With no particular authority (or endorsement), just what I’ve used:

Something I’m grateful for.

An update on a (usually work related) project I care about.

A cool program I learned about recently.

A policy that really would make life better.

What a particular emotion feels like to me, where it happens, and an image that fits it (use a list of emotions from somewhere).

What I spent the most time on that day, and how I think my relationship with it is.

A haiku (or 3) about something that happened today.

Game design ideas

My data plan on my phone expired rather suddenly (my phone was too old), and I just didn’t get a new phone number/data plan (wifi + wifi calling still worked).

It was nice having unplugged transit time.

I do those too! That’s where the ideas for new architectures, datasets, and training tweaks come from! Math is fun, and it’s fascinating that math can talk sometimes.

Edit: And I see now that we’re editing messages after people reply? Rude, no? Designing a hallucinating machine certainly doesn’t rot your brain.

I think in a non market economy I would still work on language models. It’s cool that a machine can hold a conversation.

This is from 2020; It seems like we might have more recent data and there’s been some shifts?

This is the 2025 report from the same group: https://www.nationalsurveyreligiousleaders.org/s/NSRL-report-2025-clergy-in-america.pdf

They say (page 28):

Evangelical clergy, by contrast, stand out as especially conversionist, with 82% agreeing that it is important to try to persuade people to join them. Only 35% of mainline clergy agreed that such conversion attempts are important, compared to 41% of Black ministers and 52% of Catholic priests saying that. Consistent with their more ecumenical views, mainline clergy are less likely than clergy in any other group to agree that it is important for them to try to persuade people in other religions to accept their religion instead of the person’s current one, though the differences between the mainline percentage and the Catholic and Black Protestant percentages are not statistically significant at the conventional level.

Same question in the new report is here; seems like it’s from the same data round though? So that’s a bit confusing:

There is an additional question, on how this varies for ‘primary’ ministers vs others on page 77; feels like it should be broken down by religion first, but I haven’t looked closely.

As I watched the pepfar drama, I’ve become more sympathetic to the overhead fees of a good charity. How the charity works internally isn’t that relevant, compared to how much good gets done per dollar donated (eg, if $1 feeds 10 people in an area and no other charity can do better, I’m not all that sad if the ceo got 90 cents; they did good work even if they aren’t very charitable as a person).

But yes, there are probably grifts.

Housing and food for your communities.

I’m going to make a mildly stronger claim. I think this game really is quite moddable by a non-coder. What you need is to implement a different ruleset with new win conditions; everything else can be done with copying existing files into the correct file structure. New win conditions are specified by a pretty boring JSON file, docs here:

https://github.com/yairm210/Unciv/blob/master/docs/Modders/Mod-file-structure/5-Miscellaneous-JSON-files.md#victorytypesjson

See here for an MVP for a mod of this type (probably replaces/strips away too much, but you should be able to find the vanilla files in the github linked in the OP):

https://github.com/yairm210/Unciv-minimal-base-ruleset/tree/master

Which is all to say, this is much easier than doing address lookup imo.

By testing it out in the app?

I’ve also tried getting AI to program really simple things, like using js to find particular elements in a webpage (which I don’t control and involves far too many lines). It did fine.

It’s not ready for commercial use, but it makes hacking around unfamiliar code more accessible.

Kozy asked for a different rule set; essentially changing a few numbers related to non-combat victory (shorter research times, lower policy points required, etc). Identifying these numbers in a complicated code base, especially for a non-programmer, could be very difficult. For the non-programmer, understanding how the code works isn’t very important. You just need to know what to change, and perhaps make sure you don’t change more.

I think this is exactly a case where getting a novice programming friend to make a mod would make sense. Equivalently, to vibe code.

That sounds like something easily modded; like a couple of integers somewhere. It would be cool to do (and seems vibe-code accessible if a model can hold the full script in context?)

You can be selective with this power; works well for a lot of folks. Have a smallish in group where you’re always upstanding, enjoy all the benefits that our tribal brain craves, and also enjoy the material benefits.

Some folks would have easier access to their drugs though. Pre war on drugs might have some benefits

Thank you!

I think it’s kinda interesting still, in that it shows people are (must?) drive so much. But yes, agree that per cap seems like the wrong statistic for any kind of safety.

Maybe more fair would be transit fatalities per mile traveled (any method)?

downvoters: is it wrong?

The students read Tolkien, then invent their own settings. The judge thinks this is similar to how claude works. I, nor I suspect the judge, meant that the students were reusing world building whole cloth.

I would love to see the source on this one. It sounds fascinating.

I agree. It really doesn’t look like AI is the thing that broke. More like the education system, or something about social media.

As a civil matter, the publishing houses are more likely to get the full money if anthropic stays in business (and does well). So it might be bad, but I’m really skeptical about bankruptcy (and I’m not hearing anyone seriously floating it?)