Ypermythos

@BenShindel Black markets

Can someone (genuinely) explain to me why we can't just keep raising taxes on "bad" things? It feels that we really do not tax "bad" things enough. The federal tax on cigarettes is only like 15%!? Even the highest federal+state combo taxes on cigarettes are <100%.

@eigenrobot Cool story bro. But why no vegetables??

well i've finally figured out why this makes people angry i hadn't realized until now how many people "enjoy" food in the same way they "enjoy" travel, which is to say "instrumentally and hollowly as a mechanism for reinforcing their fragile self-conceptions"

@rubenhassid Workers definitely know what’s best. That’s why they invented the mechanical loom, cotton gin, steam engine / railroads, tractors, ATMs, PCs and self checkout! Oh, wait… 🤔🤔🤔

The future of work isn't about humans vs. AI. It's about humans WITH AI.

BREAKING: Stanford just surveyed 1,500 workers and AI experts about which jobs AI will actually replace and automate. Turns out, we've been building AI for all the WRONG jobs. Here's what they discovered: (hint: the "AI takeover" is happening backwards)

@martin_casado Like if Plato couldn’t discover the secrets of the universe through thinking alone then why would people expect an LLM to do it..?

@martin_casado Interesting but seems slightly misleading / burying the lede. Isn’t the important conclusion that we need verifiers in the RL loop, which seems consensus AFAICT?

Another excellent post by prof Misra on the hard information theoretic limits of AI, including: - limits to long term agentic behavior - limits to long term reasoning - limits to self improvement

@repligate By the time I’ve processed what you’ve revealed the tweet is old so I figure you won’t check it

Why do I get so little interesting engagement on my tweets? Many are widely liked/shared and could open some of the most interesting discussions about AI. But it’s very rare for me to get one substantive comment. This definitely makes me tweet less. What’s wrong, chat? scared?

@justalexoki They were retarded already, people just didn’t notice

generally sane people are legitimately going to lose their mind over things like this, this is just the beginning

@nikitabier Also his biggest achievement this year

@jdluk87 @grok - is this guy fucking with us?

Apple engineers are perfectionists. We obsess over every tiny detail you'll never notice. When building offline Siri, we spent weeks testing edge cases like airplane mode handoffs, battery drain scenarios, and voice recognition in noisy environments.

@Dan_Jeffries1 Steelmanning this: it’s to build support for UBI so the masses can have some kind of safety net.

I think what irritates me the most is folks think they say fear mongering BS like this to "get people ready." Really? How is witting an article "getting people ready" for 20% unemployment? How would anyone get ready for that? What would the actual practical goal be to stop it, prevent it, mitigate it? UBI? Some other plan? What is the actual fucking plan if you truly believe something like this and simultaneously work at a place looking to make it happen? Absurd bullshit.

If you actually believed this then you'd be morally bankrupt for working at a company looking to make it happen. That leaves only a few actual reasons for saying something like this: 1) You believe you're a part of the few, specially chosen, wise people who should have this power and can guide it fairly (you're not and you can't) 2) You want daddy government to come in and give you a monopoly through discrimatory lawfare because see number one 3) You're unabashedly evil So which is it?

@nickkokonas They already exist and call themselves robo psychologists - @YeshuaGod22

I do think there will be ‘AI Psychologists’ in the near future, for what that’s worth. What I do feel for certain is that most creators don’t yet realize what is in their hands. Those who think AI is ‘stealing’ forget how they themselves learned their craft.

Thoughts on AI (warning: long thread). I’ve been formulating these ideas for a long time…. not months, decades. But only recently has the acceleration made the immediacy of them relevant to almost everyone. They are also very incomplete. But here goes.

@ziv_ravid @grahamu Thanks, makes sense. But do we know that humans’ internal representations naturally reflect human typicality?

Good point! We measured something different: not whether LLMs can answer "is a robin more typical than a penguin?" (they can :)), but whether their internal embedding spaces naturally reflect human typicality - with typical items geometrically closer to category labels. The correlations were weak. While LLMs correctly answer typicality questions through reasoning, their embedding geometry doesn't mirror the graded prototype structure humans use. They achieve impressive performance through different representational strategies than humans.

You know all those arguments that LLMs think like humans? Turns out it's not true. 🧠 In our paper "From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning" we test it by checking if LLMs form concepts the same way humans do @ylecun @ChenShani2 @jurafsky

@LRudL_ What happens to the people who are incompetent CEOs?

In the longer-term, as AIs get closer to a Pareto improvement over even augmented humans: we should work towards control & ownership over AI being widely diffused. As G. K. Chesterton said: the problem with capitalism is that there are too few capitalists

Most reactions to the impending AI automation of the economy are: (1) denial, or (2) simplistic patches like UBI But the modern social contract is based on states & companies needing human labor. @luke_drago_ and I outline a more robust way forward in a Time op-ed.

Finding #2: The Problem But LLMs struggle with fine-grained semantic distinctions. They can't capture "typicality" - like knowing a robin is a more typical bird than a penguin. Their internal concept structure doesn't match human intuitions about category membership.

@burkov @PandeGG What if they actually did a 4D chess by saving enough money on hard drives so they could buy deepmind? 🤔🤔🤔

@PandeGG That they almost lost their business because of mediocrity in the top management. They only managed to survive the LLM disruption thanks to DeepMind.

Smart people at Google created infinite storage for the world's photos to train models on them. Then a mediocre guy came to power and decided to save on hard drives.

@mlmfrisco69439 @romanhelmetguy Gotcha. Would-be genociders should make sure to eliminate 100% of their target group in order to avoid punishment. Seems legit

@romanhelmetguy I made it clear in my original comment that a sizable percentage of the oppressed peoples need to exist for reparations to be paid. It is not about punishing the oppressors it is about creating an equitable playing field.

What happened between 1453 and 1492 that makes Turkey legitimately Turkish but the USA stolen land?

@bryan_johnson @priyanshudotsol How else is he supposed to get enough money to become the 2nd most measured human..?

@priyanshudotsol It may feel heroic but you're spending down your human capacity. That posture reduces brain oxygen by 30%. The blue light suppress melatonin, fragments deep-sleep, blunts next-day insulin sensitivity. The missed sleep raises cortisol, increases visceral-fat and decays memory.

It's 4 AM guys, but builders are building. What's your excuse

@bryan_johnson Death by priapism

I guarantee I'm going to die in the most ironic way possible. I hope you all enjoy.

@MattWalshBlog Agree wholeheartedly w the rest but this part is too collectivist for me

The worst part is that the window seat passenger thinks he has the authority to control the shade without consulting the rest of the row. That window belongs to all of us. You can’t shut off our view without asking first.

I don’t understand people who keep the window shade down on a plane. We’re 35 thousand feet in the sky. You don’t want to check out the view? We’re above the freaking clouds, for God’s sake. Humans could only dream of a view like this for millennia. And you have the shade down.

@dalsault @MarioNawfal

@MarioNawfal I don't know how many people will want to be regularly hauled around like luggage though. And when many millions of people are using self-driving we will get a realistic idea of why it is not a good idea.

🇺🇸TESLA FULL SELF-DRIVING IS A MIRACLE Tesla’s Full Self-Driving isn’t a feature—it’s a technological miracle. Elon has delivered what the experts said was impossible: real-world, real-time autonomy that navigates cities, highways, roundabouts, and chaos—flawlessly. It doesn’t assist the driver. It replaces the driver. No other company is even close. While others play catch-up with demo videos, Tesla’s cars are already driving themselves—safely, smoothly, and better than most humans. This is the greatest leap in transportation since the invention of the car. Elon didn’t just raise the bar. He rewrote the future. Source: @Tesla , @SawyerMerritt