Dustin

@Liathetrader Well some people go to the grocery store themselves

Groceries for 2 people. Will last us maybe 5 days. I don’t know how people do it. This country is becoming unbearably expensive. 2k monthly groceries Not counting if you want to go to a restaurant once a week, add another 1k.

@JeremiahDJohns Dude literally changed his mind due to overwhelming data. A real heartwarming story

The YIMBY messaging victory is so total and complete that Antonio Reynoso, the guy who just five years ago was railing about 'overdevelopment', is now talking about how much we need more market rate housing to build a million new units.

@FifthRocket @mateosfo @KelseyTuoc @EN_parker I guess Matthew is saying if you inject only a few newborns and you focus on the ones who didn't need it, then how are you sure you saved lives?

@mateosfo @dabluck @KelseyTuoc @EN_parker Maybe more like vitamin k injection for newborns. A treatment applied to all to save a few. And we don’t measure that at the state level, why would you do that when we have more specific stats?

Traffic deaths have not gone down in any city where Waymo is operating. Zero lives have been saved. That said, we will know in ~ 5-10 years if Waymo is better than the best human drivers (they only release data that compares them to the worst human drivers).

@KelseyTuoc @mateosfo @FifthRocket @EN_parker Cancer doesn't seem like a good comparison. If I treat one person with cancer I save one life. If I take one waymo trip... I didn't save anyone. It's a different category of thing.

@mateosfo @FifthRocket @EN_parker This is an absolutely absurd attitude which I have never heard from an actual social science researcher, for the record. It's like evaluating a new cancer treatment by looking at state-wide cancer mortality. No one ever does this. It's not a sound experiment.

One thing people should realize is that most of reddit is fake now, so there's no reason to get too outraged about bots posting crazy things all day.

@jon_stokes Reddit has been like this for awhile. The Luigi discourse was genuinely frightening.

this seems really not good, I would say

@hecubian_devil Code that is verbose, redundant, and overly self-explanatory is often good code. You want it to be understandable and clear. Being too cute gets you into trouble. Writing prose is the opposite. Redundancy and over-explaining sounds really bad.

I’ve been using Claude Code daily this past week, and its outputs seem very impressive to me, a wordcel who can’t code. But when I see AI prose, I can tell it’s bad. Which makes me assume the code is probably as kludgy and crass as its writing is.

I will never, ever reconcile myself to using tools that have anxiety.

The ruler is beefing with the pope, his legions have fought to a stalemate with Persia, now typhus is breaking out in a major city, oh yeah it’s late antiquity out there

It’s finally April, it’s 78 degrees outside — and it’s officially outdoor dining season in NYC! But we should be able to dine al fresco year-round. We’ve introduced a bill to eliminate some of the overly burdensome regulations on outdoor dining and allow it year-round.

@kduggan16 @AoDespair

Show this to anyone who says speed cameras should be turned off at night.

@antoniogm Well not in Europe for instance

Living in NYC is so great. Where else can you get iced coffee?

@cspan @SecDuffy Guy who loves transit definitely calls it the "metro train in New York City"

.@SecDuffy: "I'm a Republican, and I love transit…You should be able to have your 14-year-old daughter get on a metro train in New York City at 7 o'clock at night and feel safe. And there's no way in hell I would ever put my daughter on the MTA in New York."

@EmilyAssembly I really think it'd be better for the city if elected officials would stop saying things they know are false to get likes on social media.

The private market approach in NYC has been to build luxury housing that no one can afford. In my district, it often sits vacant or houses diplomats a few times a year. If we’re serious about providing affordable housing, we must decommodify it and build social housing instead.

Tell him to pick up the trash.

@luke_pighetti @jeiting @openclaw If I want claude to figure out the larger picture I will it that. But don't worry it doesn't listen and continues to do things beyond what it's told anyway, sometimes even noting it knows it isn't supposed to.

@dabluck @jeiting @openclaw claude code these days will take a simple instruction and figure out the larger picture and implement things not explicitly asked for. this is one of the reasons it's so productive. can't imagining hamstringing that

I've become pretty agent brained over the last 72 hours. Here are some observations. What makes @openclaw different: 1. Self-modification, the fact that it can edit its own prompts, let's you collaborate with it to enhance it and configure it. In the field learning has eluded most tools to date. (also a huge source of security exposure) 2. HEARTBEAT and CRON tools. The fact that it has easy access to both schedule things, and just a general awareness loop. It's what makes it an agent and not just a parrot. 3. The ability to build more agents, the subagent paradigm is really powerful, maybe it's not something that survives the next 10x'ing in context window, but for now allows for some really powerful things, and creates good security boundaries. What it needs yet: 1. More context persistence, there are lots of tricky ways that you lose context. Calling into a subagent, getting a callback from a webhook, etc. You maintain agent level coherence, but any in context from the session that initiated the action is lost. There are work arounds but you kind of have to understand the moltbot architecture to use it correctly 2. It's missing a spark. The heartbeat is good, and there is some temperature in there, but I think it needs a spice mod, that just causes it to do random things once in a while. Humans do this all the time and I think it's a huge aspect of our general intelligence. I think this can be simulated, gonna try some ideas.

@hamishmckenzie You're right that you won't be able to convince me of something that isn't true! Literally the first CTA in every email is trying to break that direct relationship between author and reader. I will be convinced when you stop making product decisions that undermine your whole case

Maybe, going by your tone, you're not that open to persuasion on this, but I'll try anyway: Writers own their audience relationships on a mailing list. They can take that mailing list with them anytime they want. With the exception of Apple's IAP, which writers requested to use and Apple forces Substack to comply with, writers even control their payments relationships in a Stripe account that they can disconnect from Substack any time they want. Readers have full control, too. They can choose to read writers in the Substack app, or they can stay away from the app completely and just read in email or on the web. (This is a big advantage for writers because their distribution isn't limited to only email.) When a reader does use the app, it's better for writers, because app users subscribe to more writers, pay more on average, and can get the latest version of any post in the cleanest format possible with community features right there. No one is forced to do anything. The users have complete control.

Don't leave your audience in the hands of whoever runs TikTok or wherever else. Get your audience on a mailing list. Protect it.

@hamishmckenzie Substack would never ever attempt to co-opt your relationship with your readers except every single time a user tries to read your content. So many examples here are a few: mail.bigdeskenergy.com/p/death-by-tho…

@dabluck This is a common misunderstanding of Substack. There is an algorithmic feed in the Substack app to help you discover writers who you can then subscribe to. The discovery tools generate writer-reader relationships, they don't co-opt them.