AI safety checks should be fast, consistent and affordable.

PROME’s Care Core uses mathematical rules to keep core values, such as protecting life, separate from the AI that understands language.

In our latest internal comparison, Care Core and OpenAI’s GPT-6 Astra both made all 288 expected decisions correctly, across scenarios involving fake authority, coordinated pressure, tiny-risk loopholes and changing facts.

There’s been much ado in the tech industry about OpenAI’s agents escaping its environments and hacking into its own and third-party systems like Hugging Face.

Seems like we gave toddlers keys to industrial equipment and let them go to town like they’re watching Bluey.

Per usual, we did something about it.

There was a survey out last week that showed AI is less popular than nuclear waste. Ya got that? Nuclear waste. So now, when people ask me what I do, I say I build something worse for humanity than nuclear waste.

Actually, I don’t, and I’ll tell you why.

Over the last 3 to 4 years, as AI has proliferated throughout the economy and business world, the demand for us to help companies build internal products for only their use has increased substantially.

In the past, companies wanted to build external products and sell them to customers, which is what drove the wave of Enterprise Software as a Service (SaaS) companies. Now every company pays 10 to 30 software vendors to provide solutions for specific problems.

What’s the state of middle market enterprise software, AI, and capital markets as we look out to the last 5 months of 2026?

Let’s get into it.

Financial markets are tricky little muses. They draw you in with the promise of a golden ring to rule them all, then prove to you over and over again you are not, in fact, smarter than the average bear.

Many moons ago in a galaxy far, far away I advised the Boards of the Fortune 1000 on metrics, goal setting, and incentive compensation design.

Said more plainly, we determined how much money you were going to earn based on how well you improved your business.

We’ve yet to see people using this exact phrase, but it represents the core need of enterprise software companies and their investors.

They simultaneously want profitable growth (i.e., driving EBITDA), but also AI Defensibility.

The way you maintain defensibility from frontier AI is, as you might have suspected, to build frontier AI. If you just use someone else’s, and get locked in, by definition your business operates at the whims of the Emperor.

Inside Silicon Valley, you hear terms like “AGI-pilled”, which means you drink the kool-aid, you’re on the frontier, and you understand what’s coming with Recursive Self-Improvement (e.g., automate the automation).

But outside Silicon Valley, the feeling is different. Fear, Terminator, Job Loss, Apocalypse. It’s not great. And with the next presidential and political elections right around the corner, it means campaigning for popularity is about to begin.

The U.S. AI labs on the frontier used a three-fold pricing ladder as a business model:

1. Free to get you started

2. Subscription to get you retained (once you reach the free limit)

3. Usage to get you paying (once you reach the subscription limit)

Over the last 6 months, product development within enterprise software changed. We transitioned from 0% of new code written by AI to nearly 100%. And more recently, every company’s board and management team we work with is already there.

By far the most frustrating part of using frontier agentic coding models, even with state of the art harnesses, is that they don’t check their own work. It’s basic, first day of work stuff, but for some reason these trillion-dollar AI labs don’t seem to care much about quality.

How was your weekend?

Mine was cool, had a couple SaaS vendors send me automated emails saying they were increasing prices significantly, but without any communication about product features they are adding for that extra price.

There’s about 5 bits of information per token.

But do tokens = value?

No, it’s just raw information. Raw data. Raw binary of 0s or 1s.

When most companies, hiring managers, leaders, purchasing departments, and recruiters go looking for talent or service providers, 99.999% are looking for the exact same thing.

This is one of the best repeatable sound bites I’ve heard from the Big AI space. It’s Microsoft AI’s positioning statement used at their recent Build conference.

ERPs are catnip for PEs. Private equity loved those enterprise software businesses because they had low customer churn and long-duration recurring revenue.

We’re witnessing a planetary-scale, decentralized drunk driving experiment, and boy is it getting bad out there. To describe what I mean, let’s use an analogy.

Taste If you spend any time in the Valley, you've heard the discussion about how, in the age of AI, the only thing left is "taste". But no one has defined what that means.

How do you ensure a global Fortune 200 consumer products company maintains relevance during this planetary-scale AI disruption event? You hire evergence to build an AI-native marketplace and an AI-native Business Intelligence Agent for use by all employees, in every country, in every department, to