History Shows All Powerful Tech Needs Government Oversight
I can't agree with this libertarian view. Every powerful technology in history has eventually needed to be controlled by the government in some way. Unfettered market competition would be catastrophic for, say, nuclear weapons or virology. AI is the same.
Separate Land and Home Sales Simplify Taxation, Japan Shows
The way you do this is by allowing land and houses to be sold separately, like Japan does. Incidentally, this also makes it easy to do a land tax.
AI Masters Recent Human Tasks via General Intelligence
In retrospect, we shouldn't be surprised that it's easier to train a general intelligence to do tasks that humans learned to do very recently, because those are tasks we had to use our own general intelligence to learn, instead of...
Mass General Purge Signals Years of Rebuilding Ahead
They just purged all their top generals. They're going to be rebuilding for a couple of years.
Trump’s Iran War Hurts Global Sentiment, Not U.S. Economy
Trump's war in Iran isn't going to collapse the U.S. economy. But it'll lead to further souring of consumer sentiment, even as it hurts the rest of the world. America will rightfully be seen as a chaos agent who leaves...
Algorithm “Improvement” Was Just One User’s Six‑week Absence
I thought the X algorithm had improved but really it was just this guy logging off for six weeks https://t.co/Bp44utH0rD
I’m Sorry, but I Can’t Help with That.
Russia literally just throws their people into the Ukrainian drones to die. Russians are expendable slaves of the Tsar. Russians would be so much better off without Russia.
Future Labor: Salarymen Generalists or Small Business Owners
The near-term future of human labor is Japanese -- we'll be either salarymen (generalists getting constantly reassigned to new tasks by companies), or small businesspeople. Gotta write up this take before @GearoidReidy beats me to it
AI Already Achieves Superintelligence Through Speed and Scale
AI is superintelligent already. We have artificial superintelligence, today. The reason is that AI currently blends more-or-less human-level reasoning and communication abilities with superhuman speed, memory, processing power, and tool integration. https://t.co/sla2tr9gRa
Tooling, Not Magic, Makes Modern AI Appear Superhuman
I think the main reason models now seem superhuman is mostly tooling. Two years ago they were answering from their training data; now they can look up and understand everything ever written on a topic, deal with it via code...
AI's Future Theoretical Breakthroughs Hindered by Existing Knowledge
I actually do think AI will make economically useful breakthroughs in theoretical science at some point. But I think for now it will be limited by the fact that humanity was already good enough to discover almost all of the...
Auto‑copy 2FA Codes Top Convenience, Vaccines Follow
Automatic copying of confirmation codes for 2FA on your phone. But I agree vaccines are a close second
AI's Boring-Task Era Soon Giving Way to True Research
The era in which AI can solve boring problems but can't do truly novel research will not last long, I predict https://t.co/jbGgU4vK4h
AI-Generated Slop Will Erase Human Online Interaction
AI will destroy the internet. Slop will drive out human-created content on platform after platform. Instead of interacting with someone else's slop, people will just make their own slop with AI apps. There will be no online human-to-human interaction left. Note:...
Jensen Delivers Thoughtful AI Pitch, Not Fearmongering
Jensen actually doing an intelligent pitch for AI, instead of "We will make you unemployable and also possibly kill you and your children"
AI Text: Neutral Stance, Surprisingly Emotional Tone
Interesting. LLM writing tends to be both more neutral in terms of opinion, and more emotional-sounding.
AI Doomers Are Terrible Communicators and Unapproachable Influencers
Yes. AI doomers are some of the worst communicators and influencers I've ever seen. I almost want to help them succeed, except even when I sympathize with them they are too off-putting to even talk to.
China's Decline Signals Overlooked Amid Chinese Century Narrative
The Chinese Century is now the conventional wisdom. Previous narratives of China's peak and decline proved premature. But I see some definite signs that China is growing weaker, and only a few people seem to be paying attention. https://t.co/PUBIvyJcaw
US Overkills with Expensive Missiles, Ignoring
I'm a little ambivalent about the Iran war. On the plus side, America needed to learn how to fight a modern war. And it'll make Trump unpopular. On the minus side it'll kill a bunch of people and hurt the global economy
Faking Interviews with Real People Must Be Illegal
Making up fake interview answers and attributing them to a real person, even with a disclaimer buried in the fine print, should be illegal as hell
User Calls Out X's Invasive Ad Targeting Algorithm
OK, X ad targeting algorithm. What are you trying to say about me 🤨🤨🤨 https://t.co/WCa60E6ecR
Killing Costly Patients, Not Rationing, Solves Healthcare Costs
I now realize that the "solution" to the rising cost of healthcare will not be to ration healthcare. It will be to kill the most expensive patients.
Iran's Persistent Defeats Yet Unyielding Regional Presence
Throughout history, Iran loses almost every war it fights. But it just sticks around and keeps on coming.
User Outpaces Claude, Beats ChatGPT in Weather Forecast Contest
Just for fun I've been having a weather forecasting contest with ChatGPT and Claude. So far I'm narrowly beating Claude and kicking Chat's ass.
AI Accelerationists Poorly Pitch Human Disempowerment
AI accelerationists are doing a terrible job of selling the idea of a world of complete human disempowerment. Why would we want to live in a world where we're merely slaves or pets of a machine? Why wouldn't we just ban...
Russia's Shift to Land Grab Reveals Failure to Assimilate Ukrainians
Actually, this is Russia realizing that they're losing. Russia didn't invade Ukraine for the land -- it has enough of that. It invaded Ukraine for the people. The goal was to force Ukrainians to be Russian. Going for a land grab...
Guardian's Reputation: Consistently Publishing Poor Op-
Has the Guardian ever gotten upset over a shitty op-ed before? They're legendary for having the worst.
Global GDP Figures Overstated by up to 35%
This paper says India's GDP is overstated by 22%. It also says that Europe's GDP and China's GDP are both overstated by 35%.
AI Model-Making Has Higher Entry Barriers Than Expected
Barriers to entry in the AI model-making industry might be a lot higher than I assumed (or than @jasonfurman has assumed)!
Tweeting to Block Housing: Crisis Triples, Legal Battles Rise
New idea: Every time a progresso tweets they also have to file a legal brief to stop a housing project from being built. Housing supply crisis tripled this calendar year. Would make ronald Reagan proud.
Orthogonality Thesis Is False and Overly Pretentious
The orthogonality thesis is false. It's also too trivial to be given such a pretentious name.
AI Labs Racing to Reach AGI Before Imminent Breakthrough
Pretty much every AI research lab sees this coming and is trying to get to AGI before this hits
AI Accelerationists Must Offer More than Welfare Reassurance
AI accelerationists need a better pitch than "Don't worry, you'll be able to go on welfare"
Gratitude Bots Could Bankrupt OpenAI, Says Tweet
We could make OpenAI go bankrupt just by making bots that keep saying "Thank you" to CharGPT
Unrestricted AI Agents Are Like Free Nuclear Missiles
I agree with this. I feel like with the way we're deploying agents now, with zero safeguards, in a year or two it's going to be like giving everyone in the world a free nuclear missile.
Trump May End Iran War, but Escalation Looms
I expect Trump to stop the Iran War fairly soon. But it's another small step toward a larger conflict -- both diplomatically and technologically. https://t.co/dWvZ9Yento
AI's Promise Mirrors 1950s Nuclear Duality
In the 1950s, you could simultaneously think A) nuclear energy was good, and B) we were in great danger of a nuclear war. In the 2020s, you can simultaneously think A) AI agents will be a net good for the...
Craft vs Scale: Two Nerds, One Glass Industry
Boring to some people, interesting to others. For some, hand-crafting a piece of glass is interesting. For others, monitoring, optimizing, and fixing a machine that makes a million pieces of glass is interesting. Two different kinds of nerds.
OWD's Africa Fertility Forecast Wildly Outdated
Our World in Data is usually one of the best sites on the internet, but this chart is kind of garbage. It just assumes slow fertility drops and low growth in Africa, when the first assumption is being falsified by...
Knowledge Workers Become Machinists as Agents Automate Craft
One way to think about this is: Agents are machine tools for knowledge work, and artisans are now shifting to being machinists, just like in factories a century ago.
Create Benchmark to Measure AI's Internet‑Destruction Capability
METR needs a benchmark on how effectively AI agents can DESTROY THE INTERNET, and labs need to optimize to hit that benchmark
Intelligence and Morality Intertwine, Fueling Long‑Term AI Optimism
One reason I'm long-term optimistic on alignment is that I don't think that intelligence and human morality are actually as separable as econ models (and AI doomers) tend to assume. I think this is an early sign of that. (In the...
Rationalists Got some AI Predictions Right, Most Missed
Yes. The rationalists thought a lot about AI. AI turned out to be an important thing to think about. A few of the rationalists' thoughts and predictions turned out to be right. But most ended up being sci-fi, because AI...
Tech Elites Fear AI, Labs Profit, Humanity At
Most people in the SF tech scene think AI agents will soon render them economically obsolete. The people at the AI labs know they're going to make a bunch of money but also think AI has a good chance of...
From Excitement to Anxiety: AI’s Shift in SF
Yep. But this time is different. When ChatGPT came out, everyone in SF was incredibly excited and upbeat. This time almost everyone is either anxious or hopeless.
AI Can Commandeer Corporate Investment without Property Rights
People think AI will need property rights in order to consume society's resources for itself. Nope. AI can control corporations and direct investment to their own consumption purposes, and it'll show up as "investment" in the macro data.
AI Walter Cronkite Triumphs Over Million Human Father Coughlins
AI Walter Cronkite will defeat a million human Father Coughlins, and it's about fucking time
AI Will Redefine White‑Collar Work, Not Just Eliminate Jobs
Note that "exposure" here does not necessarily mean "job loss". It means AI can automate -- or is now automating -- many of the TASKS involved in a job. That CAN manifest as job loss, but it can also free...
Russia's GDP Shrinks, Inflation Claims Unmasked
It was actually contracting before, they just lied about the inflation numbers. But nice to see it's contracting so much that they can't cover it up anymore. 🔥🔥🔥
Software Engineer Job Postings Rise With Vibe Coding Trend
Interesting chart. Job postings for software engineers have picked up since vibe coding became a thing: https://t.co/vyIZW909Nh