Empowering the Expert: Give Smart People the Tools to Do Smart Things

If you have spent any time on X (formerly Twitter) or browsing tech journals recently, you have likely encountered a relentless barrage of hyperbolic claims. You know the ones:

• Anthropic asserting that AI will dismantle and replace entire industries within a few months.
• Sam Altman (OpenAI) suggesting that massive compute scales could unlock the cure for cancer.
• Elon Musk claiming that compilers will become obsolete because AI will simply write binary directly.
• Anthropic warning that their Mythos model is a terrifying "cyberweapon" that the general public is too fragile to handle.

The "Secret" Narrative

There is a recurring subtext to these announcements. The narrative suggests that AI is faster, cheaper, and tireless. Consequently, anyone who expresses skepticism—usually the actual subject matter experts—is labeled a "dumb luddite" who is failing to adapt and is destined for the "permanent underclass."

Conveniently, "adapting" usually involves a specific financial commitment:

1. Paying a $200/month personal subscription.
2. Convincing your employer to sink $4,000/month into tokens so you can be "more productive."

---

Artifacts $\neq$ Work

We must distinguish between the result and the process. Historically, tools did not replace the professional; they expanded the professional's horizon.

As the formula suggests: $\text{Total Value} = \text{Process (Understanding)} + \text{Artifact (Output)}$

AI companies focus exclusively on the Artifact, ignoring the immense value found in the Process. In software engineering, the real bottleneck isn't typing code; it's understanding the business requirements, designing the system, and maintaining the long-term health of the codebase.

"The hard part is knowing if something is good, what tradeoffs it makes, and the consequences of a particular implementation or choice."

The Prompting Fallacy

Some argue you can simply write a "perfect prompt" to encapsulate a complex business problem. However, anyone who has dealt with outsourcing over the last three decades knows that this is a fantasy.

The idea that a probabilistic system will replace a deterministic tool like a compiler is, frankly, absurd.

The Marketing of "Cyberweapons"

Among legitimate security professionals, the release of Mythos didn't inspire awe—it inspired annoyance. Anthropic took a fascinating technical capability and wrapped it in "cyberweapon" marketing. This does two things:

1. It devalues the painstaking work of actual pentesters and security researchers.
2. It obscures the actual utility of the model behind a veil of reckless hype.

Why the Hype?

There is a financial motive here. Currently, investor appetite for AI is astronomical. However, selling a "vision of the future" only works for so long. Eventually, investors will stop caring about the mythology and start caring about the margins. AI firms are racing to convert hype into capital and customer dependency before the bubble shifts.

---

A Vision for Complementary AI

I envision a future where AI complements human expertise rather than threatening to replace it. We have seen this pattern before with game development tools.

The Unity/Unreal Engine Effect:

• The Downside: A flood of low-quality "asset-flip" games.
• The Upside: Solo developers and tiny teams with genuine skill could suddenly build massive, complex dream projects without needing to write a custom engine from scratch.

Midjourney Medical is a step in the right direction. Providing doctors with non-invasive, high-fidelity insights into the human body doesn't "cure cancer" automatically, but it makes the expert dramatically more effective at diagnosis and treatment. (Though, the "wellness spa" branding shows how quickly technical utility is swallowed by consumer futurism).

Final Request to AI Executives

Instead of threatening the workforce with "clankers," try this:

• Stop marketing replacement.
• Start marketing empowerment.
• Build tools that enhance the expert's reach.
• Focus on the process, not just the artifact.

Bottom line: Give smart people the tools to do smart things.