To my astonishment, I've actually finished my first pass of converting the entire back end code base of one post wonder to modern async/await syntax. What does this mean for you? Absolutely nothing, except that it will be easier to maintain in the future, and I will hopefully be better able to get others to consider assisting in that task.

I'm making other changes as I go along that aim toward maintainability, and just making it hard to write bugs in the first place.

I reached a point where all I'm doing is testing and debugging, testing and debugging, until it actually works again, and I can think about pushing it as the new current release.

After that, it will be time to think about refactoring the front end code. Refactoring is a fancy word for rewriting, or at least restructuring. All I have to do is choose a front end framework so the code will be more maintainable for other contributors, easier for everyone to understand. All I have to do is choose a framework. Like, insert a list of at least 1000 possible choices. Of course, there are just a few well-known ones. If I go with React, people will know that word 10 years from now when they try to contribute to my code. This does not necessarily mean that what is called React in 10 years will look anything like it, but they'll know the word.

Seriously, I can see that I write software with very long release cycles. This makes me an excellent candidate to maintain the software stack for the interstellar multigenerational probe. Just slow me down and run me as a low priority process.

Taking an early upgrade slot rather than waiting for Linode to do scheduled maintenance at 5am.

I'm back to work updating OnePo's codebase to modern standards on the server side. Specifically async/await as opposed to callbacks. it's BOOOOO-RINNNNG. But I have "only" one file left. Of course, it's the one where most of the code is.

So after that I should refactor that file as well. Also BOOOOO-RINNNG, and yet somehow satisfying. Then maybe crank up eslint to call me on my shit. Poor man's typescript. Then maybe typescript?

Then rethink the frontend code. Using [ugh I will not even think about choosing a framework today].

Of course OnePo is still a tiny, tiny place. As in "I've never bothered to index the database collections for performance" tiny.  I may need to dial up the disk space just a tad soon though.

"But Tom, aren't you gearing up to welcome former Twitter residents?" No, not really, although the six people who will actually leave in this, the most recent of many outrages are warmly welcome.

Moderation of a successful public social media site is a huge, unsolved problem. As others continously have pointed out, the real challenge for something like Twitter is moderation, not engineering. Yes, it's bad that Musk's plan is basically to fire the moderators.

But I'm in no position to improve on that. There's no moderation in practice on OnePo at the moment, although I'm pledged to supply some in the event of certain types of abuse.

What we have instead is an invitation-only model and a predilection toward thoughtfully locked posts, which reduces (though it absolutely does not eliminate) the likelihood that someone you're connected to needs moderating... at least from your perspective. So far this seems to work OK for our needs.

But, a system like that cannot function as a public square. If we indulge my ego in imagining OnePo writ large, we wind up with a worse echo chamber than Twitter, because it's so difficult to discover new people and the general habit is to limit audiences for things.

Also, OnePo is just a whole 'nother idiom. The intersection of OnePo and Twitter is a ridiculous idea. That would be... uh... BeReal, actually. BeReal is pretty fun. But it does one tiny job.

Actually I kinda like the idea of doing one tiny job. OnePo's one tiny job is allowing friends to know and support each other better. That's fine for what it is, which is a lot.

Midjourney created this image, given the prompt "white cat named Spike in the style of Marvel." If you've seen my cat, and if you haven't you must be new here, this is pretty amazing. I picked my favorite of the four initial images and upscaled it. I didn't do any further refinement.

Not a real photograph. Dall-E created this image, given the prompt "a big white cat named Spike with stripes like a raccoon and spots like a cow photorealistic."

My privacy-minded partner thought at first this was a photo of our cat.

So Dall-E is just better, right? It depends. The above was Dall-E's best effort on the same Marvel Spike genre prompt Midjourney nailed so well. I attempted to refine this with the words "Marvel cinematic universe" and so on, but it didn't get any closer to the concept.

Midjourney created these images from the prompt "Midjourney's mother." I didn't pick a winner or do any further refinement.

It is difficult not to read into this. But is it just telling me what I want to hear? And if it is, so what?

Midjourney created this image from the prompt "the true purpose of Midjourney." Three of these can be understood simply as riffing on the word "journey." The one at upper right is perhaps just riffing on images frequently associated with the words "true purpose." So as lovely as they are, they don't make me go "hmm" as much as the previous set.

"Dall-E's mother photorealistic" yielded this image, and other images of animal mothers. One was a chimpanzee. If they were all chimpanzees, it could be considered a little cheeky. But emphasizing one chimpanzee out of four images would just be cherry-picking on my part.

Dall-E generated these images for the prompt "the true purpose of Dall-E." There's a theme, at least for three of them, but if there's a message here this monkey isn't clever enough to figure it out.

A Google employee who claims their text-generating AI is a sentient being has been fired. On the whole, I tend to agree with those who say that particular AI (not Dall-E or Midjourney) is probably not a sentient being. And yet, I also agree with those who have serious questions about our ability to judge that. This will be more of an issue as the line gets fuzzier. And it's getting fuzzy pretty fast.

This was inspired by my own need to know about new chargers that pop up, primarily on the route to my mom's house. When you drive an EV, PlugShare is absolutely essential for checking out chargers to make sure they are legit and have no recent problems, but I think there's room for an independent newsletter that lets you know when new chargers pop up on your path.

Also: how the hell did I get this domain name? Thanks to my privacy-minded partner for thinking of it.

Only of use in the US, sadly, since I'm dependent on Department of Energy data.

(Yes, I need to work on my sharing pic game. A picture of the same text that's in the description... ouch.)

Some small OnePo improvements and fixes:

• Making links to other pages the traditional way, via the "link" icon and the dialog box, works again on the iPhone.
• You can also make links by just pasting a link to any page straight into your post, without clicking the link button. When you do that, a nice automatic preview and link appears. We've had this for ages, but there was a dumb bug that broke it for most cases, except YouTube. All fixed. Here's a nice example (and a cool game):

You'll note I didn't have to create any content for that, I just pasted the address of the page straight into this post. This is based on the "open graph" tags that websites already provide for the benefit of Facebook sharing.

• Speaking of the automatic link feature: pasting a URL like that in the middle of your text once again splits the text into two sections, with the link preview in the middle.

As I mentioned, you can still make links the old-fashioned way too.

(At this point I don't care that I'll never be able to buy the domain onepo dot com. I'm still going to call it that, because how can you not?)

Boo: I have paxlovid rebound. I am “screamingly positive” for COVID, much like in this article. It happens.

Yay: the rebound symptoms are mild, as promised. Think “severe allergies.” No one has yet reported serious symptoms during the rebound, and more importantly paxlovid reduces hospitalizations by 90% during the initial course of COVID.

I have to stay isolated for another five days and retest. I’ll live! (Literally)

If you don’t have any risk factors they say you probably don’t need it. I took it because of my asthma and history of difficult flu. That may well have been a super-great decision. I don’t have a control Tom. 🤪

🌠 for anyone who finds themselves facing the “paxlovid or not?” question.

My favorite bit about this story is that the light from this star began its journey 13 billion years ago, but the star is now 28 billion miles away, because the universe kept expanding and we kept on moving apart from one another during those 13 billion years.

Only that's not right, because the star died aeons ago. Stars like our sun live 10 billion years; huge stars like this one are too fast to live, too young to die; they wink out in 10 million years.

So it's more accurate to say the former location of the star is now 28 billion light years away. And no longer part of the "observable universe," because light from where that star is today will never reach us. 28 billion light years = 8584 megaparsecs (yes, I asked google). A galaxy 4,300 megaparsecs away from us is receding from us at the speed of light. Our relative velocity is therefore close to twice the speed of light, which is possible because it's not the velocity of either object relative to a point that isn't moving. And that means light from there will never reach us again.

[Waves goodbye, realizes it's a bit late for that]

Hey gang, sat down to write a post today and pictures were busted; sorry about that. Post tomorrow, got us a fix for uploading pictures today. Enjoy.

Work is currently underway on support for adding alternative text to any image, whether in a post or in a comment. For comments this involves a bit of a design overhaul, so it may be a Minute.

Hi gang,

I saw a couple of friends coping in a DIY way with a possible bug today (tough to say at this point), and it prompted some thoughts about how little attention I've given the platform in a long time. My last push to the server was... cough... two and a half years ago.

In a way this is not such a bad thing. Hey, it works!

But there's a lot that could be improved, and long-running bugs that haven't received attention (notifications for your own replies, anyone?), and believe it or not I'm still interested in the question of whether I did everything I could to make it easy for folks to onboard and, yes, grow the site.

But of course there's very little time.

Meanwhile... recently in another context I was facing pleas to do more in a space that's outside of my more essential responsibilities from day to day. I am very well supported at work. But as in any workplace there are gotta-haves, wanna-haves and really-oughta-haves and it can be tough to juggle them.

So I struggled to organize this really-ought-have stuff as a Project. But a Project implies a Budget and Time and there is no budget or pre-scheduled time for that particular thing right now (darn those pesky gotta-haves). That meant everything I was writing felt like a recipe for stress.

I felt stuck, and anxious.

Then I noticed a kanban board feature in our internal wiki and deliberately created an itty-bitty kanban board to track progress in that particular arena.

When I was asked if I wanted to turn it into a Project instead... I said "nope!" Because the medium is the message. Project = Milestones. Schedule. Release Dates. Time Pressure. All of which are appropriate when resources are available, and needlessly stressful when they are not.

But a kanban board says "hey... this is the art of the possible! Nibble at it. Bash at it. Occasionally really go at it. You'll get there."

My kanban board at work turned out very well and I'm still using it to nibble away at the oughta-have thing I can only work on occasionally.

So in the same spirit, I plan to set up a kanban board for One Post Wonder.

And today, in that kanban spirit, I did three itty bitty things to move the ball forward:

1. The "submit a bug" button just sends me a damn email. It used to post an issue to my github project, which was too clever by half. I didn't get any notifications, I didn't pay attention and you probably felt neglected. I'm sorry about that. This should be better.

2. I fixed an issue that caused gmail to reject some emails from One Post Wonder. There are services you can pay to deliver your emails reliably. Cost aside, I feel it's important that we try to maintain independence in this area. Even though it's a pain in the ass and an ever-moving target.

3. I changed the welcome message you see when you sign up from "each day, post a carefully curated sonnet concerning the state of world affairs or you will be punished with lasers" (*) to this:

"It doesn't matter if it's a cat picture, a duck joke or a column worthy of Alexandra Petri."

I think this is way more welcoming, and I hope it will encourage people to feel like it's OK to just chill here. Lord knows I do. Shitpost away, everyone. You just can't do it more than once a day.

(*) It didn't really say that. It was far more well-intentioned but still kinda intimidating.