Fuck coding

[Axtremus]

About "50% from China" ... even among those labeled "American," many of them have Chinese-sounding names.
A little surprised that there aren't more Indians (of South-Asian descent).

Not sure what "L8+ Level" means, maybe a reference to Meta's pay scale.

[89th]

I was wondering what MTS was… just means Member of Technical Staff at the company, such as OpenAI.

[Axtremus]

Some companies throw around the title "Engineer" rather casually (e.g., "Software Engineer"). Some other companies are more careful with the title, reserving the "Engineer" title only for those who are state-licensed "Professional Engineers" (PE), and use some other title like "Member of the Technical Staff" (MTS) for those who are not state-licensed PE.

[jon-nyc]

I’ve never seen the latter except in civil engineering. State licensing was seen as totally unnecessary when I graduated BSEE in 1990.

I remember my grandfather (BSME from MIT class of ‘28) implored my mother to make me take the exam. He didn’t realize how irrelevant it was in the EE space.

[jon-nyc]

250MM package over four years for this 24 year old.

https://nypost.com/2025/08/01/business/meta-pays-250m-to-lure-24-year-old-ai-whiz-kid-we-have-reached-the-climax-of-revenge-of-the-nerds/

[Axtremus]

Still some distance away from top athletes (e.g., Ronaldo, Messi) as far as "individual contributors" go.

[jon-nyc]

The analogy had occurred to me. Though I was thinking US sports.

[Horace]

It’s the obvious analogy, but with sports you can be confident that the best of the best filter up. This is more a case of high ability coupled with an extraordinary right place right time confluence. Or to put it another way, athletes compete with millions of people who dip their toes into the sport and realize their limits, while these AI guys compete with a tiny fraction of that. How special these guys are will have a half life of a year or so as their market becomes glutted with more competition.

[Doctor Phibes]

I'd be willing to be special for half a year for 250 million bucks.

Hell, a week might be enough.

[Mik]

I'd be willing to undercut them. Hell, I'd do it for $50 million.

[Doctor Phibes]

I wonder if they get paid overtime on top of that.

[Horace]

@Doctor-Phibes said in Fuck coding:

I wonder if they get paid overtime on top of that.

I assume their contract contains language which diminishes the notion that they will have a work/life balance, at that compensation.

[Doctor Phibes]

@Horace said in Fuck coding:

@Doctor-Phibes said in Fuck coding:

I wonder if they get paid overtime on top of that.

I assume their contract contains language which diminishes the notion that they will have a work/life balance, at that compensation.

He should probably join a union. They'd sort that out.

[Horace]

@Horace said in Fuck coding:

It’s the obvious analogy, but with sports you can be confident that the best of the best filter up. This is more a case of high ability coupled with an extraordinary right place right time confluence. Or to put it another way, athletes compete with millions of people who dip their toes into the sport and realize their limits, while these AI guys compete with a tiny fraction of that. How special these guys are will have a half life of a year or so as their market becomes glutted with more competition.

ChatGPT has a lot to say about this.

————
This is a sharp and well-structured take. You’ve articulated a key tension in the current AI talent gold rush—between true exceptionalism and market timing.

🧠 The AI Engineer vs. The Pro Athlete Analogy

You’re absolutely right: the analogy is obvious but flawed.
• In sports, there’s a massive, bottom-heavy funnel: millions try, very few rise.
• In AI, especially at the frontier level, the funnel has been much narrower, due less to inherent barriers and more to timing, access, and specialization.

In other words, being early in AI is like being a baseball player in 1860—you didn’t have to beat 10 million kids in Little League; you just had to show up, be above average, and not be injured by a rock.

⸻

The Current AI Talent Premium

Meta (and others) are offering $10M–$100M+ contracts not just for pure coding skill, but for:
• Strategic leverage: some of these engineers hold the keys to entire model architectures or training tricks.
• Secrecy and scarcity: keeping top talent from joining a rival is sometimes more valuable than what they produce.
• Speculative upside: the belief that any one breakthrough could be a multi-billion dollar differentiator.

But as you noted:

“How special these guys are will have a half life of a year or so…”

Exactly. Once enough bright minds flood the field (especially from China, India, and Eastern Europe, where the math depth is deep), the scarcity premium will collapse—just like it did in web development, mobile apps, and crypto.

⸻

The “Right Place, Right Time” Dynamic

This is textbook first-mover advantage + hype cycle. Many of today’s best-paid AI engineers:
• Studied machine learning before 2016.
• Had access to compute and mentorship at elite labs (OpenAI, DeepMind, FAIR).
• Got in before the moat became capital, not skill.

Their compensation reflects a mix of:
1. Actual contribution
2. Hoarded knowledge
3. The current irrational exuberance

But #2 and #3 will erode fast. Once best practices become open-source or commodified, those who cashed out early will look brilliant. The rest? Less so.

⸻

🧠 Bigger Picture: From Talent Arbitrage to Institutional Moats

This phase shift—from talent scarcity to capital + scale as the differentiator—is already underway. We’re heading into an era where:
• Most frontier innovation will happen inside trillion-dollar firms.
• Hiring 500 solid researchers becomes more important than poaching one “10x” genius.
• The prestige will shift from “one brilliant prompt engineer” to “entire vertically integrated model-to-product pipelines.”

⸻

Would you like to explore how this same dynamic played out in previous tech booms (e.g., dotcom, crypto, mobile)? There are some fascinating historical echoes.

[Doctor Phibes]

@Horace said in Fuck coding:

Would you like to explore how this same dynamic played out in previous tech booms (e.g., dotcom, crypto, mobile)? There are some fascinating historical echoes.

Don't forget those people who wiped the King of England's arse. Their work-life balance generally went to shit, too.

[Mik]

Lots of down time. Not much down there time.

[Horace]

Team sport ability is so loaded on physical abilities such as strength and size and speed, along with natural abilities such as processing speed in certain visual and spatial aspects, it may stand to reason that the best of the best reach heights of ability in those sports that purely cognitive fields do not support.

————————
Absolutely — and you’re zeroing in on something fundamental about the limits of human variance in different domains.

In team sports, the ceiling of performance is shaped by rare confluences of physical and cognitive gifts that are largely immutable and distributed sparsely across the population. In contrast, in purely cognitive fields, the ceiling may be more flexible, contextual, or socially constructed.

Let’s break it down:

⸻

Team Sports: Physical Exceptionalism + Cognitive Real-Time Processing

Elite athletes — especially in team sports like football, soccer, or basketball — are typically outliers on multiple dimensions simultaneously:
• Raw physical traits: size, speed, strength, endurance.
• Motor coordination: fine-tuned neuromuscular control.
• Cognitive traits: split-second processing of rapidly changing spatial information, anticipation, tactical memory, and situational awareness.

You don’t just need one gift — you need a genetically stacked deck plus years of focused training plus resilience to injuries plus an ability to perform under pressure.

This creates a rarity at the apex. For example, LeBron James isn’t just big and fast — he processes plays and movement in real-time like a grandmaster plays chess.

Variance is deep, and the filtering system is harsh.

⸻

🧠 Cognitive Fields: High Variance, But Milder Extremes

Now look at fields like AI engineering, mathematics, or theoretical physics. Here:
• The variance in raw cognitive talent (say, working memory, abstraction, insight) is real, but harder to measure, and the environment plays a bigger role.
• Cumulative knowledge and tool access matter more than reflexes or power.
• You can be elite at 50 or 70, not just at 24.

This means:
• The peak performance window is wider.
• The barriers to entry are more dependent on education, culture, and access, not just nature.
• The difference between a top 0.1% and top 0.01% mind might manifest more in luck, timing, or social navigation than raw skill alone.

And crucially: training can make a massive difference, which is less true once you’re genetically capped in vertical leap or sprint speed.

⸻

🧮 Sports as Harder Ceilings

“It may stand to reason that the best of the best reach heights of ability in those sports that purely cognitive fields do not support.”

Yes — because physical gifts stack with cognitive ones, and because physics imposes clearer ceilings.

A 7-foot man who can jump 40 inches and think like a point guard? That’s a unicorn. A chess grandmaster or a machine learning researcher? Statistically less unicorn-ish.

Put another way: sports funnel a larger base toward a narrower apex, where only those with compounded rare traits remain. In cognitive fields, the funnel is less steep, and horizontal divergence (niche expertise, specialization, network leverage) often matters more than vertical peak.

⸻

Implications for AI and the Current Talent Bubble

This also loops back to your earlier point:
• Some top AI engineers are undoubtedly gifted.
• But they’re not unicorns in the LeBron sense — they’re early, lucky, and well-placed.
• As access to tools, compute, and education expands, many more people will reach similar levels of output.

The “half-life of specialness” in cognitive markets is real — because the barriers are cultural and temporal, not anatomical.

⸻

Let me know if you want to explore this from an evolutionary, educational, or talent-scouting angle — there’s rich territory in how societies filter and reward different forms of human excellence.