The thing is, we already have advanced refactoring tools. If I want to change a function signature and update it everywhere, that's a solved problem, as least in Visual Studio. It's right 100% of the time and I don't need to check its work in excruciating detail. Why would I ask an LLM to do it?

I think Novo Nordisk investors should be rightly angry with the management who let it lapse. But whatever.

You are correct. When asked about falling birthrates, almost everyone says the problem is financial. But that argument has a big hole through the middle of it: people had far more kids when people were a lot poorer in both absolute or relative terms. What people really need to raise kids is space. If you want to crash the birth rate, pack people together in apartments and high density housing. Apartments are fine for the first few years after you're out of school and building up some savings while working on your career, but then people need space to start a family. All these billionaires complaining about birthrates need to understand that efficient housing is the antithesis of 2.1 children per woman. Efficient housing should be inexpensive, so you can save enough for a down payment on a real house.

Musk would start up a company around this to produce more of his children since he can't find women willing to help him with that anymore.

You would be shocked how easy it would be for him to find women willing to have his babies, even now. Women have high standards, but "lots of money" beats most of them.

I think that a slow decline in population over many generations would be a reasonable goal. I worry more about a sudden decline, especially after such a large group is heading into retirement across the west. We're going to go through a period where there's a lot of elderly people who need care, and that's going to suck up resources, which leaves less resources for everyone else, and that naturally leads to a drop in standard of living. There's a huge difference between slow or even flat growth, vs. a declining economy. Human psychology reacts very violently to any perceived decline. That's why leaders are always obsessed with constant growth... it keeps people content. Decline causes violence and fighting over increasingly scarce resources.

There's a big difference between the Blue Origin flights that just pop you up above the Karman line to technically be in space and experience a ballistic trajectory for a few minutes of weightlessness, vs. going on a trip to the ISS, which is in orbit. The delta-v required to reach orbit is an order of magnitude more. This is definitely real astronaut stuff, especially if he has a job to do once he reaches the ISS.

Translation: unless you're watching tons of reaction videos for the newest marvel movies (bleh), or those videos where a journalist gets a dozen successful women directors to sit around a round table and talk about how difficult it is to be a successful woman director, you'll never get a date.

If they're accurate, right? Like... you actually care if the technology can generate accurate summaries? Because at this point there's lots of evidence that AI-generated summaries contain lots of mistakes.

It's so weird that so many people are ignoring the massive accuracy issues of LLMs and have this misguided idea that you can just trust the output of a computer because... well, it's a computer. It's literally using random numbers in its text generation algorithm. Why not just use astrology?

Wait, the US intelligence community is putting a bunch of their files through an unproven technology that's renowned for making mistakes, and they're just going to trust the output? Time to get the popcorn!

There are professions where your job is mostly bullshitting: politician, journalist (specifically writing editorials), sales, fiction author, HR, middle manager, etc. Those jobs are going to be impacted, or in some cases, replaced by LLMs. But those people (in particular the journalists who write these articles) seem to think that everyone else's job involves just as much bullshitting as theirs. But there are lots of jobs where being correct matters: engineering, technicians, warfighting, surgeons, construction, etc. The danger is that people are going to try to use LLMs to do these jobs and it's going to be dangerous.

I've realized that LLMs are not "artificial intelligence." They frequently make incorrect statements that exude confidence and authority. They're "artificial bullshitters."

It seems like if you're so poor that you can't afford to buy groceries in cash, you should consider using the food bank. That's what it's there for.

If LLM coding is just a better stackoverflow, then that's hardly going to replace programmers. Typically if you increase the ability of programmers to do their jobs efficiently, then you increase the demand for programmers. That's how the Jevons Paradox works. But this is just an incremental gain at best. Not a huge improvement.

AI-assisted coding has been around for a little while now, and the gains people are claiming are absolutely massive. Yet where's the evidence? How come I haven't seen any meaningful change in any applications that I use on a daily basis? Where's the guy cranking out big new popular apps at an astonishing rate and just raking in the cash? The answer is simple: it's a scam. These claims are being made by people who are trying to get investment dollars, or who've received investment dollars and need to continually reassure their investors that the money didn't go to waste. Don't buy into it. At the very least demand some evidence.