thomascgalvin

Full Self Driving determines that it is about to strike two pedestrians, one wearing a Tesla tshirt, the other carrying a keyfob to a Chevy Volt. FSD can only save one of them. Which does it choose ...

/s

Writing a text editor is like writing a compiler; the point isn't adoption, the point is in the learning

> The modern Apple feels like their hardware teams way outperforming the software teams.

There aren't a lot of tangible gains left to be made by the software teams. The OS is fine, the office suite is fine, the entertainment apps are fine.

If "performance" is shoving AI crap into software that was already doing what I wanted it to do, I'd rather the devs take a vacation.

There are a lot of people who prefer to fight from guard.

I use Mermaid and such for a lot of technical documentation, but this seems like it's going to be much more straightforward, especially for quick and one-off diagrams.

Very nice.

For me, an important distinction is whether or not a human is reviewing the edits suggested by an AI.

I toss all of my work into Apple Pages and Google Docs, and use them both for spelling and grammar check. I don't just blindly accept whatever they tell me, though; sometimes they're wrong, and sometimes my "mistakes" are intentional.

I also make a distinction between generating content and editing content. Spelling and grammar checkers are fine. Having an AI generate your outline is questionable. Having AI generate your content is unacceptable.

My experience with GenAI is that it's a significant improvement to Stack Overflow, and generally as capable as someone hired right out of college.

If I'm using it to remember the syntax or library for something I used to know how to do, it's great.

If I'm using it to explore something I haven't done before, it makes me faster, but sometimes it lies to me. Which was also true of Stack Overflow.

But when I ask it to so something fairly complex on it's own, it usually tips over. I've tried a bunch of tests with a bunch of models, and it never quite gets it right. Sometimes it's minor stuff that I can fix if I bang on it long enough, and sometimes it's a steaming pile that I end up tossing in the garbage.

For example, I've asked it to code me a web-based calculator, or a 3D model of the solar system using WebGL, and none of the models I've tried have been able to do either.

While this is generally true of a certain American political party, you see hard cutoff lines even in states like Massachusetts, where the majority of voters and politicians seem to be actively trying to make things better for people.

The real answer is that gradients are hard, and clear lines are easy. A shocking number of Americans don't understand how our income tax brackets work; they believe that if you cross the line into a higher income bracket, your entire income is taxed at that new, higher rate, and you end up losing money overall.

Massachusetts, which has the highest percentage of college educated citizens in the entire country, also shot down ranked choice voting last year, because the math was too hard.

That's certainly part of it. Another significant factor is that many people have come to the conclusion that they simply cannot afford children. It costs something around $30k to give birth in the US, the cost of housing grossly outpaces inflation, wealth concentration is at or near an all-time high ...

Many people are struggling to keep themselves afloat, and adding a family to that mix would be ruinous.

The issue with GenAI is that is typically sounds correct, and presents its opinion in a way that seems authoritative, but might be wrong... and the only ways to know that it' wrong are 1. to try its suggestion and fail, or 2. to look at it with existing expertise.

I see this frequently when generating source code; it'll get me 80-90% there, but I need to fix, massage, and adjust. This is fairly easy, since the code-test-code loop is already part of a normal development cycle.

When it's a matter of law, though? You don't really get those kinds of do-overs, and if the judge catches you citing a hallucination in your argument? I doubt that would go over well.

I'm still running an Intel-based MacBook Pro from 2019, and it's trucking along just fine. It's not as fast as the M* chips, obviously, but it's still perfectly serviceable for software development. I'm tempted by the new Pro, but I don't consider it a necessity.

If I picked up a new Mini or MacBook today, I'm reasonably confident it, too, would last me five years or more.

> Tell someone in the 1600s about the concept of cellphones and they would say it’s impossible to ever instantaneously communicate with someone around the world.

I don't know that they would.

In the 1600s, they'd just invented the printing press, so communication was already undergoing a fairly massive change. A book that once took a lifetime to copy could now be printed in three days.

From there, it's a short walk to "this book that used to take a lifetime to copy can now be printed more or less instantly." And from there, we can start laying the groundwork for what we would eventually call telecommunications.

Cell phones rely on technology that people in the 1600s hadn't discovered yet. Interstellar travel would require us to not only discover new technologies, but technologies that actively violate what seem to be pretty hard physical laws.

If the speed of light really is a universal maximum, and something like an Alcubierre warp drive isn't possible, leaving this solar system is vanishingly unlikely.

In Kotlin, an infix method can be called without the `.` operator, and without the parens surrounding arguments.

    100 money "USD"

is the same as

    100.money("USD")

`Number.money(currency: String): Money` is an extension method; Kotlin provides syntactic sugar which allows you to "add" methods to existing classes. The above is basically equivalent to:

    fun money(count: Number, currency: String): Money

> It's unprecedented and it's hugely beneficial to our species.

"Hugely beneficial" is a stretch at this point. It has the potential to be hugely beneficial, sure, but it also has the potential to be ruinous.

We're already seeing GenAI being used to create disinformation at scale. That alone makes the potential for this being a net-negative very high.

Plus bonus compensation for the loss of skills while you aren't working.

Right to work is 100% anti-union. In reality, it's "the right to work without any protections."

The idea used to sell it is that you don't owe your employer any loyalty, and they don't owe you any loyalty; either of you can walk away at any time.

In reality, this creates a completely asymmetric power dynamic. You can be fired at any time, for any reason, which means you have little power to advocate for better compensation, safer working conditions, and so on.

I generally don't even like syntax highlighting for Markdown. I find having the text jump around because it became bold or italicized distracting, and it's even worse when a header is suddenly a larger size and everything below it is pushed down the screen.

Markdown is simple, and I like working with it simply.

I suppose there is some value in having the more complicated aspects, like images and tables, checked for syntax so you can see if you've made a mistake, but I run a macro to generate those anyway, so I'm fairly confident in the result.

I have used markdown to write books, and it worked just fine. Now, these were novels, not technical documents, but I never encountered something that couldn't be solved with markdown plus some occasional HTML.

I've also generated quite a bit of technical documentation using markdown, and Pandoc's extensions[1] allow you to include pretty much any formatting you might require, complicated math and syntax highlighted code blocks included.

That markdown can then be transformed into HTML, a Word doc, a ePub, a PDF ...

There needs to be a very compelling reason for me to reach for something other than markdown.

[1] https://pandoc.org/MANUAL.html

Recidivism is a feature, not a bug. Like most institutions, America's criminal justice system is bent towards self-perpetuation, not problem solving.

I kind of think we have proven that most people are not, and cannot be, capably responsible for their own retirement. The 401k was an experiment, and that experiment has largely failed.

The answer is probably a more generous and better funded Social Security, except that fund keeps getting raided for other purposes.

There are only so many ways to write "C, but better," or "C++, with fewer warts."

Any languages that want to be recognizable to people comfortable with the C-lineage are going to converge on a lot of points.

I agree that a lot of these components are more distracting than useful. I do like the animated border around buttons, though; that would be very useful to draw the eye towards the "default" option on a page, as long as it was the only one.

The Wim Hof method is explained here:

https://www.wimhofmethod.com/breathing-exercises

Java has always been conservative about maintaining backwards compatibility and introducing new keywords, but the java.* and javax.* packages have been growing larger and larger since day one.

I do think Go and Kotlin have kicked the Java community into high gear, though. A lot of the recent (or upcoming) features are direct answers to features found in those languages.

A Link to the Past remains one of my favorite games to this day. I consider pretty close to a perfect 2D adventure game.

Inbox was a legitimate improvement on "normal" email, and I still miss it.

My initial reaction was also "if I can talk to your software, that's an API."

Reading through the article, though, the author isn't advocating for "no APIs," they're advocating for minimalist APIs that use agent detection to "do the right thing."

Chrome tells the API that it can accept HTML, so the server sends data formatted inside of a web page, with an HTML table.

Curl doesn't send that header, so the server sends unformatted CSV. But you could send an Accept header to get the HTML if you wanted.

The benefit of this is that for most use cases, this will "just work."

The downside is, if you want to view the CSV in a browser, or the web page in curl, you need to know (or guess) how the server is deciding what to send you, and take the correct action. An API documented with OpenAPI, while more complicated, explicitly tells you what you can do.

This article is basically one long "be careful, that hammer can hit your thumb."

While that's true, it doesn't mitigate the hammer's usefulness, nor does it propose a better alternative.

Technology is hard. Good tools make it easier, but they never make it trivial. You still need to know what you're doing.

I would vastly prefer a couple of boring years.

I 100% understand why veterinarians are in a crisis.

We had to let our pupper go a few weeks ago. The vet had to basically sit there and watch us while we processed the fact that we were about to pay her to kill our best friend.

For us, it was the worst day we had experienced in years. For her, it was Tuesday. She had other people waiting in the room next door, and had to go from solemn to bright and cheery over the span of ten footsteps, and she has to do that every day.

Her job is often to put a very real price tag on the life of a beloved companion. "I'm sorry, but keeping him alive will be $10,000... or we can humanely put him down for much, much less."

It's grim business. Necessary, but grim.