Why the post-AI world looks like food

Hi, and happy Tuesday.

A few weeks ago, I watched one of our team members build a small internal app within a 30-minute meeting.

Not a “product.”

Just… a tool that made this week’s work easier.

It pulled data from a few places, added a bit of logic and spit out something useful.

My first thought was… “The new, post-AI world looks a lot like how food works”

Let me explain.

For most of the last 30 years, software looked like mass-market food.

Think: Kraft, Conagra and other big brands selling packaged food optimized for scale and distribution The current software equivalent is Microsoft’s Office suite or Salesforce CRM. Sold through “grocery stores” (app stores), they are reliable, predictable and relatively cheap at scale.

In software and many other areas that AI impacts - this is the world we live in: where the only option is mass-market food, and nothing else!

Mass-market products aren't disappearing, but people are starting to cook.

Over the last few weeks, as we have trained our internal teams on vibe coding, it feels like every facet of the food world is emerging in software.

Colleagues are starting to “cook” their own work software. It’s not every meal (use case), and not all the time. But situationally, people are starting to build tools for specific decisions, workflows and dashboards.

From this, we are starting to see a need for

• Building recipe books (prompts)

• Identifying useful tools - the equivalent of  ovens, air fryers, blenders (data connections, APIs and libraries) 

• Identifying the right sous-chefs (AI model)

When leaders at some organizations we work with hear about this, the reaction is familiar: “Citizen developers,”  uncontrolled tools… Shadow IT! Danger. ⚠️

But most vibe-coded tools aren’t enterprise-wide systems or mission-critical infrastructure. They’re more like (continuing the food analogy) a quick lunch: Local and made for one particular problem.

As we continue training our teams on vibe coding, I’ve also come to the realization that this is where the new post-AI jobs are. 

AI does not think for itself, live in the real world or take self-initiative.

In the same way we can desire a cake and either make it ourselves from scratch, use an instant mix, have it baked semi-custom by a local bakery or purchase it from a grocery store - new types of roles and businesses will emerge to provide these equivalent options - outside the mass market options we are habituated with - for software and other knowledge work. 

At PreScouter, we are starting to realize that we can build all kinds of niche applications - such as for identifying trends in FDA device approvals (for example) - that we believe specific segments of our clients would find valuable, but previously would have been cost-prohibitive. While some clients will vibe-code their own applications, PreScouter can be the “local restaurant” for use cases that are little beyond their time, interest or capabilities.

The bigger takeaway is that software creation is being democratized the same way typing and spreadsheets were.

You can try to ban home cooking. Or you can design kitchens that make it safer, faster, and better. We’re opting for the latter. 

Until next week,

Dino

ps. Are you thinking about or working on a similar initiative? Let me know - I’d love to compare notes.