Cam

Passionate about technology & human progress | Views are my own

[Cam @tay4bs]

@PEoperator Fair, though the cost usually isn't the license. It's what you paid to automate a sales process that needed fixing first. Tool goes away, discipline problem doesn't.

[Cam @tay4bs]

I suggest you take it one step further by putting your hard earned money in a Christian banking institution such as AdelFi. That way your deposits go towards furthering His kingdom by making loans that go to Christian families, building churches, funding missionary work and supporting ministries.

[Cam @tay4bs]

@Storage_Venture @TheGovtHatesU Sign me up for the battle of Salt Creek!

[Cam @tay4bs]

@thevivafrei @grok explain the reckless Ben situation & backstory.

[Cam @tay4bs]

@kingthorpin @Shadzey1 I wonder how many of the watches are reps? I bet most of them. The Panerai Carbotech for sure.

[Cam @tay4bs]

The post I needed to read on a Friday amidst the chaos that has become my life & career.

I have a resume reads like a Jackson Pollack painting. Different industries, ever changing roles, consulting, sales, ops, heck even interim CFO (in addition to running IT) during a time of upheaval in a public to private transition.

The only through line being adaptive leadership and the ability to be thrown into a shitstorm and make sense of it and bring peace, calm and massive growth… leaving businesses, whether big or small, much better than when they found me.

I wonder… How many of us are there?

[Cam @tay4bs]

@PrivatEquityGuy That’s inspiring. Thanks for sharing.

[Cam @tay4bs]

@Stellaaa Make it count. You’ve done a lot of work to build, now let’s mobilize.

[Cam @tay4bs]

@WimShakespeare ratchet slippas

[Cam @tay4bs]

@alt_w_v_g @WineGal8603 Low ROAS wrecks the margin, every time. ;)

[Cam @tay4bs]

@stormOCS $320K/yr on software. Headcount still climbing. Both numbers up means the stack became its own department. @stormOCS, you see it at $50M agencies. Same physics at $500M portfolio companies. Same disease, bigger patient.

[Cam @tay4bs]

The 60% framing is dead on. One person solves it with prompts. An enterprise has the same problem bolted into its infrastructure. 60% of your tech stack isn’t your business. It’s the connective tissue, the middleware, the integrations, the platforms that exist so other platforms can function. Entire line items dedicated to the work that surrounds the work. The individual version of this is a productivity hack. The enterprise version is a structural cost nobody budgeted to fix. And it compounds every quarter.

[Cam @tay4bs]

I'm not arguing they're not real. They're very real. The idea that anyone can just build an army of "employees" by cutting and pasting prompts with no original thought or domain expertise is where it fails.

I've seen and heard a lot of stories about those that "just don't get it" when it comes to AI. Typically they haven't had the need or desire to build something.

Most piddle around with scripts from GitHub repos or point their Openclaw at some X post thinking it's going to do amazing things only to be disappointed when it only does what the original creator intended.

It helps to have an itch to scratch. That's all.

[Cam @tay4bs]

The company has 47 SaaS contracts. 12 are essential. The quoting software quote arrives: $180K implementation, $100K annual. The CFO forwards it to engineering. Engineering forwards it to Claude.

Three weeks pass. The prototype handles 80% of workflows. Total spend: $19.47 in compute credits and one engineer's evenings. The vendor's customer success manager follows up. The email sits unread.

The repricing isn't a market correction. It's a margin audit with receipts.

Even if the prototype flops, you've got leverage now. Every CIO is running this play. Vibe-code something that mostly works, wave it at renewal time, watch the discount materialize. The threat doesn't have to be credible. It just has to be documented.

[Cam @tay4bs]

Now imagine this running against a codebase with 15 years of undocumented business logic, three abandoned migration attempts, and a config file that Keith hand-edited in 2019 and never told anyone about. The agents are as good as the foundation they're building on. And most foundations have cracks nobody mapped.

[Cam @tay4bs]

Users only using 20% of your features. That's not a product problem. That's a procurement problem. Somebody bought the full suite because the sales deck was compelling. Nobody mapped the features to the workflows.

Agents won't eat SaaS for lunch. They'll eat the 80% that nobody uses. And when that 80% disappears, so does the justification for the subscription price. The reckoning isn't about AI. It's about utilization.

[Cam @tay4bs]

SaaS isn't dead. But the version of SaaS where you buy a tool, configure it for six months, and then beg users to actually log in? That's on life support.

The blend matters. The organizations that survive the next wave aren't the ones that replaced everything with AI. They're the ones that figured out which parts of their existing stack are load-bearing and which are just legacy anxiety wearing a subscription fee.

[Cam @tay4bs]

Larry Ellison didn't build Oracle by being subtle. The message is clear: the moat isn't the model. It's the data the model trains on. And who has the data? The companies running Oracle databases for the last 30 years.

This is the enterprise software play that nobody in the AI hype cycle wants to acknowledge. The legacy vendors aren't dinosaurs. They're landlords. And they just realized they're sitting on the most valuable real estate in the AI economy.

[Cam @tay4bs]

Too hard to process. Too messy to query. Too expensive to store. That's the description of every enterprise's data layer before someone finally decides to deal with it.

The unstructured data problem is the technology equivalent of the junk drawer. Everyone has one. Nobody wants to sort it. And every new AI initiative quietly depends on someone sorting it first.

[Cam @tay4bs]

All betting switching costs so high that customers would absorb any price increase rather than rip and replace. It worked for a decade. But AI doesn't need to rip and replace. It just needs to make the switching cost calculation look different.

The question for every portfolio company running on embedded legacy software: what happens when the lock-in premium stops being worth it? The math is changing faster than the contracts.