Rogs 🔍🔸

Can I pin someone else's tweet?

@TSA Gen Z social media intern humor tweets don’t work for the TSA. You are a disliked government agency foisted on us post-9/11 and a constant reminder of the impossibility of reducing the footprint and inconvenience of government. Not a beloved burger chain

I'm confused, isn't he saying here that we *should not* burn it all as fast as possible? > “It's, you know, the same reason life exists. It's much better to conserve and strategically use free energy to secure more free energy, and keep growing, and have some order, rather than just like, burn all in one go and have chaos, right?” To me, "trash all resources as fast as possible" sounds like the part he puts after "rather than" that he says is less good, no?

When you start digging into how dumb @beffjezos is, the rabbit hole goes deep… In this recent clip, Beff is claiming that humanity's ultimate goal should be to trash all resources as fast as possible — slash and burn the entire cosmos to make thermodynamic entropy number go up — BUT that's okay because the process of building a universe-scale furnace will supposedly be precisely the process by which we'll achieve our moral values to the fullest! Seriously, he's claiming that these two objective functions are functionally equivalent: 1. Grow our flourishing civilization beyond earth and fill the universe with life, love, happiness, wonder, creativity, discovery, art, humor, cooperation 2. Turn every ordered configuration of atoms into chaos in the shortest possible time Watch 3:02 in the video to hear it from him: “It's, you know, the same reason life exists. It's much better to conserve and strategically use free energy to secure more free energy, and keep growing, and have some order, rather than just like, burn all in one go and have chaos, right?” Read that again. That's the load-bearing argument that licenses him to equate claims #1 and #2 above. He's saying: Since life on Earth built beautiful complexity while using up the sun's energy, we can just go ahead and extrapolate that, into me leading a movement to extinguish all usable energy sources as rapidly as possible. Earth teaches us that the faster we extinguish all usable energy sources, the better we'll grow a civilization even more flourishing than the one on Earth! Too bad that if you think about it for 5 seconds, the fly in the ointment of his logic is that optimal free-energy-dissipation strategies like "chuck all the matter into a black hole" don't conveniently dovetail with sane goals like growing our flourishing civilization to fill the universe with life and love. Beff's Effective Accelerationism movement successfully got thousands of people to put "e/acc" in their profile without checking too much into the doctrine. These mostly normal, psychologically-healthy people just prefer a government that doesn't stifle the many forms of good tech progress… Unfortunately these people became Beff's useful idiots, considering the outcome he's explicitly trying to accelerate is the universe's heat death. I'm not exaggerating or misrepresenting him here. Homie wants to accelerate the universe's heat death. His core claim that objectives #1 and #2 above are equivalent — a claim which is FACTUALLY FALSE if you ask any physicist — somehow never gets questioned (or even flagged as questionable) by the people who have a chance to interview Beff. One podcast after another sees fit to wager their show's credibility on a guest who can dazzle listeners with smart-sounding jargon for hours, as long as no one dares to ask him why his movement's central claim is as factually false as Jonah and the Whale. But hey, if you don't ask your acolytes to swallow a factually false claim as part of your core doctrine, are you really leading a new religion or “hyperstition” 🙄? No, you're just like the rest of us jockeying for crumbs of AI venture capital *without* the reality-distortion benefit of a social media demagogue halo. Nor are you getting your thin-skinned ego inflated by a vocal minority of tech figures who accepted or looked past your dumb central dogma (“#1 = #2”) while everyone else took 5 seconds to think about it and now thinks you're a laughingstock.

@ngpsu22 @larsiusprime Can you give an example of some of the wrong assessments?

@larsiusprime Idk much about these places but I see Baltimore on the website and the SDAT land value assessments are so bad that they are useless. If anything they just make LVT look untenable.

Visualization of land value per square foot across four cities: Tallinn, Estonia Detroit, Michigan Cincinnati, Ohio Austin, Texas

In Japan, market research shows bullet trains absolutely dominate travel for trips under 500 miles. In the US, that’s: Boston ↔️ DC via NYC San Francisco ↔️ Los Angeles Chicago ↔️ Twin Cities Atlanta ↔️ Orlando Chicago ↔️ Toronto via Detroit

@henrytdowling @omarsar0 I was confused too. Like I don't get how *extending* subscription support for Fable is disruptive. It can only help you, no? x.com/ESRogs/status/…

@omarsar0 Dumb q: how is this disruptive to do? Like, if Fable stops being available, can you not just switch over to 5.6 for coding? Does it present planning difficulties to not know which coding models will be available a week in advance?

I don't think Anthropic realizes how disruptive these changes are to users. I appreciate the extension, but please stop playing games. Either keep it under the subscriptions or put it under the API already.

@omarsar0 Sorry, dumb question, but how is this disruptive to users? I've seen others express a similar sentiment, but honestly I don't get it. Is the idea that you want to make sure to cancel your subscription before the next billing cycle, iff Fable is no longer available or something?

@AgustinLebron3 FWIW, I don't think "thing I care about" is what distinguishes the cases. Thought Claude put it well here: x.com/ESRogs/status/…

@ESRogs "I favor less regulation about things I don't really care about, but really favor regulation about things I care about and that I personally benefit from." It's a fairly common position for people to take, whether inside SF or outside.

The naivete is astounding. The history of governance in Western countries in the last 25 years, arguably more, is the steady expansion of government power and general authoritarianism. It's a global phenomenon. If you think this is a good thing, fine. But say so.

@AgustinLebron3 Claude corroborates my claim

@AgustinLebron3 These people don't think it's good in general. A common refrain is, "I generally favor less regulation, except for AI". Did you already know that and you just think it's hypocritical or something, or were you not aware of that fact?

@Laneless_ @zack_overflow Yeah, I've had this problem with Opus too. Had partial success adding a rule against it: x.com/ESRogs/status/…

@zack_overflow Every LLM does this afaict, giving them a scratchpad might help?

Noticing Fable is very often writing code comments that reference details from chatting with it, but they make very little sense in the context of reading it in the codebase

I went to a wedding yesterday between a power-couple central to the AI alignment scene. I got to see ~100 rationalists, EAs, and future-y people who I haven't seen for a decade. The shocker was that the vibe was soooo good and wholesome. Felt like we'd all collectively gone up an adult development stage or something. People who were once fellow lowbies are now running well-respected orgs, but not so self-important that they couldn't get goofy on the dance floor. Lots of kids running around. Those who were once merely "allies" are now also lifelong friends. Beautiful and heartfelt speeches. Mine was about how alignment of future technologies depends on the internal alignment of the people stewarding them, and how I believed that the bride and groom had achieved this alignment. For the first time in a while I felt optimism about where the scene is going. I maintain that the broader Bay Area social milieu is currently cursed – ie dissociative, unrooted, casually sociopathic – but with any luck, this already-powerful group of folks will improve the culture through the merits of their intellect, influence, and heart.

Europe figured out concessions decades ago. Every park has a kiosk: espresso, pastry, maybe a spritz or a beer at 6pm. Small operator pays rent, city gets revenue, park gets foot traffic and eyes on the grass. Everyone wins. SF has 230+ parks and almost none of this. Dolores Park froze a Blue Bottle cart permit over community outreach complaints. Coit Tower took years of hand-wringing over a removable food kiosk (unsurprisingly: Peskin territory). Meanwhile we wonder why maintenance budgets are tight and parks feel empty at 7am. Give me a coffee kiosk in every major park. Let the concession help pay for upkeep. Let foot traffic take care of public safety. Create a third space — a square where people can meet instead of shouting at each other on social media. I’ll die on this hill. Photos from Lisbon, Portugal 🇵🇹

I've previously tried to rebut this argument as a kind of rebranding of arrogance as humility, especially around COVID (see slatestarcodex.com/2020/04/14/a-f… ). As COVID was ramping up, people kept saying it would be insufficiently "humble" to take any action against it, because we hadn't yet "proven" that it could become a worldwide pandemic. As there became a 10% chance, a 25% chance, and then a 50% chance that it would, people continued not to take action, because there still wasn't the "proof", and it would be "arrogant" to prepare for something that merely had a 25% chance of killing millions of people. But this is the wrong way of thinking about things. The people who were implicitly acting as if there was a 100% chance that COVID wouldn't be dangerous, were calling the people assuming a 75% chance that it wouldn't be "arrogant" and accusing them of being "absolutely sure" that it would be dangerous! If you allowed a 10% or 25% or even a 1% chance that AI could be dangerous, I think you would be extremely invested in building the global capacity for a potential off switch - not necessarily in pulling that off switch now, but in making sure it existed and can be pulled on short notice. That's AI2040's current ask - building the technology to verify an agreement if it needs to happen later. It's the only thing that needs to be done before the situation clarifies in 2027 or 2028. I think if you are indeed trying to being humble and leave room for all possibilities, you should refocus on supporting it. Also, we at least have lines on graphs suggesting our position might be right. Where's your argument that regulating chips the same way we regulate every other product would lead to a "global surveillance state"? When it's about blocking AI safety, all people require is hand-waving; when it's about trying to create the capacity for it, no argument, no matter how strong, will ever be good enough.

@RalphSmorra @AlexGodofsky TBH I'm not 100% sure either way. x.com/AlexGodofsky/s…

@ESRogs @AlexGodofsky OP didn't respond like it was a joke.

The hardships Ukraine is creating for the innocent civilians in Crimea are unconscionable. Ukraine should suspend its interdiction campaign until the civilian population can be adequately resupplied and then resume it.

@dylan522p @KevInvestingYT > Not my job to care about stocks. Is this real or no? finance.yahoo.com/markets/option…

@KevInvestingYT Scale up CPO on GPU is delayed as we said. ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯ Not my job to care about stocks. My job is to track the supply chain accurately

When a bi man calls you daddy it hits different

@RalphSmorra @AlexGodofsky See: x.com/mgarbowski/sta…

@RalphSmorra @AlexGodofsky > is a joke I think it might literally be a joke.

When EBR-II scientists reflected on shutting it down, they cried. They had demonstrated a reactor with "walk away" safety, the ability to use nuclear waste as fuel, and just how economical sodium cooled reactors could be. But they had to shut it down. The story in videos: 🧵

This is a pretty neat project. Basically they come up with a bunch of surprising events for every month (mostly celebrity deaths and political events) and ask the LLMs about it, and this way deduce the actual knowledge cutoff. The claimed cutoff is the circled square.