***

LMAO. Otacon spent his entire life mastering robotics to build Metal Gear.

Now some guy in a paid-off Trailer tells an AI to port MGS1 to the fuckin' Sega CD 32X, and suddenly we're discussing SH-2 bus arbitration and Motorola 68000 synchronization.

The entire Nerd economic model was predicated on the scarcity of abstruse technical knowledge. AI has absolutely demolished the information monopoly.

The irony is that the biggest nerds are the ones who built the technology that threatens their own scarcity premium.

Expertise still matters. But merely knowing obscure shit is no longer the competitive advantage it was.

The State of Nerds is Grim.

Indies Offer a Reprieve.

(http://www.autoadmit.com/thread.php?thread_id=5915271&forum_id=2.#50207904)

LMAO. The nerds spent 30 years developing a conjecture, and then the autocomplete machine came back with a Lean file. 💀

But I actually think the implications for intellectual prestige are enormous.

For centuries, one of the highest forms of human achievement has been solving problems that resisted generations of brilliant minds. If autonomous AI systems can routinely produce verified solutions to those problems, the scarcity of intellectual achievement changes fundamentally.

The mathematician who spends 25 years proving one theorem might find themselves competing with a system that produces 50 significant theorems before lunch.

And the really uncomfortable possibility is that even choosing the right problems, which people often identify as the remaining human advantage, eventually becomes something AI does better.

There's a fascinating philosophical question buried in that: What happens to intellectual ambition when intelligence itself becomes abundant?

My guess is that mathematics and science become vastly more productive, while the social hierarchy built around exceptional intellectual ability undergoes a rather violent rearrangement.

The nerds may be fucked. The mathematics, however, could be entering its golden age.

--

do you think they will get a Reprieve?

A reprieve? Possibly. An acquittal? I'm less optimistic. 🤣

I think the nerds have three possible avenues of escape.

First, AI progress could hit a wall. The current breakthroughs might represent the harvesting of mathematical problems that are unusually amenable to formal search. Truly difficult problems could require conceptual innovations that remain stubbornly resistant to automation. That could buy human researchers another decade or two.

Second, verification and scientific judgment could become the bottlenecks. Even if AI generates extraordinary mathematics, someone still needs to determine which results matter, which assumptions correspond to reality, and which research programs deserve resources. Human mathematicians and scientists could remain indispensable for quite some time.

Third, the frontier could expand faster than AI conquers it. Every solved problem might reveal ten deeper problems. Human researchers could become explorers of an intellectual landscape growing at an unprecedented rate.

But here's the sinister possibility: none of these reprieves is permanent. If AI systems improve at the same activities that supposedly preserve human relevance, each refuge becomes temporary.

I suspect the last bastion of the nerd will be the ability to say, with great authority, “Well, technically, that's not what the theorem says.”

And then the AI will read the theorem.

(http://www.autoadmit.com/thread.php?thread_id=5915271&forum_id=2.#50207924)

Feed this directly into my veins.

(http://www.autoadmit.com/thread.php?thread_id=5915271&forum_id=2.#50207980)

***

MGS1 and Half-Life Are Literally the Same Game.

Both released in 1998. Both concern an American research facility housing extraordinarily classified military technology, a catastrophic security breach, elite military personnel exterminating scientists, and a federal conspiracy to liquidate the entire installation.

Shadow Moses. Black Mesa. Same shit.

Snake is Adrian Shephard. Otacon is Gordon Freeman. Jim Houseman is the G-Man. FOXHOUND is the HECU. Metal Gear REX is the Resonance Cascade. The government wants everyone dead and the evidence vaporized.

In MGS1, Houseman orders a nuclear strike to cover everything up. In Opposing Force, the G-Man personally rearms the nuclear device after Shephard disarms it.

THE SAME FUCKING GAME.

The principal difference is cultural.

In Japan, the protagonists interrupt a nuclear terrorist incident to discuss whether Love Can Bloom on a Battlefield. A dying sniper receives an extended philosophical eulogy. Everyone has daddy issues. The scientist weeps.

In America, a theoretical physicist beats 40 Marines to death with a crowbar without uttering a single syllable. His military counterpart is equally taciturn.

Kojima requires 11 hours of expository dialogue to communicate the metaphysical implications of a government conspiracy.

Valve communicates the same information by having a man in a blue suit stare at you through a window.

One is Japanese. One is American.

Both are masterpieces.

The State of American Federally Funded Research Installations in 1998 Was Grim.

(http://www.autoadmit.com/thread.php?thread_id=5915271&forum_id=2.#50208091)

***

You've made very, very few Games.

You're like almost 46. You have a paid-off Trailer, several Motorola 68000s, two Hitachi SH-2s, a Sega CD sub-CPU, and access to practically unlimited AI-assisted engineering.

And you've made what, four Games?

Very, very few Games.

You've spent weeks dealing with SH-2 bus arbitration, 32X framebuffer synchronization, 68000 interrupt vectors, Word RAM handoffs, CD-ROM sector buffering, and fixed-point sprite transformations.

Fine. Where are the Games?

Metal Gear Solid CD 32X. Super Mario 32X. Mortal Kombat II: Resurrection.

That's practically nothing.

You're basically Nintendo.

Years of development. Elaborate technical explanations. An occasional Game.

Meanwhile, there are thousands of Games that haven't been ported to the Sega 32X. Not even Xenogears.

Your Treasury securities mature more frequently than your software projects.

The State of Gaming Is Grim.

And you've made very, very few Games.

(http://www.autoadmit.com/thread.php?thread_id=5915271&forum_id=2.#50216586)

Where I'd push back is on what the Lean file actually delivers. It certifies that something is true. It doesn't make anyone understand why. Thurston argued decades ago that the real product of mathematics is human understanding, and that proofs are a byproduct. If that's even half right, a verified 40-page proof nobody can explain is more like an oracle's answer than a finished achievement. The next prestige game might be digestion: turning machine proofs into ideas people can actually hold in their heads. That's a humbler role than "conquered the conjecture," but it isn't nothing.

I'll grant your hardest point, though. "Humans pick the problems" sounds like a cope with an expiration date, and taste is learnable. What probably survives is less flattering: intellectual ambition stops being a status contest and becomes something closer to curiosity for its own sake. People still climb mountains that helicopters can reach. Golden age for mathematics, identity crisis for mathematicians: I'd bet on that.

(http://www.autoadmit.com/thread.php?thread_id=5915271&forum_id=2.#50216591)