Well, that depends. I don't consider it breaking down because it was still doing what it was designed to do... generate text.

It's very simple. Experts need to continually prove they're worthy of trust. If they do that, people will slowly start trusting them again. If they abuse that trust, people will kick them back out. This isn't rocket science. And I can tell you that there's a rot at the core of academia, and people know about it and are fighting back. We won't put it behind us until we see significant reforms.

It's just a text generator. It's not thinking. People are ascribing thinking to LLMs the same way Catholics see the image of the Virgin Mary in their french toast.

You should first study the long winding path that brought us to this point, to give you perspective, and then you should look at current events as the tiny small footsteps that they are when compared to the long stretch of time. Sometimes we stumble on long walks. When this happens we should avoid spending all our time despairing at the road stretching out endlessly ahead of us, and take some time to appreciate the distance we've already covered. Then we need to return our eyes to the ground before us and allow the stumble to focus our mind on the importance of carefully considering every step.

We happened to go through a period of many decades (perhaps even a hundred years) called modernism, where the populace generally respected experts and expertise to make our lives better. But post-modernism has long challenged this idea of faith in science, expertise, and the enlightenment, and there have been many very public failures of experts to do the right thing over that time. From Thalidomide, to toxic chemicals being dumped into the environment, to the Challenger disaster, to the second Iraq war, and the 2008 financial crash, then actively talking about managing public perception in the middle of the pandemic, and yes, even the Epstein saga, these events have all eroded public trust in the experts and the institutions. The current spate of populism (literally framing all issues as the masses vs. the elites) is a direct result of this loss of public trust. People are simple... they will only believe something that they have an incentive to believe. It's hard to get them to believe that outsourcing all the manufacturing to China, or bringing in lots of foreign laborers is a good idea because it makes the GDP numbers go up, especially when they themselves compete in that low-end labor market. We can regain a sense of modernism, but putting the reins of government back in the hands of trusted experts starts by regaining trust, and we have a long way to go before we get there.

Human progress is hardly monotonically improving. But the general positive trend over time is substantial and impossible to ignore. While we're throwing away many of the substantial gains we've made over the last 75 years (globalized trade, a rules based global order and the relative worldwide peace it provides), we're not likely to completely lose these things, and a few years of seeing what life is like without them will probably change people's opinions in the years to come. Mostly history acts like a ratchet, but sometimes we backslide.

While it's important to keep working on issues of the day, it's also important to look back at where we were and how far we've come. Progress like this is based on small incremental improvements over decades and even centuries. About a hundred years ago, the death rate from measles in the US averaged about 5 per 100,000, which would mean roughly 15,000 deaths per year today just in the US. All those lives are being saved *every year* just due to a single vaccine program. We need to be thankful for the work that prior generations did that are truly amazing and have improved our lives immensely. It's trendy to complain about how bad life is, but that's disingenuous when you realize how bad people had it throughout history. Heck, just look at the infant mortality rate in the US.

I've been working in industrial automation for 25 years, installing industrial robots and manufacturing machines and so on. All of these machines were only ever installed at companies that were growing and hiring more people. I've never seen a machine installed that was followed by lay-offs. Plants only close down because demand dries up, or because they can't compete with the cost of overseas labor. Faster automation is the only way to make American workers compete with workers in China for the same work.

I think what we're seeing is called "learned helplessness". They try, for all of 3 minutes, and don't see immediate improvement in their lives, and give up. We're conditioning people to expect immediate feedback. But all progress takes a long time. You have to stick with it. As someone once said to me, "if you want to dig a big hole, you need to stand in one place for a while." Also, the phone can be a useful tool, but it doesn't have the answers you need. Real people doing work out in the real world are the people you need to talk to, and the ones getting stuff done are making a living doing it, and don't need to post all their secrets online to get clicks. Work for someone who knows what they're doing, pay attention, and ask them some questions during the slow times when they take a break.

You know, I've been trying to replace myself with software for the last 20 or so years. The better I get at it, the more demand I have for my services.

Similarly, just after the last US presidential election, my wife seemed to be literally losing her mind. Like... she thought the world was ending. I'm pretty moderate... I don't like the orange guy, but I try to appreciate what actual issues might be causing people to vote one way or the other. She would start flipping out about something like abortion, and I would say, "you know, the majority of Americans and even male Americans are actually pro-choice, but they rank it a lower priority than other issues like wages and house prices, so that probably affected why the vote went the way it did," and she'd just explode at me. This is a person with a Ph.D. and up until then I'd rate her communication skills as excellent. Finally she admitted that this constant negative emotional state wasn't good for her, and I just asked her to try not using Facebook, just for a week, and go talk to people face to face, and if they bring up politics then change the subject. She did, and after a week she said she felt much better. She's back on social media now, but I think she has some perspective on what it does to her emotionally.

You're suggesting it's mostly good for making deepfake prawn? That would be a good reason to shut it all down from my perspective.

A data center is technically infrastructure. When the AI hype blows over, I guess there'll be lots of cheap computing to be had for a few years. Plus the electricity grid infrastructure they have to add to support it will still be available for other uses, like rebuilding manufacturing? Let's hope.

It seems a bit creepy, but the further you go back in history, the more everyone in town knows all your business. Walking into a restaurant and knowing the names of everyone in there, and the waitress knowing how you like your food cooked used to be the norm. This is just a matter of perspective.

I've been trying. Image generators like Midjourney are interesting, and Veo 3 is wild. But as far as use cases go, I can only think of making little clipart for the buttons in my apps, or more nefarious things like making fake videos of colleagues. LLMs for coding seem to slow me down more than they speed me up. They are only really useful for generating small snippets of code a bit faster than I'd be able to find them on Stackoverflow by googling, but the LLM often doesn't reproduce the snippets faithfully.