loudmax

Viktor Orban's Hungary is illiberal. North Korea is a dictatorship. One is bad, the other is worse.

The danger of conflating the two is that you're setting people up for false expectations. If you go around telling people that Hungary is a dictatorship just like North Korea, and then they visit Hungary, they might away with the notion that Orban is not so bad. Nobody is starving, and they won't happen to notice the reporters in prison. You run the danger of undermining your own credibility.

That was more or less the case during Trump's first administration. I think a lot of normie Republicans were hoping for a repeat of that. The ones that aren't in denial are being gravely disappointed.

The term for this is extortion.

The last time tariffs were this high, it led to rampant corruption as companies would pay off customs officers. This was one of the reasons for switching to an income tax. For the current administration this possibility counts as a major opportunity to generate personal wealth.

But this isn't the only reason for the policy. For someone who is at heart a coward, bullying and brandishing raw power over others is its own reward. That reason enough for the policy, and damn the consequences for the nation.

Incidentally, this concept of modifying one's ecosystem comes up in Richard Dawkins' book The Extended Phenotype, which is a follow up to The Selfish Gene. One of the illustrative examples is beavers turning fast moving streams into convenient slow moving fisheries by building dams. The ecosystem the beavers built is "natural" in the sense that beavers are animals and it's in their nature to chew through trees. But presumably, at least some other animals were pushed to extinction when their habitat was modified by the beavers.

The "Game Design Otter" action figure seems to come with a pair of flashlights. I bet that's a residue from the previous prompt about illuminating the tablet with a flashlight.

That is a very interesting point about how little use of AI most of us making day to day, despite the potential utility that seems to be lurking. I think it just takes time for people and economies to adapt to new technology.

Even if technological progress on AI were to stop today, and the best models that exist in 2030 are the same models we have now, there would still be years of social and economic change as people and companies figure out how to make use of novel technology.

It's pretty clear now that a suggestion algorithm that prioritizes engagement at all costs has toxic side effects. This does not prove that all promotion algorithms, much less all social media, is inherently toxic.

Mastodon is a social media platform that's not built on a content promotion algorithm. Mastodon has a lot of group think, and you can absolutely build yourself an echo chamber full of political virtue signalling if that's what you're into. But I wouldn't characterize this as "toxic," it just reflects what you put into it. There's no algorithm to pump it into your face if you don't want it.

I'd love to see a version of Mastodon, or some social media platform, that manages to promote sensible discussion algorithmically. I don't know how this would work, but I wouldn't go so far as to say that this proves that it's impossible.

Ironically, the one American car company that was interesting to Europeans was Tesla.

European governments may be reluctant to put a tariff on Tesla because of Elon Musk's political association with the Trump government. This is literally how fascism works, and appeasing the bully is distasteful, but that's realpolitik. It will be up to European consumers to reject Tesla because the brand is now toxic.

Trump's businesses have repeatedly declared bankruptcy. Don't rule out ineptitude.

What the median right-of-center rich American wanted was tax cuts, with maybe a little bit of culture-war anti left stuff thrown in for spice. They are not pleased with abandoning the economic world order that Reagan did so much to promote. Sensibly designed, focused tariffs can be good for the economy, but that's not what's being rolled out.

The people who actually want these blunt universal tariffs fall into three categories:

  * The uneducated
  * The deluded
  * The grifters

The uneducated want tariffs because they think it will restore their purchasing power and middle-class status (which it won't).

The deluded want tariffs because they've bought into the cult, and the reckoning with reality if they back out now is too painful to contemplate.

The grifters actually benefit from selective exemptions based on political loyalty and having the money flowing into the right coffers.

MCP is Anthropic's Model Context Protocol (https://www.anthropic.com/news/model-context-protocol), which intends to be a standard method for AI models to communicate.

Amazon is a one of Anthropic's main backers, hosting their infrastructure, much as Microsoft is a big backer of OpenAI by hosting ChatGPT on Azure.

"A girl's got to have her standards." (Also from 80's memory)

It depends on coworkers, it depends on the office environment more generally, and it also depends on the type of work you're doing.

I'm not a particularly social person, but in general I'd rather work in an office near my coworkers than remotely. But, I'd rather work remotely than work in a large open office space. A cubicle might be okay, depending on how many people are sharing the same room.

Another big factor is your commute. Spending an hour or more on a daily commute is very taxing, especially if you're raising a family.

But with a population of 2.7 million. That is the more relevant factor here, and by this metric they're on the side of tiny.

And that makes their ability to develop a functioning satellite that integrates into a launch vehicle all the more impressive! Kudos to Botswana's engineers.

  On Feb. 15, 2023, EGodly thanked Coristine's company for its assistance in a post on the Telegram messaging app.

  "We extend our gratitude to our valued partners DiamondCDN for generously providing us with their amazing DDoS protection and caching systems, which allow us to securely host and safeguard our website," the message said.

This isn't about whether or not Coristine should go to jail for EGodly's criminal activities. This is about whether Coristine has the maturity and moral integrity to be accessing government systems.

Having read the article, Coristine does not come across as blameless. When you run a business, you shouldn't be gathering criminals as your clients.

If you have a short daily commute, I recommend considering plug-in hybrids.

I got a PHEV Prius Prime in 2020 and I'm delighted with it. The battery has a range of about 20-25 miles, depending on the weather. My daily commute is about 10 miles each way, so I can make it to work and back without needing gas, and I only need to fill the tank three or four times a year. In my case it works out really well: driving an electric vehicle 95% of the time, and never any range anxiety.

The way things are going, we're likely to see how things turn out without the global nonproliferation system. I'm not looking forward to it.

I went to elementary school in Belgium from 1978-1982. I had the sense there was some national pride in having had a major African colony, but maybe Leopold II wasn't a benevolent ruler. Unlike Leopold I or Albert I, who were depicted quite heroically. I didn't learn quite how far from benevolent Leopold II really was until much later in life.

You jest, but that's absolutely how the MAGA folks talk about it in their echo chamber. It's an alternate reality.

He never will, and neither will his most ardent supporters. Television is their reality, and they're very selective about only watching channels that confirm their biases.

I wouldn't hold owning a Tesla against anyone, especially if they bought it before 2024. Lots of people bought their Teslas before Elon Musk became full-on red-pill woke.

Now that Cybertruck is another story, there's no excuse for that monstrosity.

I use LLMs all day and they've replaced a lot of what I'd previously searched for. Even so, I don't think I'm ready yet for an LLM as default search engine. LLMs are useful for generating information based on what they know about. For example, generating useful commands for tools like `ffmpeg` or `jq`. They're not better than a search engine for finding factual information, like "What is the population of Canada?" LLM hallucination is still a thing.

LLMs replace some of the work we'd previously had to rely on search engines for. They're not a replacement for everything, and they shouldn't be used as such.

One use case I imagine is skilled workmanship. For example, putting on a pair of AR glasses and having the equivalent of an experienced plumber telling me exactly where to look for that leak and how to fix it. Or how to replace my brake pads or install a new kitchen sink.

When I hire a plumber or a mechanic or an electrician, I'm not just paying for muscle. Most of the value these professionals bring is experience and understanding. If a video-capable AI model is able to assume that experience, then either I can do the job myself or hire some 20 year old kid at roughly minimum wage. If capabilities like this come about, it will be very disruptive, for better and for worse.

This reminds me of rental car companies deciding they should switch their vehicle fleets to electric.

A few months ago, I visited San Jose for a wedding. When I picked up my rental car at the airport, the only options were electric, even though I had specified that I wanted an ICE vehicle tat the time I made the reservation. During the four day trip I wound up visiting five different charging stations (some of which slow charged, so weren't able to replenish the battery in the time available), and I had to install three different apps. I still have like $20 of unused credit between them. I spent several hours waiting for the car to charge, not to mention making major detours looking for a fast-charging station. If I were to guess a part of the world you'd expect to find the best possible electric vehicle charging infrastructure, San Jose wouldn't be far off the mark. But my trip wound up being dominated by range anxiety.

I drive a PHEV (plug-in hybrid) for my commute from home to work and I love it. Electric is the future, and it's great that electric vehicles and charging stations are becoming more common. But renting an electric car in a strange city today is about the worst possible scenario for a short range vehicle. You don't know where the charging stations are, the charging stations require different apps, your hotel might not have a charger, and so on. The people making decisions at car rental companies should know this!

I agree with your broader point, but I really wish people wouldn't refer to MAGA Republicans as "conservative." Right-wing, or reactionary, certainly. But there's nothing "conservative" about undermining the rule of law and ignoring the Constitution.

By the same logic, I don't like calling far-left cancel culture "liberal." I'm afraid our partisanship has blinded us to genuine liberal and conservative virtues.

The bottleneck for running LLMs on consumer grade equipment is the amount of VRAM your GPU has. VRAM is RAM that's physically built into the unit and it has much higher memory bandwidth than regular system RAM.

Obviously, newer GPUs will run faster than older GPUs, but you need more VRAM to be able to run larger models. A small LLM that fits into an RTX 4060's 8GB of VRAM will run faster there than it would on an older RTX 3090. But the 3090 has 24GB of VRAM, so it can run larger LLMs that the 4060 simply can't handle.

Llama.cpp can split your LLM onto multiple GPUs, or split part of it onto the CPU using system RAM, though that last option is much much slower. The more of the model you can fit into VRAM, the better.

The Apple M-series macbooks have unified memory, so the GPU has higher bandwidth access to system RAM than would be available over a PCIe card. They're not as powerful as Nvidia GPUs, but they're a reasonable option for running larger LLMs. It's also worth considering AMD and Intel GPUs, but most of the development in the ML space is happening on Nvidia's CUDA architecture, so bleeding edge stuff tends to be Nvidia first and other architectures later, if at all.

Fans of James Cameron will remember the POV of the terminator deciding how to respond to "Hey buddy, you got a dead cat in there or what?"

Played for laughs, but remarkably prescient.

How do Switch users feel about the joycons?

I'm not a gamer, but the original Switch joycons always struck me as overly complicated and expensive. It should be cheaper to manufacture and sell Switches with the controllers attached. Indeed, this is what they did with the Switch Lite. For games that take advantage of joycon functionality, Nintendo could have sold something like an updated Switch version of the Wiimote as an optional accessory.

Do users who are happy with their Nintendo Switch have a favorable opinion of the joycons, or would you be happy without them?

I tend to agree. I used to run Call of Cthulhu in high school, and there was a lot of preparation involved in running a good scenario. Some people are good at winging it and making stuff up as they go along. I never was. The best experiences for me and my players was always when I'd meticulously designed my own scenario, or used a store-bought scenario and read it thoroughly. Either way, hours of prep.

I would love to have an AI as an assistant Dungeon Master (or game master, or Keeper, or what have you). That is, one person in a group of players maintains the role of a master storyteller, but the AI is ready to fill in details or suggest ways to get the players back on track. This would probably be tedious if you're interacting with the LLM entirely through text, and having to manually keep it up to date with the story. But it could work well if you have a model that understands spoken language listening in on the game and generating cool images and making private suggestions to the game master.