --> 00:09:18.443
The, typically the ones who have figured out how am I going to observe my system, and then like, what actions am I going to take, in response to things that I observe, right?
00:09:18.504 --> 00:09:21.563
They're, they're thinking more productively from the ground up.
00:09:21.923 --> 00:09:24.683
They thought about it before they started building it.
00:09:25.193 --> 00:09:26.183
Most teams don't.
00:09:26.604 --> 00:09:32.033
Most teams get excited by the speed, by the result that comes so fast.
00:09:32.094 --> 00:09:37.524
They feel like magic and defer the hard part, and that brings us.
00:09:38.174 --> 00:09:38.804
To Amazon.
00:09:40.544 --> 00:09:47.174
We told you about Amazon's AI coding situation at the top, but here's the part that sticks with Greg.
00:09:47.594 --> 00:09:51.344
It's not that these outages happened at some sloppy startup.
00:09:51.614 --> 00:09:56.684
They happened at one of the most disciplined engineering organizations on the planet.
00:09:57.504 --> 00:10:03.684
new technology comes along and just does something different enough that it gets through all of that stuff, right?
00:10:03.684 --> 00:10:11.004
All of the decades of IP work and tooling that have gone in, evades all of them and causes an outage, right?
00:10:11.004 --> 00:10:15.445
I mean, that just demonstrates just how different the approach to quality needs to be.
00:10:15.504 --> 00:10:17.815
The AWS incident disproves that, right?
00:10:17.815 --> 00:10:23.634
Because all aware of how disciplined that organization is, right, and stuff that evades.
00:10:23.684 --> 00:10:30.794
So many layers of controls and protections and right best practices and just causes an issue, right?
00:10:30.945 --> 00:10:37.215
What it tells me is the activities required to properly deal with ai, software quality is different.
00:10:37.725 --> 00:10:40.529
Different, that's the word Greg keeps coming back to.
00:10:41.575 --> 00:10:44.725
Not worse, not better, different.
00:10:45.504 --> 00:10:46.315
And here's why.
00:10:47.095 --> 00:10:59.875
For the last decade or so, software engineers have lived in a world where observability the ability to see what your system is doing and diagnose when something breaks has been essentially solved.
00:11:00.384 --> 00:11:04.585
Cloud platforms hand you logging metrics and alarms.
00:11:04.825 --> 00:11:06.325
You don't even have to think about it
00:11:07.674 --> 00:11:09.355
in the last, decade or so.
00:11:09.534 --> 00:11:13.105
It's turned into like, this is not really a thing we think about anymore.
00:11:13.105 --> 00:11:13.284
Right?
00:11:13.284 --> 00:11:14.995
Of course you're gonna have log ingestion.
00:11:14.995 --> 00:11:17.304
Of course you're gonna have all of this at your fingertips.
00:11:17.424 --> 00:11:19.044
We've sort of lost that muscle.
00:11:19.315 --> 00:11:21.289
AI telemetry though, looks different.
00:11:21.559 --> 00:11:22.490
It's a different animal.
00:11:23.125 --> 00:11:24.985
You can't approach it the same way, right?
00:11:24.985 --> 00:11:31.865
If you're expecting just to throw it into your, log storage system and detect stuff you're in for a world of hurt,
00:11:32.495 --> 00:11:38.195
they've lost that muscle because AI generated code doesn't behave like human written code.
00:11:38.615 --> 00:11:41.554
It's what engineers call non-deterministic.
00:11:42.095 --> 00:11:45.934
The same input can produce different outputs at different times.
00:11:46.414 --> 00:11:51.215
You can't replay a failed transaction and watch it failed the same way twice.
00:11:51.965 --> 00:11:54.875
And Greg has a story that makes this concrete.
00:11:55.414 --> 00:12:02.565
Let's say that, a bunch of people have changed and updated the systems that the AI is using to retrieve, let's say a travel policy.
00:12:02.745 --> 00:12:07.394
And one, day somebody makes a pretty trivial change to your document repository.
00:12:07.720 --> 00:12:13.419
And then all of a sudden your AI starts retrieving a, travel policy that's three years old instead of the one that's from last year.
00:12:13.629 --> 00:12:18.460
There is a bug and a bunch of people saying, no, there's not a bug, because nothing changed and nothing's wrong.
00:12:18.639 --> 00:12:22.720
And then whoever's responsible for the AI is saying, well, the outcome is bad.
00:12:24.054 --> 00:12:26.184
Half the team says nothing changed.
00:12:26.575 --> 00:12:29.304
The AI team says the output is wrong.
00:12:29.995 --> 00:12:39.174
Nobody can prove who's right because nobody has the tools to trace how a trivial upstream change cascaded into a bad outcome.
00:12:39.654 --> 00:12:51.024
And this, this gap is where Greg lands on what he considers the fundamental difference between AI generated software and everything that came before it.
00:12:51.355 --> 00:12:53.004
Traditional software, it doesn't work like that.
00:12:53.254 --> 00:12:55.985
that 80% of the work to get that productivity game.
00:12:56.315 --> 00:13:02.794
With AI systems, you get the productivity only having done 20% of the work, and that's what's driving people up the wall.
00:13:03.065 --> 00:13:04.774
That's the fundamentally different thing.
00:13:05.524 --> 00:13:10.804
With traditional software, you do 80% of the work before you see any results.
00:13:11.315 --> 00:13:13.684
With ai, you see results.
00:13:13.764 --> 00:13:17.875
After 20%, the prototype looks finished.
00:13:18.174 --> 00:13:30.315
It feels finished, but the observability, the governance, the debugging, the edge cases, none of that's been done, and now nobody wants to do it.
00:13:31.159 --> 00:13:33.440
Because the finish line is right there.
00:13:34.009 --> 00:13:38.299
It's this sort of like, you already got 80% for free and now you have right the last mile.
00:13:38.389 --> 00:13:39.919
You're almost at that finish line.
00:13:40.039 --> 00:13:41.929
It's like the last mile of the marathon.
00:13:41.960 --> 00:13:44.960
Marathon is like up this, huge incline, right?
00:13:44.960 --> 00:13:50.629
So you've made it and you just see it over there, and then right around here is a little path that goes around the mouth and you're just like.
00:13:51.205 --> 00:13:52.945
Sure you don't wanna take the shortcut, right?
00:13:53.365 --> 00:13:54.955
You're almost at the end of the marathon.
00:13:54.955 --> 00:13:56.455
You could just sweep it under the rug.
00:13:56.455 --> 00:13:57.654
Probably nobody will know.
00:13:58.345 --> 00:14:08.335
And there's another problem, even when engineers try to do the right thing and review AI generated code before it goes out, the code itself fights them.
00:14:08.830 --> 00:14:14.139
Are people really gonna read a bunch of code that is more difficult to read than hand coded stuff?
00:14:14.289 --> 00:14:15.580
And that's a key problem, right?
00:14:15.580 --> 00:14:18.700
Because you're like, the engineers know they need to review every line.
00:14:18.970 --> 00:14:20.559
They know they should, but they don't.
00:14:23.259 --> 00:14:28.840
The code is harder to read, the incentive to skip the review is enormous.
00:14:29.289 --> 00:14:31.600
And the result in Greg's word.
00:14:32.245 --> 00:14:33.085
Is a snowball.
00:14:33.654 --> 00:14:37.404
All of these little bad micro decisions just snowball into something.
00:14:37.404 --> 00:14:38.904
We're like, Hey, congratulations.
00:14:38.904 --> 00:14:41.095
Now we have a system that is dangerous.
00:14:41.095 --> 00:14:42.534
It's unmanageable, right?
00:14:42.625 --> 00:14:47.274
And yet everybody's using it, or yet a bunch of people are in fact using it and say it's useful.
00:14:50.034 --> 00:14:54.865
And Sean Larkin back at Softchoice puts the problem in different terms.
00:14:55.615 --> 00:15:03.565
Vibe coding created something wonderful, a new class of builders, but it didn't create the other thing those builders need.
00:15:04.345 --> 00:15:10.495
What I'm worried about with Vibe Coding specifically is it's great that it created a lot of citizen developers, right?
00:15:10.495 --> 00:15:11.365
That's a wonderful thing.
00:15:11.965 --> 00:15:13.554
What it didn't do or doesn't do.
00:15:14.080 --> 00:15:20.889
Effectively is create a bunch of citizen product managers that understand how to take a product to market.
00:15:22.029 --> 00:15:25.990
Citizen developers, not citizen product managers.
00:15:26.679 --> 00:15:27.850
That's the blind spot.
00:15:28.720 --> 00:15:30.730
So what do you do about it?
00:15:39.610 --> 00:15:41.830
Here's something we keep hearing from our customers.
00:15:42.159 --> 00:15:48.159
Cloud storage for backups has gotten really complicated, expensive, and risky all at the same time.
00:15:48.309 --> 00:15:54.669
You are trying to protect your backup data from ransomware, but architecting a repository that's truly secure, it takes expertise.
00:15:54.669 --> 00:15:56.590
Most teams just don't have on staff.
00:15:56.649 --> 00:16:04.299
And even when you find a solution, cloud storage costs spiral in ways that feel impossible to predict with store charges, address fees.
00:16:04.570 --> 00:16:08.230
API calls the bill shows up and it's nothing like what you budgeted.
00:16:08.649 --> 00:16:09.490
That's the problem.
00:16:09.490 --> 00:16:11.259
Veeam Data Cloud Vault was built to solve.
00:16:12.279 --> 00:16:18.429
Veeam Data Cloud Vault is a fully managed, secure cloud storage resource designed specifically for backup.
00:16:18.700 --> 00:16:21.279
You tell them how much you need, they handle the rest.
00:16:21.460 --> 00:16:25.629
No configuration, no infrastructure management, and no guessing on costs.
00:16:25.809 --> 00:16:32.440
Vault runs on zero trust architecture, always imputable, always encrypted, and always air gap from production.
00:16:32.649 --> 00:16:36.220
And it prices on a simple, flat per terabyte model that includes API.
00:16:36.220 --> 00:16:38.500
Calls Restore and egress.
00:16:38.740 --> 00:16:39.789
No bill shock.
00:16:40.090 --> 00:16:40.990
It's worth noting.
00:16:41.330 --> 00:16:46.759
82% of organizations say they're struggling with cloud spend and security all at the same time.
00:16:46.970 --> 00:16:48.500
Veeam Vault addresses both.
00:16:48.710 --> 00:16:56.750
If you're ready to stop guessing and start protecting, click the link in the description to sign up for Veeam's AWS demo series and see Vault in Action
00:16:59.570 --> 00:17:01.789
Act three, the guardrails.
00:17:03.129 --> 00:17:06.759
Okay, so we've established the excitement and the risk.
00:17:07.299 --> 00:17:16.960
The question for an IT leader listening to this, someone running a team at a mid-size company, fielding requests from the business side to move faster.
00:17:17.380 --> 00:17:23.319
Watching developers adopt AI tools, whether it's sanctioned or not, is practical.
00:17:23.920 --> 00:17:25.269
Where do I say yes?
00:17:25.964 --> 00:17:27.045
Where do I say?
00:17:27.045 --> 00:17:27.644
Not yet.
00:17:28.244 --> 00:17:29.085
How do I start?
00:17:29.505 --> 00:17:35.444
When our producers asked Sean Larkin about this, he went straight to the mistake he sees most often.
00:17:35.775 --> 00:17:45.075
We talked about the progression from ideation all the way through to production and how a lot of people skip the critical step of making sure you're building the right thing.
00:17:45.345 --> 00:17:51.045
'cause we don't wanna create a perfect solution to a problem that doesn't exist or that nobody wants to really use.
00:17:51.069 --> 00:17:55.180
Whatever it is you're building, so you wanna make sure you're building the right it before you build it.
00:17:55.180 --> 00:17:55.599
Right.
00:17:55.960 --> 00:18:00.130
Build the right it before you build it.
00:18:00.339 --> 00:18:00.670
Right.
00:18:01.029 --> 00:18:09.789
Softchoice has developed what they call an executive alignment session or EAS, designed to help organizations work through exactly this.
00:18:10.930 --> 00:18:14.259
Not the technology, the strategy underneath it.
00:18:14.740 --> 00:18:18.519
The why before the what Ron Espinsa has seen.
00:18:18.519 --> 00:18:19.720
The difference it makes
00:18:20.170 --> 00:18:24.460
the EAS is such a big deal is taking that time to measure twice, cut once.
00:18:24.670 --> 00:18:28.329
So let's measure twice before we do any cutting and make sure that's the cut we wanna make.
00:18:28.599 --> 00:18:33.640
'cause AI allows you to make lots of cuts and it could be one that has far reaching implications.
00:18:34.029 --> 00:18:37.835
Measure it twice, cut once and from the technical side.
00:18:38.545 --> 00:18:43.855
Greg Whalen says, the organizations that actually make it to production share a discipline.
00:18:44.214 --> 00:18:48.355
They define what good looks like before they start building it.
00:18:49.345 --> 00:18:57.414
If I'm only gonna realistically look at five things in the morning, like what are those things that I'm gonna look at that indicate healthy, not healthy or something else?
00:18:57.414 --> 00:19:00.565
People having a problem defining what those things are.
00:19:00.644 --> 00:19:02.444
Right, because they haven't fully thought through what?
00:19:02.444 --> 00:19:03.434
They're really gonna build
00:19:03.825 --> 00:19:05.384
five things in the morning.
00:19:05.775 --> 00:19:07.125
That's Greg's test.
00:19:07.605 --> 00:19:17.865
Before you build anything with ai, before a single line of code gets generated, can you name the five metrics you check each morning to know if it was working?
00:19:18.285 --> 00:19:23.263
If you can't answer that question, you are not ready for production and that's okay.
00:19:23.414 --> 00:19:25.184
That's what experimentation is for.
00:19:25.809 --> 00:19:34.390
Vibe coating is awesome for the rapid prototype phases where you're testing out an idea and you don't have to have perfect data, perfect authentication, perfect.
00:19:34.569 --> 00:19:35.349
Anything really.
00:19:35.539 --> 00:19:38.450
can do a lot without those elements being in place.
00:19:38.779 --> 00:19:39.769
It does have a home.
00:19:39.859 --> 00:19:41.150
It's just not in production.
00:19:41.200 --> 00:19:42.640
I would say it's great for imagineers,
00:19:43.240 --> 00:19:50.950
it's great for imagineers and Ron's advice to organizations still on the fence about getting started is blunt.
00:19:51.369 --> 00:19:55.059
If you're waiting for the dust to settle, you're just gonna end up as part of that dust.
00:19:55.240 --> 00:19:58.119
I think companies and people have to start getting themselves dirty.
00:19:58.180 --> 00:19:59.710
They have to start putting their hands in it.
00:19:59.710 --> 00:20:00.789
You have to start trying it.
00:20:00.940 --> 00:20:02.710
Get your hands dirty.
00:20:03.160 --> 00:20:04.150
Experiment.
00:20:04.539 --> 00:20:09.009
Build the baseball app, but know the line between the prototype.
00:20:09.339 --> 00:20:11.950
And production and invest in the parts.
00:20:11.950 --> 00:20:20.140
Nobody wants to do the observability, the testing, the governance, because that's where the actual value compounds.
00:20:20.740 --> 00:20:31.000
Greg's view is that Amazon's current fix, having senior engineers review every line of AI generated code works for now, but it won't scale.
00:20:31.240 --> 00:20:33.069
AWS's solution is practical.
00:20:33.190 --> 00:20:33.910
Hey guys.
00:20:34.349 --> 00:20:38.339
Everybody has to review the code that's practical for today.
00:20:38.640 --> 00:20:41.220
Prove AI looks at that problem and says, that's not realistic.
00:20:41.220 --> 00:20:42.630
Nobody's gonna do that, right?
00:20:42.630 --> 00:20:43.950
Let's look in five years.
00:20:43.950 --> 00:20:46.319
Do you think people are gonna like, that's even worse?
00:20:46.500 --> 00:20:50.579
The answer, Greg believes, isn't more human review.
00:20:51.240 --> 00:20:59.400
It's better AI observability systems that can watch what AI builds and flag problems before they reach production.
00:21:00.119 --> 00:21:02.400
That's what prove AI is building.
00:21:02.875 --> 00:21:05.460
And it's a space that barely existed a year ago.
00:21:06.115 --> 00:21:12.234
But today, for the IT leader listening to this right now, the playbook starts with something simpler.
00:21:12.954 --> 00:21:15.984
Something Ron's 14-year-old put into words,
00:21:16.075 --> 00:21:21.565
I'm gonna quote my actual 14-year-old son, and he said, I love what you're doing with ai.
00:21:21.565 --> 00:21:26.275
I love what I can do with ai, but AI can't conceive of something new on its own.
00:21:27.954 --> 00:21:30.625
It can't conceive of something new on its own.
00:21:31.164 --> 00:21:33.115
The creativity, the judgment.
00:21:33.500 --> 00:21:38.539
The decision about what to build and why and whether it's safe, that's all yours.
00:21:38.869 --> 00:21:41.390
AI just made the building part faster.
00:21:42.109 --> 00:21:45.769
Your job is to make sure the thinking doesn't get skipped.
00:21:49.855 --> 00:21:52.974
Here's what I keep coming back to from these conversations.
00:21:53.755 --> 00:21:55.315
Vibe coding is real.
00:21:55.674 --> 00:21:56.694
It's not a fad.
00:21:56.994 --> 00:22:01.795
Even self-described skeptics with decades of experience are calling this genuine.
00:22:02.394 --> 00:22:04.134
The speed is extraordinary.
00:22:04.585 --> 00:22:07.134
The creativity it unlocks is extraordinary.
00:22:07.845 --> 00:22:11.924
Ron built a working app in an afternoon without writing a line of code.
00:22:12.585 --> 00:22:16.335
Sean sees an entire generation of new builders emerging.
00:22:17.984 --> 00:22:23.595
The pilot is dead long, live the pilot, but the speed is also the risk.
00:22:23.924 --> 00:22:29.714
AI gives you 80% of the result after 20% of the work and the remaining 20%.
00:22:30.319 --> 00:22:36.680
The observability, the testing, the governance, the question of what success even looks like.
00:22:37.309 --> 00:22:40.849
That's the part that separates a demo from a production system.
00:22:41.059 --> 00:22:43.339
It's the part everyone wants to skip.
00:22:44.029 --> 00:22:46.910
It's the marathon shortcut that Greg described.
00:22:47.089 --> 00:22:54.470
So if you are an IT leader listening to this and thinking, how do I let my team experiment without losing control?
00:22:54.619 --> 00:22:57.670
How do I say yes to the speed without saying yes to the risk?
00:22:58.500 --> 00:23:00.420
Soft choice can help you figure this out.
00:23:01.200 --> 00:23:18.990
Their executive alignment sessions are designed to help organizations define the why before the what to make sure you're building the right thing before you build it right, and you're developing a lighter AI powered version that you'll be able to access on your own terms.
00:23:19.470 --> 00:23:24.869
Visit softchoice.com/e to learn more.
00:23:25.650 --> 00:23:27.779
That's softchoice.com.
00:23:28.224 --> 00:23:30.535
Forward slash EAS.
00:23:33.625 --> 00:23:38.875
The catalyst was reported and produced by Tobin Dalrimple and the team at Pilgrim.
00:23:38.875 --> 00:23:41.394
Content Editing by Ryan Clark.
00:23:41.515 --> 00:23:46.795
With support from Philippe Demas, Joseph Beyer, and the marketing team at Softchoice.
00:23:47.125 --> 00:23:54.654
Special thanks to Greg Whalen, Ron Espinsa and Sean Larkin for sharing their expertise and insights.
00:24:00.369 --> 00:24:03.730
Thanks again to Veeam Data Cloud Vault for supporting today's episode.
00:24:04.029 --> 00:24:09.789
If unpredictable cloud storage costs and backup security keep showing up on your radar, this one's worth a look.
00:24:09.849 --> 00:24:13.240
Click the link in the description to sign up for their AWS demo.
00:00:02.399 --> 00:00:12.689
This episode of The Catalyst is brought to you by Veeam Data Cloud backup, fully managed secure cloud storage for your backups with no surprise bills and zero configuration headaches.
00:00:12.960 --> 00:00:17.670
Click the link in the description to sign up for their AWS demo series and see it in action.
00:00:20.519 --> 00:00:23.789
Here's a story about building something from nothing.
00:00:25.679 --> 00:00:27.929
When I first heard the term, I'm like, I'm gonna test this out.
00:00:27.929 --> 00:00:30.399
And I, have, I run a baseball org, as you've heard me talk about.
00:00:30.638 --> 00:00:35.500
I was like, you know what would be cool is if I had an online tool that I could get my guys to use, what might that look like?
00:00:35.500 --> 00:00:38.918
So I actually went in, I started asking Gemini, how might I do this?
00:00:39.399 --> 00:00:43.088
And I gave it some scenarios and different things I wanted to learn about, and it just.
00:00:43.134 --> 00:00:44.695
Started popping out this app for me.
00:00:44.695 --> 00:00:45.564
Do you mean like this?
00:00:45.564 --> 00:00:46.585
And I was like, oh, wow.
00:00:46.823 --> 00:00:47.723
I was blown outta my seat.
00:00:47.723 --> 00:00:48.323
I'm like, okay.
00:00:48.924 --> 00:00:50.344
That's Ron Espinsa.
00:00:50.814 --> 00:00:53.155
He leads Softchoice Google category.
00:00:53.545 --> 00:00:55.973
He's not a developer, he's not a coder.
00:00:56.454 --> 00:01:02.814
But in a single afternoon he built a custom baseball analytics app player profiles.
00:01:03.250 --> 00:01:11.379
Speed ratings, defensive positioning, game scenarios, all by describing what he wanted in plain English.
00:01:11.799 --> 00:01:12.760
I don't like these colors.
00:01:12.760 --> 00:01:13.269
Change that.
00:01:13.299 --> 00:01:13.719
Okay?
00:01:14.019 --> 00:01:15.849
And it went and changed all the color schemes.
00:01:16.179 --> 00:01:20.079
What if I could factor in instead of just offense and defense?
00:01:20.290 --> 00:01:24.040
What if I could put in player profile and it recreated the app with the player profile?
00:01:24.219 --> 00:01:26.500
Next thing I know, I had the ability to put in a player's.
00:01:27.125 --> 00:01:34.444
Name, size in terms of height and weight, their speed ratings, how they perform offensively, how they perform defensively.
00:01:34.745 --> 00:01:36.334
And then I could say, where's the play?
00:01:36.575 --> 00:01:38.885
And there was no code that I generated.
00:01:39.694 --> 00:01:42.275
Ron's baseball app is a weekend project.
00:01:42.635 --> 00:01:55.055
It's fun, it's harmless, but that same technology, AI that writes code from conversation is now inside enterprise software teams at companies of every size.
00:01:55.810 --> 00:02:02.680
And in December of 2025, it potentially caused a problem that was anything but harmless.
00:02:04.120 --> 00:02:14.020
According to the Financial Times, Amazon's own AI coding tool, an agent called Kiro was given production level access by an engineer.
00:02:14.379 --> 00:02:21.400
The AI autonomously decided to fix a bug by deleting and recreating an entire live environment.
00:02:21.939 --> 00:02:35.205
The result, a 13 hour AWS outage by March, 2026, more AI assisted deployment failures reportedly cost Amazon millions of lost orders.
00:02:36.145 --> 00:02:45.835
The same executive who pushed engineers to use AI coding tools was forced to mandate senior engineers sign off on every AI assisted code change.
00:02:46.254 --> 00:02:48.775
Amazon called it controlled friction.
00:02:55.044 --> 00:02:58.284
From Softchoice, a worldwide technology company.
00:02:58.645 --> 00:03:00.384
This is the catalyst.
00:03:00.895 --> 00:03:01.594
I am Katey Teekasingh.
00:03:02.664 --> 00:03:05.335
This season, we're doing things a bit differently.
00:03:05.694 --> 00:03:14.305
We're making audio documentaries, real stories from the front lines of it, exploring the challenges of small teams chasing big dreams.
00:03:14.514 --> 00:03:17.224
Today's episode a technology that lets.
00:03:17.460 --> 00:03:26.159
Anyone builds software by describing it in plain English, has gone from a weekend hobby to the dominant force in how code gets written.
00:03:26.340 --> 00:03:28.620
It's already inside your organization.
00:03:29.009 --> 00:03:32.490
The question is whether anyone's watching what it's building.
00:03:32.879 --> 00:03:40.199
We are calling it the Vibe Coding episode, act one, the vibe.
00:03:42.580 --> 00:03:52.030
In February, 2025, a former open AI researcher named Andre Cari posted about a new way he'd been building software.
00:03:52.569 --> 00:04:03.639
He described it as fully giving into the vibes, telling an AI what he wanted, accepting whatever code it produced, and barely looking at the results.
00:04:04.099 --> 00:04:07.579
He called it vibe coating within months.
00:04:08.090 --> 00:04:17.389
Collin's Dictionary named it their 2025 Word of the Year when our producers sat down with Sean Larkin, AI principal architect at Softchoice.
00:04:19.120 --> 00:04:20.019
He put it simply,
00:04:20.439 --> 00:04:28.810
the way I understand vibe coating, it was, there's a term that was coined by Andre Carte who said, I developed this application over the weekend.
00:04:28.810 --> 00:04:34.000
And so he coined the phrase vibe coating to mean something that's a lower level effort.
00:04:34.180 --> 00:04:35.860
Typically done over the weekend.
00:04:36.218 --> 00:04:38.980
Something you would definitely not put into production.
00:04:39.129 --> 00:04:43.990
It's the sort of experimentation, tooling, rapid prototypes you might build on a weekend.
00:04:44.500 --> 00:04:46.694
Weekend projects, experiments.
00:04:47.678 --> 00:04:49.718
Things you'd never put into production.
00:04:50.468 --> 00:04:51.728
That was the original idea.
00:04:52.629 --> 00:04:54.939
But vibe coding didn't stay on the weekend.
00:04:56.348 --> 00:05:03.879
According to industry surveys, over 90% of US developers now use AI coding tools daily.
00:05:04.209 --> 00:05:08.858
Roughly 40% of all codes written globally are now AI generated.
00:05:09.399 --> 00:05:17.048
When our producers asked Greg Whalen, CTO of prove ai, a company building observability tools.
00:05:17.288 --> 00:05:22.629
Specifically for AI systems, whether this felt like a real shift, he didn't hedge
00:05:23.079 --> 00:05:23.649
I'll say that.
00:05:23.649 --> 00:05:26.829
I'm, at this point, I think it's a real thing.
00:05:27.069 --> 00:05:32.918
The amount of time it saves a professional is immense and I'm, I'm naturally skeptic, right?
00:05:33.098 --> 00:05:35.649
You know, I come with decades of experience, right?
00:05:35.649 --> 00:05:38.619
Generally, I'm resistant to these types of things.
00:05:38.889 --> 00:05:41.019
I'm definitely not calling this a fad, right?
00:05:41.019 --> 00:05:41.379
I'm calling.
00:05:41.408 --> 00:05:47.079
It's just that the time saved to bring applications to life is, immense.
00:05:47.468 --> 00:05:48.339
Not a fad.
00:05:49.000 --> 00:05:59.199
That's notable coming from Greg, who's been working on generative AI systems since the 1990s, back when the term meant something completely different.
00:05:59.860 --> 00:06:01.538
But Greg draws a sharp line.
00:06:02.228 --> 00:06:17.139
There's a difference between a professional who uses AI to code faster, someone who could write the code themselves, and a vibe coder, someone who has no engineering background, no ability to debug, no.
00:06:17.139 --> 00:06:19.598
Plan to support what they've built.
00:06:20.029 --> 00:06:29.569
Coding it explicitly means that the individual who's asking for the app is somebody who has not experienced software development, right?
00:06:29.569 --> 00:06:37.309
Somebody who's not prepared to operationalize, somebody who's not prepared to necessarily run or support the application that they're building.
00:06:37.999 --> 00:06:44.538
Sean Larkin back at Softchoice sees this distinction reshaping what work actually looks like.
00:06:45.139 --> 00:06:46.668
The shift isn't about writing.
00:06:46.788 --> 00:06:47.238
Code.
00:06:47.658 --> 00:06:50.029
It's about what kind of worker you need to be.
00:06:50.538 --> 00:06:58.968
The whole thinking from valuing what a knowledge worker does and having specific knowledge 'cause that's out there in the LLM to creatives.
00:06:59.089 --> 00:07:01.999
How am I going to now use all of these tools?
00:07:02.064 --> 00:07:06.113
To create something new and different, be a category creator and create brand new things.
00:07:06.204 --> 00:07:10.103
This is really me because this is where human creativity comes into the picture.
00:07:10.464 --> 00:07:20.814
Knowledge workers become creatives and for Ron Espinsa at Softchoice, all that creative speed changes the math on how companies test ideas.
00:07:21.113 --> 00:07:23.423
The old expression, the king is dead long live.
00:07:23.423 --> 00:07:24.983
The king applies here.
00:07:24.983 --> 00:07:28.858
The pilot is dead long live the pilot, just the notion of how we think about it is gone.
00:07:29.449 --> 00:07:38.689
We can roll out 20 pilots, 50 pilots, a hundred pilots in the time that it took to do one in the past, and we can adjust on the fly really quickly, which makes it no longer a pilot.
00:07:38.689 --> 00:07:39.920
It makes it production ready.
00:07:42.769 --> 00:07:50.839
The pilot is dead long, live the pilot, but even Ron, the non coder who built an app in the afternoon.
00:07:51.425 --> 00:07:52.415
Catches himself,
00:07:52.714 --> 00:08:02.915
and I don't wanna suggest that vibe coding takes the place of, of real production ready, but it allows you to do that rapid prototype in such a way that you can now say, okay, I'm ready to go code.
00:08:03.245 --> 00:08:04.444
Ready to go code.
00:08:05.074 --> 00:08:05.855
That's the pivot.
00:08:06.214 --> 00:08:14.194
Because what happens after the prototype, after the demo, after somebody says, this works, let's just ship it.
00:08:14.735 --> 00:08:17.134
That's where the story gets complicated.
00:08:21.274 --> 00:08:21.964
Act two.
00:08:22.593 --> 00:08:23.404
The blind spot.
00:08:25.173 --> 00:08:27.875
Here's what happens when the vibe meets reality.
00:08:28.384 --> 00:08:39.693
According to industry estimates, roughly half of all proofs of concepts never make it past the prototype stage, and the ones that do, Greg Whalen sees a pattern.
00:08:40.144 --> 00:08:49.024
I think there's a lot of usage of it, which means a majority, you can assume a majority are in fact using it or experimenting with it, but are they using it?
00:08:49.125 --> 00:08:54.254
For production, critical types of stuff, and that's where we find a lot of breakdown.
00:08:54.553 --> 00:08:58.154
A majority experimenting a minority in production.
00:08:58.724 --> 00:09:03.914
And the teams that do make it to production, Greg says they share one thing in common.
00:09:04.365 --> 00:09:08.984
The individuals who are getting to production and able to productionize, right?
00:09:08.990 --> 00:09:18.443
The, typically the ones who have figured out how am I going to observe my system, and then like, what actions am I going to take, in response to things that I observe, right?
00:09:18.504 --> 00:09:21.563
They're, they're thinking more productively from the ground up.
00:09:21.923 --> 00:09:24.683
They thought about it before they started building it.
00:09:25.193 --> 00:09:26.183
Most teams don't.
00:09:26.604 --> 00:09:32.033
Most teams get excited by the speed, by the result that comes so fast.
00:09:32.094 --> 00:09:37.524
They feel like magic and defer the hard part, and that brings us.
00:09:38.174 --> 00:09:38.804
To Amazon.
00:09:40.544 --> 00:09:47.174
We told you about Amazon's AI coding situation at the top, but here's the part that sticks with Greg.
00:09:47.594 --> 00:09:51.344
It's not that these outages happened at some sloppy startup.
00:09:51.614 --> 00:09:56.684
They happened at one of the most disciplined engineering organizations on the planet.
00:09:57.504 --> 00:10:03.684
new technology comes along and just does something different enough that it gets through all of that stuff, right?
00:10:03.684 --> 00:10:11.004
All of the decades of IP work and tooling that have gone in, evades all of them and causes an outage, right?
00:10:11.004 --> 00:10:15.445
I mean, that just demonstrates just how different the approach to quality needs to be.
00:10:15.504 --> 00:10:17.815
The AWS incident disproves that, right?
00:10:17.815 --> 00:10:23.634
Because all aware of how disciplined that organization is, right, and stuff that evades.
00:10:23.684 --> 00:10:30.794
So many layers of controls and protections and right best practices and just causes an issue, right?
00:10:30.945 --> 00:10:37.215
What it tells me is the activities required to properly deal with ai, software quality is different.
00:10:37.725 --> 00:10:40.529
Different, that's the word Greg keeps coming back to.
00:10:41.575 --> 00:10:44.725
Not worse, not better, different.
00:10:45.504 --> 00:10:46.315
And here's why.
00:10:47.095 --> 00:10:59.875
For the last decade or so, software engineers have lived in a world where observability the ability to see what your system is doing and diagnose when something breaks has been essentially solved.
00:11:00.384 --> 00:11:04.585
Cloud platforms hand you logging metrics and alarms.
00:11:04.825 --> 00:11:06.325
You don't even have to think about it
00:11:07.674 --> 00:11:09.355
in the last, decade or so.
00:11:09.534 --> 00:11:13.105
It's turned into like, this is not really a thing we think about anymore.
00:11:13.105 --> 00:11:13.284
Right?
00:11:13.284 --> 00:11:14.995
Of course you're gonna have log ingestion.
00:11:14.995 --> 00:11:17.304
Of course you're gonna have all of this at your fingertips.
00:11:17.424 --> 00:11:19.044
We've sort of lost that muscle.
00:11:19.315 --> 00:11:21.289
AI telemetry though, looks different.
00:11:21.559 --> 00:11:22.490
It's a different animal.
00:11:23.125 --> 00:11:24.985
You can't approach it the same way, right?
00:11:24.985 --> 00:11:31.865
If you're expecting just to throw it into your, log storage system and detect stuff you're in for a world of hurt,
00:11:32.495 --> 00:11:38.195
they've lost that muscle because AI generated code doesn't behave like human written code.
00:11:38.615 --> 00:11:41.554
It's what engineers call non-deterministic.
00:11:42.095 --> 00:11:45.934
The same input can produce different outputs at different times.
00:11:46.414 --> 00:11:51.215
You can't replay a failed transaction and watch it failed the same way twice.
00:11:51.965 --> 00:11:54.875
And Greg has a story that makes this concrete.
00:11:55.414 --> 00:12:02.565
Let's say that, a bunch of people have changed and updated the systems that the AI is using to retrieve, let's say a travel policy.
00:12:02.745 --> 00:12:07.394
And one, day somebody makes a pretty trivial change to your document repository.
00:12:07.720 --> 00:12:13.419
And then all of a sudden your AI starts retrieving a, travel policy that's three years old instead of the one that's from last year.
00:12:13.629 --> 00:12:18.460
There is a bug and a bunch of people saying, no, there's not a bug, because nothing changed and nothing's wrong.
00:12:18.639 --> 00:12:22.720
And then whoever's responsible for the AI is saying, well, the outcome is bad.
00:12:24.054 --> 00:12:26.184
Half the team says nothing changed.
00:12:26.575 --> 00:12:29.304
The AI team says the output is wrong.
00:12:29.995 --> 00:12:39.174
Nobody can prove who's right because nobody has the tools to trace how a trivial upstream change cascaded into a bad outcome.
00:12:39.654 --> 00:12:51.024
And this, this gap is where Greg lands on what he considers the fundamental difference between AI generated software and everything that came before it.
00:12:51.355 --> 00:12:53.004
Traditional software, it doesn't work like that.
00:12:53.254 --> 00:12:55.985
that 80% of the work to get that productivity game.
00:12:56.315 --> 00:13:02.794
With AI systems, you get the productivity only having done 20% of the work, and that's what's driving people up the wall.
00:13:03.065 --> 00:13:04.774
That's the fundamentally different thing.
00:13:05.524 --> 00:13:10.804
With traditional software, you do 80% of the work before you see any results.
00:13:11.315 --> 00:13:13.684
With ai, you see results.
00:13:13.764 --> 00:13:17.875
After 20%, the prototype looks finished.
00:13:18.174 --> 00:13:30.315
It feels finished, but the observability, the governance, the debugging, the edge cases, none of that's been done, and now nobody wants to do it.
00:13:31.159 --> 00:13:33.440
Because the finish line is right there.
00:13:34.009 --> 00:13:38.299
It's this sort of like, you already got 80% for free and now you have right the last mile.
00:13:38.389 --> 00:13:39.919
You're almost at that finish line.
00:13:40.039 --> 00:13:41.929
It's like the last mile of the marathon.
00:13:41.960 --> 00:13:44.960
Marathon is like up this, huge incline, right?
00:13:44.960 --> 00:13:50.629
So you've made it and you just see it over there, and then right around here is a little path that goes around the mouth and you're just like.
00:13:51.205 --> 00:13:52.945
Sure you don't wanna take the shortcut, right?
00:13:53.365 --> 00:13:54.955
You're almost at the end of the marathon.
00:13:54.955 --> 00:13:56.455
You could just sweep it under the rug.
00:13:56.455 --> 00:13:57.654
Probably nobody will know.
00:13:58.345 --> 00:14:08.335
And there's another problem, even when engineers try to do the right thing and review AI generated code before it goes out, the code itself fights them.
00:14:08.830 --> 00:14:14.139
Are people really gonna read a bunch of code that is more difficult to read than hand coded stuff?
00:14:14.289 --> 00:14:15.580
And that's a key problem, right?
00:14:15.580 --> 00:14:18.700
Because you're like, the engineers know they need to review every line.
00:14:18.970 --> 00:14:20.559
They know they should, but they don't.
00:14:23.259 --> 00:14:28.840
The code is harder to read, the incentive to skip the review is enormous.
00:14:29.289 --> 00:14:31.600
And the result in Greg's word.
00:14:32.245 --> 00:14:33.085
Is a snowball.
00:14:33.654 --> 00:14:37.404
All of these little bad micro decisions just snowball into something.
00:14:37.404 --> 00:14:38.904
We're like, Hey, congratulations.
00:14:38.904 --> 00:14:41.095
Now we have a system that is dangerous.
00:14:41.095 --> 00:14:42.534
It's unmanageable, right?
00:14:42.625 --> 00:14:47.274
And yet everybody's using it, or yet a bunch of people are in fact using it and say it's useful.
00:14:50.034 --> 00:14:54.865
And Sean Larkin back at Softchoice puts the problem in different terms.
00:14:55.615 --> 00:15:03.565
Vibe coding created something wonderful, a new class of builders, but it didn't create the other thing those builders need.
00:15:04.345 --> 00:15:10.495
What I'm worried about with Vibe Coding specifically is it's great that it created a lot of citizen developers, right?
00:15:10.495 --> 00:15:11.365
That's a wonderful thing.
00:15:11.965 --> 00:15:13.554
What it didn't do or doesn't do.
00:15:14.080 --> 00:15:20.889
Effectively is create a bunch of citizen product managers that understand how to take a product to market.
00:15:22.029 --> 00:15:25.990
Citizen developers, not citizen product managers.
00:15:26.679 --> 00:15:27.850
That's the blind spot.
00:15:28.720 --> 00:15:30.730
So what do you do about it?
00:15:39.610 --> 00:15:41.830
Here's something we keep hearing from our customers.
00:15:42.159 --> 00:15:48.159
Cloud storage for backups has gotten really complicated, expensive, and risky all at the same time.
00:15:48.309 --> 00:15:54.669
You are trying to protect your backup data from ransomware, but architecting a repository that's truly secure, it takes expertise.
00:15:54.669 --> 00:15:56.590
Most teams just don't have on staff.
00:15:56.649 --> 00:16:04.299
And even when you find a solution, cloud storage costs spiral in ways that feel impossible to predict with store charges, address fees.
00:16:04.570 --> 00:16:08.230
API calls the bill shows up and it's nothing like what you budgeted.
00:16:08.649 --> 00:16:09.490
That's the problem.
00:16:09.490 --> 00:16:11.259
Veeam Data Cloud Vault was built to solve.
00:16:12.279 --> 00:16:18.429
Veeam Data Cloud Vault is a fully managed, secure cloud storage resource designed specifically for backup.
00:16:18.700 --> 00:16:21.279
You tell them how much you need, they handle the rest.
00:16:21.460 --> 00:16:25.629
No configuration, no infrastructure management, and no guessing on costs.
00:16:25.809 --> 00:16:32.440
Vault runs on zero trust architecture, always imputable, always encrypted, and always air gap from production.
00:16:32.649 --> 00:16:36.220
And it prices on a simple, flat per terabyte model that includes API.
00:16:36.220 --> 00:16:38.500
Calls Restore and egress.
00:16:38.740 --> 00:16:39.789
No bill shock.
00:16:40.090 --> 00:16:40.990
It's worth noting.
00:16:41.330 --> 00:16:46.759
82% of organizations say they're struggling with cloud spend and security all at the same time.
00:16:46.970 --> 00:16:48.500
Veeam Vault addresses both.
00:16:48.710 --> 00:16:56.750
If you're ready to stop guessing and start protecting, click the link in the description to sign up for Veeam's AWS demo series and see Vault in Action
00:16:59.570 --> 00:17:01.789
Act three, the guardrails.
00:17:03.129 --> 00:17:06.759
Okay, so we've established the excitement and the risk.
00:17:07.299 --> 00:17:16.960
The question for an IT leader listening to this, someone running a team at a mid-size company, fielding requests from the business side to move faster.
00:17:17.380 --> 00:17:23.319
Watching developers adopt AI tools, whether it's sanctioned or not, is practical.
00:17:23.920 --> 00:17:25.269
Where do I say yes?
00:17:25.964 --> 00:17:27.045
Where do I say?
00:17:27.045 --> 00:17:27.644
Not yet.
00:17:28.244 --> 00:17:29.085
How do I start?
00:17:29.505 --> 00:17:35.444
When our producers asked Sean Larkin about this, he went straight to the mistake he sees most often.
00:17:35.775 --> 00:17:45.075
We talked about the progression from ideation all the way through to production and how a lot of people skip the critical step of making sure you're building the right thing.
00:17:45.345 --> 00:17:51.045
'cause we don't wanna create a perfect solution to a problem that doesn't exist or that nobody wants to really use.
00:17:51.069 --> 00:17:55.180
Whatever it is you're building, so you wanna make sure you're building the right it before you build it.
00:17:55.180 --> 00:17:55.599
Right.
00:17:55.960 --> 00:18:00.130
Build the right it before you build it.
00:18:00.339 --> 00:18:00.670
Right.
00:18:01.029 --> 00:18:09.789
Softchoice has developed what they call an executive alignment session or EAS, designed to help organizations work through exactly this.
00:18:10.930 --> 00:18:14.259
Not the technology, the strategy underneath it.
00:18:14.740 --> 00:18:18.519
The why before the what Ron Espinsa has seen.
00:18:18.519 --> 00:18:19.720
The difference it makes
00:18:20.170 --> 00:18:24.460
the EAS is such a big deal is taking that time to measure twice, cut once.
00:18:24.670 --> 00:18:28.329
So let's measure twice before we do any cutting and make sure that's the cut we wanna make.
00:18:28.599 --> 00:18:33.640
'cause AI allows you to make lots of cuts and it could be one that has far reaching implications.
00:18:34.029 --> 00:18:37.835
Measure it twice, cut once and from the technical side.
00:18:38.545 --> 00:18:43.855
Greg Whalen says, the organizations that actually make it to production share a discipline.
00:18:44.214 --> 00:18:48.355
They define what good looks like before they start building it.
00:18:49.345 --> 00:18:57.414
If I'm only gonna realistically look at five things in the morning, like what are those things that I'm gonna look at that indicate healthy, not healthy or something else?
00:18:57.414 --> 00:19:00.565
People having a problem defining what those things are.
00:19:00.644 --> 00:19:02.444
Right, because they haven't fully thought through what?
00:19:02.444 --> 00:19:03.434
They're really gonna build
00:19:03.825 --> 00:19:05.384
five things in the morning.
00:19:05.775 --> 00:19:07.125
That's Greg's test.
00:19:07.605 --> 00:19:17.865
Before you build anything with ai, before a single line of code gets generated, can you name the five metrics you check each morning to know if it was working?
00:19:18.285 --> 00:19:23.263
If you can't answer that question, you are not ready for production and that's okay.
00:19:23.414 --> 00:19:25.184
That's what experimentation is for.
00:19:25.809 --> 00:19:34.390
Vibe coating is awesome for the rapid prototype phases where you're testing out an idea and you don't have to have perfect data, perfect authentication, perfect.
00:19:34.569 --> 00:19:35.349
Anything really.
00:19:35.539 --> 00:19:38.450
can do a lot without those elements being in place.
00:19:38.779 --> 00:19:39.769
It does have a home.
00:19:39.859 --> 00:19:41.150
It's just not in production.
00:19:41.200 --> 00:19:42.640
I would say it's great for imagineers,
00:19:43.240 --> 00:19:50.950
it's great for imagineers and Ron's advice to organizations still on the fence about getting started is blunt.
00:19:51.369 --> 00:19:55.059
If you're waiting for the dust to settle, you're just gonna end up as part of that dust.
00:19:55.240 --> 00:19:58.119
I think companies and people have to start getting themselves dirty.
00:19:58.180 --> 00:19:59.710
They have to start putting their hands in it.
00:19:59.710 --> 00:20:00.789
You have to start trying it.
00:20:00.940 --> 00:20:02.710
Get your hands dirty.
00:20:03.160 --> 00:20:04.150
Experiment.
00:20:04.539 --> 00:20:09.009
Build the baseball app, but know the line between the prototype.
00:20:09.339 --> 00:20:11.950
And production and invest in the parts.
00:20:11.950 --> 00:20:20.140
Nobody wants to do the observability, the testing, the governance, because that's where the actual value compounds.
00:20:20.740 --> 00:20:31.000
Greg's view is that Amazon's current fix, having senior engineers review every line of AI generated code works for now, but it won't scale.
00:20:31.240 --> 00:20:33.069
AWS's solution is practical.
00:20:33.190 --> 00:20:33.910
Hey guys.
00:20:34.349 --> 00:20:38.339
Everybody has to review the code that's practical for today.
00:20:38.640 --> 00:20:41.220
Prove AI looks at that problem and says, that's not realistic.
00:20:41.220 --> 00:20:42.630
Nobody's gonna do that, right?
00:20:42.630 --> 00:20:43.950
Let's look in five years.
00:20:43.950 --> 00:20:46.319
Do you think people are gonna like, that's even worse?
00:20:46.500 --> 00:20:50.579
The answer, Greg believes, isn't more human review.
00:20:51.240 --> 00:20:59.400
It's better AI observability systems that can watch what AI builds and flag problems before they reach production.
00:21:00.119 --> 00:21:02.400
That's what prove AI is building.
00:21:02.875 --> 00:21:05.460
And it's a space that barely existed a year ago.
00:21:06.115 --> 00:21:12.234
But today, for the IT leader listening to this right now, the playbook starts with something simpler.
00:21:12.954 --> 00:21:15.984
Something Ron's 14-year-old put into words,
00:21:16.075 --> 00:21:21.565
I'm gonna quote my actual 14-year-old son, and he said, I love what you're doing with ai.
00:21:21.565 --> 00:21:26.275
I love what I can do with ai, but AI can't conceive of something new on its own.
00:21:27.954 --> 00:21:30.625
It can't conceive of something new on its own.
00:21:31.164 --> 00:21:33.115
The creativity, the judgment.
00:21:33.500 --> 00:21:38.539
The decision about what to build and why and whether it's safe, that's all yours.
00:21:38.869 --> 00:21:41.390
AI just made the building part faster.
00:21:42.109 --> 00:21:45.769
Your job is to make sure the thinking doesn't get skipped.
00:21:49.855 --> 00:21:52.974
Here's what I keep coming back to from these conversations.
00:21:53.755 --> 00:21:55.315
Vibe coding is real.
00:21:55.674 --> 00:21:56.694
It's not a fad.
00:21:56.994 --> 00:22:01.795
Even self-described skeptics with decades of experience are calling this genuine.
00:22:02.394 --> 00:22:04.134
The speed is extraordinary.
00:22:04.585 --> 00:22:07.134
The creativity it unlocks is extraordinary.
00:22:07.845 --> 00:22:11.924
Ron built a working app in an afternoon without writing a line of code.
00:22:12.585 --> 00:22:16.335
Sean sees an entire generation of new builders emerging.
00:22:17.984 --> 00:22:23.595
The pilot is dead long, live the pilot, but the speed is also the risk.
00:22:23.924 --> 00:22:29.714
AI gives you 80% of the result after 20% of the work and the remaining 20%.
00:22:30.319 --> 00:22:36.680
The observability, the testing, the governance, the question of what success even looks like.
00:22:37.309 --> 00:22:40.849
That's the part that separates a demo from a production system.
00:22:41.059 --> 00:22:43.339
It's the part everyone wants to skip.
00:22:44.029 --> 00:22:46.910
It's the marathon shortcut that Greg described.
00:22:47.089 --> 00:22:54.470
So if you are an IT leader listening to this and thinking, how do I let my team experiment without losing control?
00:22:54.619 --> 00:22:57.670
How do I say yes to the speed without saying yes to the risk?
00:22:58.500 --> 00:23:00.420
Soft choice can help you figure this out.
00:23:01.200 --> 00:23:18.990
Their executive alignment sessions are designed to help organizations define the why before the what to make sure you're building the right thing before you build it right, and you're developing a lighter AI powered version that you'll be able to access on your own terms.
00:23:19.470 --> 00:23:24.869
Visit softchoice.com/e to learn more.
00:23:25.650 --> 00:23:27.779
That's softchoice.com.
00:23:28.224 --> 00:23:30.535
Forward slash EAS.
00:23:33.625 --> 00:23:38.875
The catalyst was reported and produced by Tobin Dalrimple and the team at Pilgrim.
00:23:38.875 --> 00:23:41.394
Content Editing by Ryan Clark.
00:23:41.515 --> 00:23:46.795
With support from Philippe Demas, Joseph Beyer, and the marketing team at Softchoice.
00:23:47.125 --> 00:23:54.654
Special thanks to Greg Whalen, Ron Espinsa and Sean Larkin for sharing their expertise and insights.
00:24:00.369 --> 00:24:03.730
Thanks again to Veeam Data Cloud Vault for supporting today's episode.
00:24:04.029 --> 00:24:09.789
If unpredictable cloud storage costs and backup security keep showing up on your radar, this one's worth a look.
00:24:09.849 --> 00:24:13.240
Click the link in the description to sign up for their AWS demo.
