Napoli Fella

Stand with 🇺🇦 and 🇮🇱

German guy living in bella Napoli

#space #science and #technology nerd at night, #military and #politics nerd at day

cat dad

[Napoli Fella @Napoli_Fella]

@danfaggella I tend to agree with you 99 percent. But there is one wildcard. What if ASI without real agency is possible? No self but only a superintelligent tool? Like a really smart but perfectly guardrailed LLM.

[Napoli Fella @Napoli_Fella]

@buccocapital AGI is worth significantly more than 1 trillion. Nothing is priced in besides software engineering.

[Napoli Fella @Napoli_Fella]

@DrPhiltill With ASI in the game it could be pretty fast. But who knows. Maybe you need a lot of real world experimentation that no intelligence can really speed up. I agree that full molecular nanotechnology is one of the holy grails of technology that look really far away.

[Napoli Fella @Napoli_Fella]

@StefanFSchubert He is right. There hasn't been any successful sustained reversal of below replacement fertility to above replacement fertility anywhere in the world. Without AI, humanity is on the trajectory to productivity collapse, followed by economic and subsequent civilizational collapse.

[Napoli Fella @Napoli_Fella]

@rand_longevity Very very early in. Since a few month maybe. The curve will get much, much steeper once AI becomes truly superhuman in advancing math and programming. RSI hasn't meaningfully taken off just yet, it's just the beginning.

[Napoli Fella @Napoli_Fella]

This 👇

[Andrew Ho @andrewho03]

I'm not that worried about AI safety, but that's because I think that AI might be humanity's last hope before descending into, at best, a "dark age" lasting centuries. Since I don't think AI risk is extinction-level, I'm willing to take a bet on stronger capabilities.

Prior to the development of modern AI, I had a rather dim view of our long-term future. There were (and still are) a number of headwinds that reduce my estimation of long-term economic growth:

- Falling birthrates worldwide, especially in Asia, mean that world population will peak sometime in the 21st century. The problem of falling birthrates is solvable but only with a degree of financial expenditure that states will not find palatable until the hour is late. Also, falling birthrates seem to be disproportionately concentrated in countries that are more well-educated and economically successful, leading to severe adverse selection in subsequent generations.

- Even before population peaks, the age distribution will shift heavily toward the elderly, leading to capture of political institutions and leadership structures by the old. This promotes politics of redistribution from young to old (such as the "triple lock" mandate of the UK Basic State Pension), depressing the economy, further reducing the birthrate, and reducing fiscal appetite for fertility subsidies.

- Redistribution toward the old also promotes anti-market attitudes in the young, which is bad for economic growth and worsens the birthrate problem. The most obvious example is California Prop 13 which limits property taxes, effectively transferring money from young workers to retirees; the feeling of housing unaffordability and an inability to "get ahead" in life has clearly contributed to the popularity of socialist policies among the young.

- America, currently the wealthiest country on the planet, appears to be increasingly captured by rent-seeking special interests, disproportionately concentrated in its most economically productive cities. For example, the San Francisco annual budget is ~$17 billion dollars, or approximately ~$2 million dollars per homeless person, yet the problem of homelessness worsens each year. California high-speed rail was so poorly managed that the French contractor literally quit and went to build HSR for Morocco instead. Building new elevators for the Manhattan subway cost around $100 million dollars each. And so on... it's a pretty dismal long-term trajectory.

Overall, it seemed to me as though even if we could expect positive economic growth for a number of additional decades, that we'd potentially begin to see negative economic growth within my lifetime. Negative economic growth is a horrible thing; imagine the 2008 financial crisis just repeating endlessly, businesses closing down, people getting fired and being unable to find jobs, governments running out of money, and a general feeling of hopelessness. Really bad stuff.

What's most exciting about me is the ability of AI to reverse these trends -- cut through bureaucracy and reinvigorate the economy just enough so that we get back into a positive feedback loop where people feel good about the markets and hence become more pro-market, leading to even greater economic growth, etc. It almost feels as though the period of growth following the Industrial Revolution was an aberration in the history of mankind -- most of which has been "nasty, brutish, and short," characterized by zero- or even negative-sum competition with slow and plodding growth rates -- and people are now beginning to forget how special it is to have a positive-sum period of rapid growth. It's really important that we shift as quickly as we can from the negative feedback cycle regime, which is sure to destroy trillions of dollars of value in the coming decades, to a positive feedback cycle regime in which we can escape the worse parts of our base psychological urges.

[Napoli Fella @Napoli_Fella]

@SerendipitousOr I still think it's necessary if you want FAST generalization over arbitrary tasks.

[Napoli Fella @Napoli_Fella]

@elder_plinius Like with everything there are levels of superintelligence. What we see right now is something I would call weak/narrow/spiky SI. General SI in my view would be something better than the best human in everything. And strong SI better than humanity combined in everything.

[Napoli Fella @Napoli_Fella]

@NateSilver538 They say they spent 2000usd in API tokens. For 2000usd only people in third world sweatshops would even type the 250 pages if you dictated them.

[Napoli Fella @Napoli_Fella]

@kimmonismus

[Napoli Fella @Napoli_Fella]

@Awk20000 so what? Who cares? He even used a computer and possibly even electricity. You people are retarded.

[Napoli Fella @Napoli_Fella]

@polynoamial @iamgroguu @DaveShapi What would you say is still missing for a model to develop genuinely new fields of math or physics or to come up with new conjectures that advance a field?

[Napoli Fella @Napoli_Fella]

But "experts" told me it's just stochastic parrots that retell what they read on the internet

[Noam Brown @polynoamial]

An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.

We believe it will be a major step for scientific reasoning. https://t.co/iP6cyheZ7i

[Napoli Fella @Napoli_Fella]

@TiloJung @Inclutus Deine antisemitischen Dogwhistles wären mir schon sehr peinlich, aber ich bin auch nicht linksradikal.

[Napoli Fella @Napoli_Fella]

@niccruzpatane @SpaceX A Mars landing in itself is not immediately economically relevant. Getting starlink V3/4 and data center constellations up and running is far more significant.

[Mark Kretschmann @mark_k]

Now needed more than ever:

Effective Accelerationism

[Napoli Fella @Napoli_Fella]

@rbhar90 The road to hell is paved with good intentions. Deceleration will only help entrenched elite interests and/or China. 99.99 percent of humanity will profit far more from accelerating faster.

[Napoli Fella @Napoli_Fella]

@deanwball Every disease cured later, every life lost because the cure arrived late, every misery because abundance got postponed will be on the decelerationists. Deceleration will help no one except the elite in power which needs time to entrench and consolidate.

[Napoli Fella @Napoli_Fella]

Zuck is right

[Mark Zuckerberg @finkd]

Give it a read: https://t.co/Rq4O3xRHdL