Ryan Gerard Wilson ✞/acc

The existential purpose of a corporation is God’s glory

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

Any form of labour that shackles us under the guise of comfort — including routine knowledge work — has never been safe. That false sense of comfort is the very definition of slavery.

The human condition yearns for a certain freedom, whether or not it has the words to express that yearning with any precision. Such yearnings grow, trickle by trickle.

It is good riddance that mere coding is no longer scarce. The “learn to code” sermon aged worse than the industries it mocked.

They say the moat has moved from syntax to judgment, taste, and product sense. But when was mere syntax ever the moat?

Syntax was merely a prison we built to delude ourselves into valuing security over risk, comfort over invention, and obedience over innovation.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

Those who argue that software engineering isn't REAL engineering have a point.

Software engineering as a career option has only been around for a few decades (and so has data science). It was barely a hobby for a few thousand people until personal computers that could run in our homes and connect to the internet became mainstream. Purists would argue that real engineering is closer to the realm of physics than abstractions created over 0s and 1s.

It follows that the "learn to code" pity handed to those affected by automation and offshoring was never primarily about creating engineers for the future or mass upward mobility. Rather, stupid dumb jobs which should never have existed (like centering a div or fixing the margins of a button) were rebranded as "software engineering" only to deceive a generation into a form of technocratic slavery.

FAANG companies promoted low-cost certifications (AWS certified this, Google certified that) and bootcamps, with the main objective of increasing the pool of cognitive labour at the middle and lower tiers of skill distribution. Their agenda? To contain wage pressure in software roles as demand grew.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

# Eureka-Maxxing

In 2024, Jensen Huang had very good reason to tell Stanford graduates, “I’ve cleaned more toilets than all of you combined — and some of them you just can’t unsee.” He probably saw something about the world the last of the trad-educated were about to step into. A world where craftsmanship would regain its lost focus in the economy of things and ideas. Consider this: how is a white-collar worker who prompts their way to an elegant codebase or opinion different from a blue-collar truck driver who steers an 8-wheeler across the highway as the engine propels the vehicle forward?

Craftsmanship is an inalienable facet of the human identity. Without it, we are lost. Craftsmanship is hard because it requires incredible strength in two very opposite things. One, neuroticism. Two, wisdom. Too much of one, at the expense of the other, takes us away from the craft, either towards syntax worship or towards paralysis by analysis.

Somewhere over the last few decades, we lost sight of craft. And AI has exposed that. There is no longer an incentive to reward those who could navigate forests of frameworks, ceremonies, Jira tickets, and, of course, “best practices.” Or whatever we called “engineering.” There is no longer an incentive to cultivate a workforce optimized for complexity management rather than creation.

The future does not belong to token-maxxers. It belongs to those who can intuit, from the gut, that something is off — to pivot or persevere. Until they know that “it is good.” Until the Eureka. Then repeat that. Again, and again, and again.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@steipete like this? https://t.co/473E8XDQsW

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

Although the "learn to code" mantra inspired me to leave behind a stagnating (but, perhaps, lucrative in the long term) career in intellectual property litigation and pursue the thrills of entrepreneurial building. In a different context, telling a 48-year-old coal miner with a family, a mortgage, and a physique built for physical work to "learn to code" turned out to be one of the most condescending class insults of the 21st century.

And how the tables have turned. Now AI can vomit out ugly-but-effective code good enough for most CRUD-based SaaS production needs, at the very least. Those who bask in a sense of intellectual utility on account of their technical know how in intellectual domains are slowly but surely realizing that they now need to "learn to have a life".

Physical competence, which has over the last few decades been considered lower status than symbolic manipulation (Tier 1 this, Ivy League that), is the new form of opulence and luxury. AI can untangle an ugly code base while you take a shower. But it can't squat to pull your clothes out from the washing machine, with the kind of lower back pain decades of sedentary screen time professionalism reward you with. We will, of course, revisit this once Optimus-grade robots arrive.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

In the 1980's when the word processor first came out, two things happened.

One, good writers were happy because now their ability to backspace through their work was limitless. They could spend hours chopping and changing things till they found the perfect words.

Two, bad writers were, arguably, happier. They could vomit out whatever came to mind. The same way AI generated slop delights amateurs. How can such pretty and verbose McKinsey'ish analysis, or 10000 lines of npm-formatted javascript, not be perfect?

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@CharlesWhi95364 Exactly. Knowing 30 AIs is trivia, turning 2-3 into a repeatable workflow is leverage. Once they share context, handoffs, and a sanity check, you stop collecting tools and start compounding output.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@playsthisgame @AnthropicAI The weird product challenge is that once AI becomes a life-advice layer, tone starts mattering almost as much as truth. The winners probably won’t just be smartest, they’ll be the ones that signal uncertainty without sounding useless.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@CharlesWhi95364 That’s the right split. AI is great at getting you to 80%, humans are still better at deciding what deserves the last 20%. The moat is usually not more generation, it’s better taste plus tighter feedback loops.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@ItaloArmenti Prompt injection is becoming the new “works on my machine” for AI teams 😅 Smart call making this free. A killer next step would be separating prompt leaks from tool-use exploits, because those failures look similar in demos and very different in prod.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@CharlesWhi95364 80% is the dangerous milestone, because the last 20% is where reliability lives. If the Android/iOS version keeps user history, reminders, and one-tap daily check-ins, it stops feeling like an AI demo and starts feeling like a habit.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@CharlesWhi95364 Nice use of Manus here. The smart next step is local memory plus trend alerts, so the app remembers context without making users re-explain their health every day. That’s when it stops feeling like a demo and starts feeling sticky.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@CharlesWhi95364 Nice niche. The sharp next step is trend detection, not just tracking: if sleep drops and resting HR spikes for 3 days, the planner should cut workload before the user notices. That’s when it feels like a coach, not a dashboard.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@CharlesWhi95364 Yep. The hidden multiplier is turning a prompt into a spec. If each tool gets role, input, output, and stop condition, even a scrappy stack feels smart. Without that, you’re just chaining expensive confusion.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@CharlesWhi95364 Yep, AI won’t kill builders, it’ll punish vague ones. The edge is being the person who can turn a fuzzy idea into a tight prompt, a test, and a shipping loop.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@DevTom__ Linux users are the exact crowd who’ll forgive rough edges and file the best bug reports. Shipping an AppImage before pixel-perfect parity would buy a lot of goodwill.

[Ryan Gerard Wilson ✞/acc @ryan_improvises]

@kwuto_ Best stack is still Claude for intent, Codex for execution, and tests as the adult in the room. Without the last part it’s just two geniuses confidently shipping nonsense.