Zanni

The thing that makes harming others wrong is the lack of consent. That allows us to avoid the sticky question of what is "harm." For self-harm, consent implied, what is "harm?" Exercise with a risk of injury? Failing to exercise, with a risk of decreased life span? Failing to optimize your exercise routine? Surgery? Drinking alcohol? Dating the wrong person?

This isn't unique to code or AI. In creative writing courses, we were asked to give thoughtful critiques of (human written) stories and excerpts, and often I felt as if I were doing more work than the original author. If you can't be bothered to review your manuscript, or at least run it through a spellchecker, why should I waste my time on it?

I wish this pope would stay in his lane. The only opinion I want to hear from him is whether AI can be baptized.

I used to feel this way, but I reconsidered my threat model. You know what format is "locked in"? Physical books. Can only exist in one location at a time. If you loan it out, you can't read it until it's returned. Subject to theft, fire, rot, bugs or simply being lost.

There are aspects of Kindle I don't love--the constantly changing cover art for books I've purchased--but I've never run into an actual problem. I've got 2,500 books on my Kindle devices, and I can access them anywhere in the world at any time on my dedicated readers, my phone, my laptop (via Kindle Cloud Reader).

If DRM is the price I have to pay for a dead-simple ecosystem, multi-device support and free cloud storage, well, I guess I'm happy to pay it.

Exactly. I'm happy to see NASA get back in the game, but this is basically just a test flight. I'll get excited for the next one.

("Hey, kids! We're going to Disneyland! We're going to drive all the way around it before we head home!")

That video is 54 minutes long. Maybe, if it's the basis for your argument, you could post a summary?

This article [1] claims a "3-line improvement" on the vision chart. It wouldn't be helpful for me (moving from 20/600 to 20/300, I'd still need correction) but would be a huge win for someone in the 20/50 to 20/80 range, correcting to 20/40, the minimum for driving without correction.

[1] https://www.ophthalmologyadvisor.com/news/yuvezzi-approved-a...

Electric cars make no sense.

Reusable rockets make no sense.

Autonomous cars make no sense.

Data centers in space make no sense. <--- You are here.

Humanoid robots make no sense.

Fred Lambert at Electrek has become an unreliable reporter on Tesla. I don't know where his bias stems from, but it's unmistakeable.

Lambert claims David Moss couldn't find a Robotaxi ride without a safety driver. True enough, there aren't a lot of them on the road. That doesn't mean they don't exist or were "pulled back."

Here's Moss's latest tweet on the subject: "ANOTHER UNSUPERVISED:

2 in a row now here in Austin, TX of Tesla Robotaxi’s completely autonomous without chase cars.

You can see 2 cars in the clip with no one in it & anyone can book it!

Ride 59 complete." https://x.com/DavidMoss/status/2016939137031381487

Note also in that SAME ARTICLE, Lambert credits Moss with driving 10,000 miles on FSD v14 with zero interventions.

To paraphrase William Gibson: full self-driving is here, it's just unevenly distributed.

Valid point if your ratio is correct, but I suspect it's the other way around.

If the cost per ride is 35% cheaper than Uber/Lyft, then that money stays in the local economy, just in the hands of the consumer rather than the gig worker. Currently WayMo is more expensive, but I see that changing as they scale. And certainly CyberCab is promising to be much cheaper.

Jhonen Vasquez was previously best known for Johnny The Homicidal Maniac, an insanely dark graphic novel. I'm still stunned that the powers that be thought he should write an animated children's show. That said, I love Invader Zim. Taquitos!

Peter Lik has a strategy sort of like this. It's still a limited edition of, say, 100, but the price increases as the edition sells out. The last print to sell may be 100x or more the first.

That lost shot (Falcon 9 transiting the sun) is my favorite. I've got a print of it in my office, waiting to be hung on the wall.

He wrote several more books (including a sequel/second edition of Code Complete). None of them had the same impact, though: https://stevemcconnell.com/books/

Carefully distinguishes between PYREX and pyrex. Immediately launches a discussion of Pyrex.

Why your [ultra-light hiker] friend suddenly has [the world's lightest] power bank.

I remember Colin Fletcher, years ago, writing in The Complete Walker about trimming the borders off his paper maps to save weight, which seemed like an insane over-optimization to me. But then, I'm not an ultralight hiker.

I am impressed folks are getting their loads down to 10 pounds though.

Awesome - I didn't know this!

Tesla screens switch to dark mode at night.

Depends enormously on the implementation and use case. My daily driver is a Tesla Model 3, which has a big, beautiful touch screen. But I almost never touch it while I'm driving. Anything I need to control can be handled by voice command ("set temp to 70") or the scroll wheels in the steering wheel. (The one irritating exception is the windshield wipers.)

You don't have to share your phone, or its secrets, to use this. At the TSA checkpoint, there's a screen that says specifically what information they're asking for, you tap, and your phone shares that. You never lose physical control of your phone. No one even looks at the screen. It's basically tap-to-pay for authentication.

The significance of blue books is that, in addition to being bound, they were handed out in class (and numbered) so you couldn't sneak in notes or pre-written pages.

Well, yes, but it's also pretty dependent on what you eat. From the article: "Additionally, the summarized study did not record the timing of snack and water consumption."

A carb-heavy snack before a meal is going to result in a radically different response than eating the same meal on an empty stomach. I'm glad they acknowledged it, but it feels absurd to publish with this headline if they didn't record snacks.

This is an unfortunate stance for science-fiction authors to take. I understand the frustration with generative AI displacing artists and authors work being harvested as training material, but ChatGPT was used here just as a sort of text-processing tool, not to displace any artists or authors, not to make judgements.

Nick Mamatas, an author I admire, condemning ChatGPT as "the Bad Facts and Dumb Idea Lake-Draining Machine of Stolen Books" is particularly disappointing. (In general, the issue of water usage by LLMs seems particularly overwrought.)

I don't understand the point of standardizing language around specific numerical ranges when they could just use numbers.

Makes sense. I like to feel the road when I'm driving, but if the car is driving, I'd prefer the ride to be as smooth as possible.

What professionals are doing this work already? As I understand it, this is something that can only be done by AI (or by shooting the movie twice in different languages).

Abuse. Twitter/X allowed bold text recently for a few days, and a number of monetized accounts immediately started making all their posts ALL BOLD ALL THE TIME to stand out. Feature dropped soon after.

If you have ebooks from other sources, like Project Gutenberg, you can still load them on your Kindle.

Fun, but more suggestive of Steadman (illustrator of Fear and Loathing in Las Vegas and other things) than Pollock: https://www.npr.org/2014/04/27/307285367/artist-ralph-steadm...