2025-09-29 20:55:53
My game Person Do Thing is now available in the US; it'll be out on Amazon soon, but if you want to get early-adopter bragging rights you can order right now on want.persondothing.com for shipping this week.
First, a caveat: I love coming up with new voting systems, but I'm very unconvinced that (in practice) this is at all a good idea.
Long story short, I tend to believe that the theoretical properties of a voting system are conditional on their legitimacy with the population, such that the best voting systems in practice are ones that are 1) familiar, 2) very simple to explain, 3) don't require maths. For real life, I largely approve of approval voting.
I also think that without a very rigorous plan and a willingness to spend your life implementing it, talking about voting systems is a lot like doing a crossword – it's a fun puzzle that vaguely mentions current events, but it has ~no actual connection to politics in the real world. This is plausibly true for most public policy discussions, but might be especially true for voting reform, because voting reform is almost-always against the interests of the current government who were voted in by the current system. There are all kinds of easy improvements that don't get implemented (cough anti-gerrymandering cough cough) because the people who decide about them are exactly the beneficiaries of the current messes. So I don't believe that writing blogposts like this has any chance of changing the world as it actually is.
But for the rest of this piece, I'm going to pretend I don't believe those things, so I can have fun thinking about voting systems.
In short: Weighted FPTP would be a parliamentary voting system where each representative gets votes in proportion to their vote-share in their district. If Anna wins her district with 60% of the vote, she gets 60 votes in parliament; if Billy wins his district with only 42% vote share, he gets 42 votes.
This has some properties:
1) it mildly counterweights gerrymandering. Gerrymandering usually works by putting Our Voters into districts with small majorities, and Their Voters into one district with a huge majority and several others with small minorities, therefore "wasting" a bunch of their votes. Under Weighted FPTP, this effect is lessened (but not removed!) because the huge-majority district delivers more votes to parliament than a small-majority district.
2) at the same time, because it's still FPTP, this system gives disproportionate reward to the winners: a candidate who gets 51% of the vote gets 51 more votes in parliament than the runner-up who got 49%. But this is both good and bad, because truly proportional systems can become dysfunctional when no faction has enough votes to get anything done.
3) it can encourage politicians in safe seats to keep working for their constituents despite knowing they're going to win anyway – even more so if the vote share used is "% of voting-eligible population" rather than "% of people who actually voted."
2025-09-26 20:31:17
This week’s episode is best paired with a hot cup of coffee and Wild Ways by Josh Ritter playing in the background. Last episode was too LLM heavy, for which I apologize. I’ve done my best to group all of that into LLM corner so as not to let it overshadow everything.
It appears we have evidence for life on Mars.
"any study of Internet culture is basically a study of crazy people”. Also: "Be careful who you pretend to be, because you are who you pretend to be.”
Dwarkesh’s advice for explaining your announcement / launches / blog posts for Twitter.
PSA: How to fold fitted sheets, via the Browser. I sent this to my wife and she very nicely said to me something to the effect of “isn’t this the same way I taught you to do it?” 🤣
“When outsiders succeed, it’s usually through reframing problems in ‘paradigm shifts’. They benefit from not being too attached to existing theories.” From a thread on outsiders solving problems.
Uri says we should not allow 18 year olds to sign long term contracts. So, so many thoughts here.
Our parenting hack of the year so far is having cut vegetables ready at the table when our kids get home from school. The percentage of vegetables consumed is up like 10x and compliance to the routine of coming in, washing hands, and sitting down at the table has risen as well. Recommended and thanks to Emily Oster for the suggestion.
German chocolate cake was invented in United States, via the Kroger App. Someone needs to figure out why the Kroger app has so many delightful facts in it. This is someone’s passion project! I'll buy you a nice bottle of wine if you find this person and introduce them to me.
80% of Swiss are satisfied with their lives. I am not sponsored by the Swiss government, but I am open to it if they are reading.
The Pangolin is the only mammal with scales. Peacock is the name of the males only; the female are peahens. The species is called peafowl. Via The Animal Book.
There’s no such thing a quality time with your kids. My mom said this to me over and over again as child. It’s quantity of time, not quality of time.
The rise of parasitic AI. This is the first moment where I’ve seriously contemplated the AIs taking over.
ChatGPT Is Blowing Up Marriages as Spouses Use AI to Attack Their Partners
“She does that to her family. She does that to her friends. She does that to me,” he lamented. “She doesn’t seem to be capable of creating her own social interactions anymore.”
I worry a lot that the sycophancy of the agents have made me less flexible with people who (of course) are less likely to defer to me. I am not sure how to measure this, but I wish I could.
Sort of a musing, but I think we owe Blake Lemoine and apology.
How to Claude and Claude Code Camp. I want to be on Claude Code Camp.
The Pope says we won’t find God in the AI.
If you are good at code review, you will be good at using AI agents. I wonder what it would look like to teach editing as a skill. Is there anyone that does this?
A promising approach to prompt injection attacks.
I can’t wait to experiment with Net Dollar.
2025-09-24 23:43:36
I've never met a human being – even (especially?) people who I've known deeply and for a long time – and thought to myself, "yeah, I'm confident I can explain why this person does what s/he does."
And yet, some big subset of biographies and magazine profiles seem premised on the idea that if you know a person's history you can meaningfully explain their major life choices. When you think about it, this is really weird.
A defining fact about a lot of modern interview podcasts is the status difference between the host and the guest.
For most of the interview podcasts I listen to, the host is much lower status than the guest – just an eager kid who invited a minor celebrity to come talk their book. As a result, the interviewer is deferential or obsequious, and the podcast is worse: it is just not fun to listen to someone suck up to somebody else, and without being pushed the interviewees say fewer interesting things.
Some interviews are the other way round: most of the guests on Joe Rogan are much lower status than he is, and it shows. I think Sasha Chapin wrote somewhere that Joe Rogan's big trick as an interviewer is feeling comfortable and status-equal to anyone, so when he has super high-profile guests it works pretty well – he'll talk to any president or nobelist as a conversation of equals. But for most of his guests, being on Rogan is just a massive opportunity, and they feel it, so you get the same over-deference problem but the other way around.
I suspect that in the legacy-media days, a big part of The Trick with interviews was that an institution could lend part of its credibility to an interviewer: maybe the guest was a bigger deal than the host as a person, but the station the interviewer worked for brought some institutional status as well.
I think the best conversations happen when both parties are at a similar status level, and (most importantly) have some fluidity in their status throughout the conversation: it's much more fun when you don't know from segment to segment who's going to play the upper hand.
2025-09-22 23:15:17
For New York’s No 1 Socialite, A.
1) Prioritize your ease of being over any other consideration: parties are like babies, if you’re stressed while holding them they’ll get stressed too. Every other decision is downstream of your serenity: e.g. it's better to have mediocre pizza from a happy host than fabulous hors d'oeuvres from a frazzled one.
2) Advertise your start time as a quarter-to the hour. If you start an event at 2:00, people won't arrive till 2:30; if you make it 1:45, people will arrive at 2:00.
3) Invite a few close friends to come 30-60 mins earlier to set up / eat dinner with you / hang out / whatever, so that when the start time approaches you’re already having fun instead of stressing that nobody will come.
4) Most people will only go to a party where they expect to know 3+ others already.
5) Use an app like Partiful or Luma that shows the guest list to invitees. Start by inviting your closest friends, get some yesses, then expand from there.
6) Send the invites in chat groups (or visibly cc’ed emails) to clusters of 4-5 people who know each other, so they can see that their friends are also going.
7) When inviting people individually, namedrop mutual friends who are invited or coming.
8) In a small group, the quality of the experience will depend a lot on whether the various friends blend together well. Follow your instinct on this, even if your instinct feels rude. It’s like cooking a dish, two ingredients can each be fabulous and still not go well together.
9) A large party is more like an Everything Soup: you mainly need to avoid ingredients that ruin the flavor for everyone else; beyond that you can mostly throw in whatever and see what works.
10) Regardless, try not to feel bad about not-inviting someone if your heart says they would make the party less-fun for others. Make peace with gatekeeping because if you don't exclude a small % of people you will ultimately lose everyone else. Someone can be a good person and a bad fit for your party, so don't think of it as a judgement on their soul. All of this is easier in theory than in practice.
11) Most events are better when roughly gender-balanced. Prioritize inviting people of the gender you’d likely have fewer of, then top up invites with the other. Once an event crosses a threshold (maybe 70%?) of male-or-female dominance, most people of the other gender are likely to decline (or just not-come to your next party) as a result. So there's ultimately two equilibria, "roughly gender balanced" and "extremely uncomfortably unbalanced," and you need to stay in the attraction basin for balance. To do this, keep your invite ratio at worst 60-40 in either direction, in order to prevent a downward spiral.
12) Co-host parties with someone you like a lot but who isn't in your exact social circle, so that your two friend-sets can intermingle.
13) Figure out the flake rate in your social circles (the % of people who will RSVP yes and flake on the day), and set your invite numbers with that in mind. In my circles, consistently 1/3rd of people who say they will be there will actually not.
14) Couples often flake together. This changes the probability distribution of attendees considerably, and so your chance of losing a quorum in a small-group setting. Small-group couple-events (e.g. 3-4 couple dinner parties) are very hard to manage in a high-flake society, as a result.
15) Create as much circulation at your party as you can. People circulate more when standing than when sitting, so try to encourage standing for those who can e.g. by having high-top tables, or taking away chairs from around tables, or leaving shelves and counter-tops open for people to rest their plates and drinks.
16) Put the food in one part of the room and the drinks in another, or spread the food and drinks out around the space, so that people have lots of excuses to move around the room.
17) If someone arrives at your party and doesn’t know anybody, welcome them and then place them with another group or person. Ideally you can pick someone they’d specifically get along well with, at second-best just someone who’s friendly and easy to talk to, but ultimately you can just insert them in any group that’s nearby and open. The main point is to prevent them having to butt in on strangers themselves, which for many people is mortifying, while your Host Privilege allows you to do it for them.
18) To leave a group conversation, just slowly step back and then step away. Don't draw attention to your leaving or you’ll be pulled back in. It feels mildly weird to do this but it’s worth it.
19) Throughout the party, prioritize introducing people to each other and hosting the people who are new or shy, even at the cost of getting less time hanging out with your best friends yourself. Parties are a public service, and the guests will (hopefully) pay you back for this by inviting you to parties of their own.
20) Let me repeat that: Parties are a public service, you’re doing people a favor by throwing them. Someone might meet their new best friend or future lover at your gathering. In the short term, lovely people may feel less lonely, and that's thanks to you. In the long term, whole new children may ultimately exist in the world because you bothered to throw a party. Throwing parties is stressful for most people, but a great kindness to the community, so genuinely pat yourself on the back for doing this.
21) The biggest problem at many parties is an endless escalation of volume. If you know how to fix this, let me know.
2025-09-15 23:15:00
This is a dumb topic for a post but it keeps bothering me and I want there to be a canonical link explaining this fallacy.
Here's my latest example: someone online referenced the Latin quote Vox populi, vox Dei – "the voice of the people is the voice of God."
Another person replied that in fact the full quote says the opposite:
|
Nec audiendi qui solent dicere, Vox populi, vox Dei, quum tumultuositas vulgi semper insaniae proxima sit. |
And those people should not be listened to who keep saying the voice of the people is the voice of God, since the riotousness of the crowd is always very close to madness. |
I see this kind of thing and think, "oh, how interesting!" – it scratches a kind of puzzle-itch in my brain, there is something magnetic about reversals. I think other people feel the same, because the Full Quote Dunk gets a billion reposts.
But then I look up the source and discover I've been lied to: that longer text is in no sense The Full Quote, it's simply one guy's critique of an existing popular phrase. Vox populi, vox Dei was a well-known saying; someone felt that it was incorrect; in order to talk about it he inevitably had to write it out; but the fact that his quote-plus-criticism is longer doesn't make it Full.
Now, Full Quotes don't have some magical truth status; it may well be that the original saying is wrong and the critique is right. In this case the guy doing the critiquing is Alcuin of York, advisor to Charlemagne and "the most learned man anywhere to be found" in the 700s. I am not even the most learned man to be found on this blog, so who am I to say whose vox is whose?
But the people dunking don't actually believe X or not-X based on what the Full Quote says; they just believe not-X for separate reasons,1 and then repost anything that agrees with their preexisting beliefs, especially if it's in Latin and makes them feel smart (but I repeat myself).
Populi: this is bad! Bad arguments are still bad even if the thing they argue for is correct. As the great Alcuin of York said, those people should not be listened to.
[^1]: most likely reason being that their opponents believe X, of course.
2025-09-12 23:05:00
I guess the fact that I'm back means that y'all didn't hate this. If you do, tell Uri and Jehan!
Matt Holden and I are doing a YouTube show about building with AI called --dangerously-skip-permissions. The first episode is “How did we get here?”. Matt and I have been having 1:1 conversations for more than a year now about what tools we’re using and how we’re using them… and now we’re having those conversations in public. I especially enjoy the way that Matt is able to connect what’s happening with LLMs today with previous eras of computing innovation. Give it a listen if that’s your thing!
Matt Holden on Markdown coding
OpenRouter has market share by LLM model. Interesting and unexpected in some ways!
On fact checking with AI. I really enjoyed this one. I have a draft blog post in my head called “Vibe Craft: How to do serious work with AI” but every time I try to write it, it falls flat. This is spiritually related to that.
Office building visits are up among people that live less than 5 miles from their office. As someone who made major life changes during the pandemic, I feel the pang of regret.
“A soldier will fight long and hard for a bit of colored ribbon.” — Napoleon
“Everyone has a plan until they get punched in the face.” — Mike Tyson
Get the original Good Tokens and other excellent items at jdilla.xyz