justin․searls․co

It’s now been 20 years of fighting with it sporadically and I still can’t figure out how to use launchctl to successfully automate tasks reliably on macOS. Such a frustrating piece of software.

If someone opens a half-baked pull request on your project and you really don't feel like doing the work to get it across the finish line, no reply will shut it down faster than telling them you'll merge it once they add tests.

GPT 3.5 is a lot worse than GPT 4

It should shock no one to learn that Open AI's newer, better language model is an improvement over the old one, but if you aren't an active user of any of this newfangled AI stuff, it can be easy to lose track of just how much better things are getting and how quickly.

If you subscribe to GPT Plus, ChatGPT will also implement the ReAct pattern for requests it thinks can be formalized, which is one way to mitigate hallucinations.

Pictured here, asking "days between 12/10 and 2/11":

  1. GPT 3.5, which gets it completely wrong with a nonsensical reasoning
  2. GPT 4, which gets it right and even has a little terminal prompt link
  3. Opening up that link will actually show you a Python script GPT 4 used to compute the answer

Cool beans. 🫘

Vision Pro with the Good Strap

Following up on my post from this morning on how to use a 3rd party "halo" strap for Vision Pro, my incredible brother Jeremy printed and sanded these adapters for me. The experience wearing this strap is night-and-day better than either of Apple's built-in straps. The headset's weight is finally where it should be.

PSA: This is the first good Vision Pro strap

UPDATE: It works! Photo here

As I mentioned in my review podcast, the two straps that ship with the $3500 Apple Vision Pro are god-awful and mediocre, respectively.

If you just spent that much money on this thing, do yourself a favor and buy two more things:

  1. A BOBOVR M2 Plus strap

  2. This 3D-printed conversion kit for connecting it to Vision Pro (you can also print it yourself)

And boom: for under $50 you'll have a comfortable way to actually use the Vision Pro. Shame on Apple for dropping the ball so badly in the name of aesthetics (what happened to, "design is how it works"?), but hat tip to Mark Miranda for pointing me to this Etsy listing.

Just survived my first flight with Apple Vision Pro. Watched 2 hours and 35 minutes of downloaded video and had 8% battery remaining.

If you've ever paid a significant amount of money for a seat upgrade, consider buying a Vision Pro instead. I literally forgot I was sitting in economy.

One of the strangest things about United Airlines is that booking a premium economy seat (of which there are only a dozen or so on this plane) has zero bearing on your boarding zone unless you buy an extra add-on. It's so bizarre because it results in United being wholly excluded from Premium Economy searches on sites like Google Flights. Feels strange paying twice as much to hang out in Group 5.

Still got it

10 pounds. 19 liters. Vision Pro, MacBook Air, spare battery, and all the clothes and toiletries I need to travel indefinitely.

Still got it.

HTML fragment caching really works!

I have somehow been using Ruby on Rails since 2005 and have never worked on an app that needed to think seriously about web request caching, probably because of my proclivity to reach for static site generators and simple asset hosting whenever anything I make will be public-facing. But the current app I'm working on is actually mostly accessible without requiring users be logged in, which means it will both (1) run the risk of having bursts of hard-to-anticipate traffic to certain pages and (2) render pretty much the exact same markup for everyone.

I'll start with the results. Here's a mostly-empty, public-facing page my basic Heroku dyno without caching:

Completed 200 OK in 281ms (Views: 201.9ms | ActiveRecord: 47.5ms | Allocations: 37082)

And now with a few lines of caching setup:

Completed 200 OK in 9ms (Views: 3.5ms | ActiveRecord: 1.6ms | Allocations: 2736)

So over 30 times faster. And that's on a very basic page. Once the site is primed with content it'll probably be even more dramatic.

Here's how to do it.

But wait, there's more…

Breaking Change artwork

v4 - Facial Computing

Breaking Change

This podcast is a month old and four episodes in and the singular event looming over all of it has finally arrived! The era of facial computing has begun!

Join me for a Vision Pro extravaganza in which I detail all of my first impressions using the device, including dozens of things that seemingly every media and YouTube reviewer missed or excluded. Listen to this podcast and you'll hear tell of bugs you wouldn't believe even if you did see them!

The headline takeaway is: Apple Vision is clearly the future, because it's clearly not yet the present. (And why I'm probably keeping it anyway.)

As always, e-mail me your reviews, reactions, and errata at podcast@searls.co and I'll absorb them into the bubbling stew of opinions I'm forming about this futuristic-and-not-necessarily-in-a-good-way computing platform.

Scant show notes follow:

Show those show notes…

I wonder if anyone has attempted to fork VS Code to run under WebKit instead of Chromium. My M2 MacBook Air should get "all-day battery life" but has never cracked 3 hours and the only running app other than Safari is usually VS Code.

This'll be an insta-purchase for me before it can get yanked from the App Store:

Juno delivers a fully native visionOS UI that taps into YouTube’s embed API, which is designed to allow videos to be embedded in external webpages. When you want to browse YouTube’s video catalog, Juno pulls up a tweaked version of the YouTube website. Apparently, the app is even clever enough to not show ads for YouTube Premium subscribers, though it remains to be seen how Google feels about a third-party developer being in control of an app for one of its biggest services on a new piece of hardware. Selig notes that he didn’t use any private/internal APIs to develop the app.

I'm not one to kink shame, but it's starting to feel like this guy's fetish is building businesses whose survival depends on the mercy of hostile platform holders.

Real beaut of a lede in the WSJ (News+ link) today:

Meta Platforms is hoping Apple’s launch of the Vision Pro can reinvigorate its $50 billion metaverse effort, which consumers have yet to widely embrace.

Love the optimism here, but I wouldn't be surprised if this ages about as well as the CEO of Palm's, "PC guys are not going to just figure this out," quote prior to the iPhone's launch.

Update: this take by Kristopher Browne on Mastodon raises a great point:

I believe them. Goggles are the last device category where Apple didn’t have designs for others to crib from. It’s a space like “smart phones” were in before the iPhone, when android was going to be a blackberry clone because it’s not like google knew how to design a device or interface.

I almost forgot about this mess. I guess the analogy really holds: just like Google had no clue that touch metaphors like "pinch-to-zoom" were going to be the breakthrough interface paradigm for smartphones, Meta (as indicated in this WSJ article) wasn't seriously chasing down mixed-reality computing of the sort implemented in visionOS until after WWDC.