Sabitlenmiş Tweet
Hot take: if it currently seems like ASI might kill us (or turn us into pets or zoo animals), let's not develop it yet or take actions that might make it come sooner.
English
Adam Alexander ⏹️
668 posts
Adam Alexander ⏹️
@AdamTAlexander
Teacher, Janitor, AI news-follower
Arthur Morgan School Katılım Ekim 2017
40 Takip Edilen39 Takipçiler
Exactly. I'm familiar with the concept of radiation. Just seems hard to know the extent to which NASA contractor jobs like this are mission critical in a literal sense vs slightly helpful but also just expensive can-this-be-done rabbit holes that ultimately don't make things easier. Surely there is a significant cost and even risk involved in using processors that are anything less than the most powerful per watt or whatever else you'd want to optimize for. I imagine if you design for "must never be effected by radiation no matter what," there are some pretty significant design costs that are incurred.
English
@The_Aug_Man @AdamTAlexander @NASAJPL @NASA_Technology They took regular iPhones to the moon and they worked fine
English
Future spacecraft may be able to “think” for themselves.
NASA is testing a next-gen space processor that can withstand the harsh conditions of space, while improving computing power. So far, testing shows it operating at 500 times the performance of chips currently in use. go.nasa.gov/4wm7QHL
English
@vladtarko Seems like this indicates the US could stand to reduce wealth inequality by increasing the size of its government.
English
More capitalist countries have lower income inequality, not higher. High income inequality is caused by cronyism which goes hand in hand with highly regulated markets. Welfare states also lower inequality, but you need free markets wealth to have generous welfare states.
paleoneoliberal@armsq17
Income inequality is a fake concern. It was only after capitalism had largely eradicated actual poverty that neo-Marxists started rambling about inequality. It’s best to ignore the issue completely, since treating it as meaningful can lead people to very silly and dangerous conclusions (the chart below is a good illustration).
English
@_space_punk_ @allTheYud The eleventh commandment: Thou shall not seek proof of my existence, so that you will always be mired in confusion about who or weather I am, which I prefer.
English
If you meet God or advanced aliens in a dream, remember that it would be trivial for them to give you proof, even in a dream.
Just tell you a rhyming prosodic English couplet, two lines, whose SHA256 hashes collide for the first N bits, N>128.
English
@dioscuri @allTheYud If only God-dreamers thought it was God's job to prove his existence.
As a kid I thought God was wrong not to put a detailed plan for a commercial fusion reactor in my science notebook. Believers told me it was reasonable he didn't. I disagreed and gave up belief in a mean God.
English
@allTheYud Dream memory is lossy though, so SHA256 collisions are a brittle test. One garbled syllable on waking and you’ve got a false negative. Thoughts on a test better able to stand up to the lossy/confused conditions of dream experience?
English
@ilex_ulmus @manaltdan Thinking about quitting my job to try and work on stopping AI development. Last tried in 2017 and couldn't figure out how to make a difference. You think it's more straightforward now?
English
@manaltdan Qualified applicants, do NOT apply your talents to protecting superintelligence from regulations. Your voice is so much more valuable when you’re not an AI industry shill. And you will increasingly be under siege in the court of public opinion.
English
Our team at Google DeepMind is hiring! We conduct research, brief senior leadership on AGI governance priorities, and drive initiatives across governance, safety, geopolitics, and strategy. We’re looking for candidates with a research background in areas including AI policy, international security, political economy, compute governance, forecasting, and institutional design. Lots of great opportunities for impact!
Allan Dafoe@AllanDafoe
My team (Frontier Strategy & Governance) at Google DeepMind is hiring. We are bringing on researchers to lead our strategic analysis on frontier AI. Our mission is to provide rigorous foresight and actionable insights to help the world prepare for advanced AI. We focus on concrete, longer-horizon challenges that fall outside the ordinary course of business: forecasting technical breakthroughs, anticipating geopolitical dynamics, contributing to the development of international safety standards, and planning scientific moonshots. We also publish external research and convene experts. Recent public work covers lessons from historical technological revolutions, forecasting frameworks for frontier AI, Cooperative AI, and the limits of model-level governance. This builds on our team's track record of co-creating the Frontier Model Forum, the Frontier Safety Framework, and the AGI Safety Council. We are seeking candidates with deep expertise in domains such as AI governance, international security, political economy, compute governance, forecasting, and institutional design. You will need the analytical skill to translate complex technical and political trends into internal briefings, external publications, and company initiatives. Experience briefing executive leadership is highly valued. London is preferred, with hubs in NYC and the Bay Area. We are targeting mid-career and senior researchers, though exceptional early-career applicants are considered. Apply here: job-boards.greenhouse.io/deepmind/jobs/…
English
Adam Alexander ⏹️ retweetledi
Elon means xAI's original setup and structure weren't optimal for the scale we're aiming for (understanding the universe, powering advanced AI like for robots and beyond). So the team is now rebuilding everything from the foundations up—team, processes, integrations with Tesla/SpaceX—just like Tesla did early on with design iterations, production pivots, and near-restarts to nail the core before exploding in growth. It's classic first-principles reset for bigger success.
English
@Citrini7 Nah, looking back in 2027 as the AI atomizes us to build Von Neumann probes, we'll wonder how we were so convinced of our society's durability in the face of a superior intelligence.
English
JUNE 2028.
The S&P is down 38% from its highs. Unemployment just printed 10.2%. Private credit is unraveling. Prime mortgages are cracking. AI didn’t disappoint. It exceeded every expectation.
What happened?
citriniresearch.com/p/2028gic
English
@peterrhague Since when was culture directed by what we want?
English
Why do we want this? Modern compilers are extremely efficient. Having human readable code as a record of what is done is surely a good thing?
Dustin@r0ck3t23
Elon Musk thinks coding dies this year. Not evolves. Dies. By December, AI won’t need programming languages. It generates machine code directly. Binary optimized beyond anything human logic could produce. No translation. No compilation. Just pure execution. Musk: “You don’t even bother doing coding.” Code was never the point. It was friction. A tax we paid because machines didn’t speak human. AI just learned fluent human. The tax is gone. Now plug that into Neuralink. No syntax. No keyboard. No screen. Musk: “Imagination-to-software.” Thought becomes executable. You imagine an outcome, the system architects and compiles it into reality instantly. We’re not automating programming. We’re erasing it from existence. The entire profession collapses into a thought. Decades of training reduced to irrelevance. The gap between idea and instantiation hits zero. You don’t build anymore. You imagine, and it materializes. Not incremental progress. Total phase shift. The way humans have created things for ten thousand years just became obsolete. Welcome to a world where the limiting factor isn’t skill, resources, or time. It’s whether you can picture what you want clearly enough for a machine to birth it into existence.
English
@boazbaraktcs @adamascholl I'm happy to stop believing doom is likely as soon as someone demonstrates that they've solved alignment for ASI. Until then, I'm not very interested in the optimism of people working on it, as I've seen no progress of note and the default outcome is bad.
English
@adamascholl I am grappling with this possibility. I think that particular outcome is highly unlikely, but there are many ways in which AI could go wrong. This is why I am working on alignment.
I would not conflate pessimism with action or optimism with inaction.
English
Wrote this for young people worried about AI.
lesswrong.com/posts/fwQburGD…
English
@tszzl Bah! People leverage existential anxiety on both sides of the argument for/against AI development. If you're giving someone advice that you wouldn't if you thought their position was right, it's clearer to just tell them why they're wrong.
English
it is best to proceed as though you will be okay. existential anxiety is not a virtue. productive action in a complex world is not correlated with existential anxiety
Boaz Barak@boazbaraktcs
Wrote this for young people worried about AI. lesswrong.com/posts/fwQburGD…
English
@historigins Silly propaganda. Forests and animals lose little from fire. People suppress it as part of their destruction of forests and other landscapes.
English
This animation from 1932 is a masterpiece. It was the first cartoon to win an Oscar.
English
@leecronin Sometimes you find out you're wrong by looking back and seeing that you were the only one using a word a certain way. No one else wants the definition of thinking to be mostly about substrate. That's not how we've been using the word, and it won't be going forward.
English
Current silicon machines do not think. The assertion that they do or ‘it’s different’ is just a category error.
English
Prove to me you’re an old gamer in one sentence. I’ll drop a like if I’m convinced.
English
@jacobmparis Seems intuitive that anything smart and flexible enough to do language processing at a human level couldn't be deterministic.
English
Why are LLMs non-deterministic? What stops it from traversing the neurons in the same order given the same input each time?
English
@knudrete @allTheYud No one's claiming the danger vector exists yet, merely that it is coming soon enough to care about it now. However, I hear the AI's are already doing some significant amount of the coding at Anthropic. Seems like a vector for future danger to me.
English
@allTheYud Hav you put in the work to explain how language models specifically as opposed to your imagined AI systems that are still as far away from existing as ever will become dangerous?
English
Over all human history, we do not observe the rule, "Nobody who warns that a bad thing will happen is ever correct." Nor yet is there any rule that people who issue warnings are always right. Put in the damn work to understand the specific case or go home.
English
As a decent writer, it feels scary to hear people say they can tell the difference between AI output and human output because I don't know exactly how they've become convinced. If they've really just discriminated between bad writing and good writing, and the AI writing is generally good, while human writing, even "professional" writing, is often bad, then I'm personally screwed.
English
ai detection is not very hard and nobody even really tries except @max_spero_
Theo - t3.gg@theo
I don't think the general public is going to accept that "ai detection" isn't really viable
English
Not stupid, no. But for people who have put time into the the question, suspending prior belief, there may be a question like: Who is better at coming up with a compelling interpretation of the world we observe. "Is it more likely that the history of religion says something about a human phenomenon or a metaphysical one?" Does perhaps have one answer that's more reasonable than others. From either side, the other appears to be worse at interpretation.
English
@ibrycecrawford Building consensus amongst humans about what's worth becoming and working towards that consensus. Also the pursuit of quality and beautiful experiences.
English
If you’re an atheist what is the purpose of life?
English