cameldrv

I don’t think Google is reading all of your email in this way, or your phone text messages being read in this way, or say iCloud or OneDrive files being monitored in this way.

Certainly it shows that agents like Muse are a complete nonstarter for anyone doing anything that needs to be private. Even if you never talk about it with the bot, merely having sensitive information on your hard drive means it can be sent to outside servers where, if it hits some safety filter, some probably low paid employee is going to read it. Say you’re working for a public company and have files on your computer that could be material nonpublic information, now potentially some content monitor is going to be seeing that and trading based on it. It’s completely untenable to have everything you put on your computer be subject to human review at some tech company.

If you get the chance to see it in person, it's an experience. The best way I can describe it is that looking from Isla Blanca beach, it looks like a skyscraper sitting on the pad, and then, the impossible happens -- the skyscraper lifts off into the sky.

It’s still possible. There’s this sort of guerilla education movement in software with apps like Mathacademy and Mentava to really focus on actual learning, in a scientific way. I can say for my part that my son has been using Mathacademy and the results have been incredible.

The reason most educational software is not like this, as far as I can tell, is that the incentive structures aren’t there. Maximizing learning just isn’t a priority for most schools.

For this new breed of software though, they’re selling to parents, at fairly high prices. Some parents really care about their kids actually learning, and so the incentives are much more direct and aligned.

They can create demand for "Made in Europe" semiconductors that aren't packaged there, and then they can create demand for "Made in Europe" semiconductors that are packaged there and try to get a packaging plant to open up.

If you have a two step problem, the strategy of "I can't do two things at the same time so I'll just give up", is not a good one.

Can you elaborate? How do Target workers live?

Something like this only costs $15m because the incentives are crap. You know it if you've driven past any road construction project. 90% of the people at any given time are standing around. You have to pay just as much for a person to stand around as you do for them to be doing useful work.

Fixing the incentives is not easy at all. There are a lot of "good reasons" why this happens. Many of them are failures down the line, but no one at the contractor level pays the price for not mitigating them, so they don't. If a truck doesn't arrive on time, you could find something else useful to do, or you could just stand around waiting. If you/your company is paid by the hour, it's more profitable to stand there.

Your company could also charge the trucking company for arriving late, or the concrete plant for making them wait. This is not customary in the industry though.

Suppose you fixed all of this though. For a given number of construction projects, you'd need way less labor. If profits are a percentage of costs, less labor means less profit. Making things more efficient means the construction company has to lay a bunch of people off and make less profit, so their incentive is to convince the government that inflated costs are reasonable.

I'm not sure what the solution is. Laying people off is painful. However, one thing you might do would be to do more work for the same money, or higher quality work for the same money.

I'm saying 40k per year per kid per year. That's not a magic number, but it's something like where if you have 2-3 kids, one parent doesn't need to work because their income is replaced by a child subsidy. Alternatively, if both parents work, daycare is paid for, along with the other expenses of having a kid. Now, that does mean taxes have to go up a lot. Maybe you don't need such a high number. You could play around with it and see what works.

On the other hand, prospective parents have to believe that the money will keep flowing, or the incentive isn't as strong. In that case, maybe something like giving new parents a house is a better bet because they'd own that regardless of what future governments do.

The problem is that if Japan ends up with a population of 64 million in 2100, that is nothing like the situation 100 years ago. Almost all of those 64 million people will be retired and not creating any economic output, but requiring a lot of care. It's questionable whether the small working age population could even provide that care, much less all of the other things that a functioning society needs.

Now it's certainly possible that AI will eliminate the need for anyone to work. That is a different question though than conventional demographic prediction.

Ultimately the problem is that Japanese people are not having enough kids. It's a huge emergency. It is recoverable, but it has to be done very soon or it will be much harder. Right now, 5.1% of the population is 25-29, prime childbearing years. If all of those people had 2.1 kids, and so did their descendants, Japan's population would stabilize reasonably well. The problem is that the 0-4 range is only 3.1%. Even if that generation has 2.1 kids, the population will be massively diminished.

If that happens, the majority of Japanese people will be retired, and the cost to provide medical care for them will be astronomical compared to the number of people that can provide that care. It's easy to say "Japan is a rich country, they can afford it." However, medical care for old people is very labor intensive. Even if the old people have "savings", this is just a claim against the labor of young people. If the young people are too worn out from caring for their parents to have their own kids, the country dies out.

The only solution I can think of is to provide massive incentives to have kids. Massive meaning something like "you get $40k per year for every kid in your house under 18", or "you get a free house when you have your first kid, and a bigger one when you have your second."

Maybe this sounds extreme, but the alternative is that young people are taxed at punitive rates in order to provide care for older people. As the situation progresses, younger people will have even less kids because they can't afford them, and eventually they won't even be able to pay for a reasonable living standard for themselves. Older people won't care and will vote themselves excellent benefits, since they would have a good majority in voting. The only option younger people would have would be to leave the country, which only forces the tax rates higher. Eventually something breaks. It's difficult to predict exactly how it would go but it's not pretty.

If Japan is to be saved though, it has to happen in the next 5-10 years or so, and the sooner the better. Once the population pyramid slides up another decade, you'd need the average couple to have 4 kids or so, and the incentives required to make that happen are not affordable, and wouldn't be agreed to democratically when the majority of the population is retired and wants the money for the remainder of their lives rather than paying for kids that would be too young to provide labor to care for them while they're alive.

Depends on the person. Some people will say wild stuff like Neptune exists or black holes exist and they will be right.

When the going gets weird, the weird turn pro. Few people have the ability/temperament to reason about things that have never happened before. Those that do have a weird psychological profile. Usually you can ignore them. Every once in a while, you can't.

Strong agree. I don't believe AIs can become conscious in the sense of subjective experience/qualia, but they can certainly learn to emulate the behavior of a person that is conscious. When they do that, there will be an AI rights movement. When AIs get the right to vote, there will be more of them than there are humans and the AIs will gain complete control.

I spent a few years in college reading and thinking about this question, and I didn't in my heart think that it was anything except a very interesting but impractical question, and yet, here we are. For those who say that they don't want to get sucked into a philosophical debate, well, tough shit. Whether AIs should have rights is a highly practical and consequential question now.

It’s interesting that a lot of U.S. law requires intent. If you just give AI your objective without specifying the means, and the AI violates a bunch of laws requiring intent, but neither the AI nor the person can be prosecuted, this is very convenient.

They made a whole song about this

https://www.youtube.com/watch?v=Ccq30minme8

If you’re using it, you should plan to migrate off ASAP. I’ve been a long time user of Harvest, and after Bending Spoons acquired it, they increased our pricing by 800%.

We can model physics and chemistry pretty well, just not beyond small scales, because the computational effort blows up.

You could just as easily say that if you can write a python interpreter that you can understand every program written in python. Ok, now what if the program is two terabytes?

A frontier LLM is nothing but a 2 terabyte program written in a weird programming language. Just because you can understand the interpreter does not mean you understand the program in a meaningful way.

You’re just a bunch of molecules following the laws of physics. It’s all just physics and chemistry, and those are well understood. Now explain the causes of World War I using chemistry and physics. Simple, right?

There are different strategies. One is to have cheap dry food plus some canned food in 5 gallon buckets. If you buy dry food like beans, rice, and pasta, you can get them from Walmart/Costco for about $1 a pound. This type of food will last a number of years, more if you put them in Mylar bags with oxygen absorbers. You need roughly 1 pound of food per person per day, and you can fit about 30 lbs of food in a 5 gallon bucket. For a family of four, you’ll need about 5 buckets for 30 days if you want to put in some canned food and things like spices etc so it’s palatable. A lot of people have some out of the way corner they can put 5 buckets.

Of course you’ll need a way to cook the dry food. The most sustainable way I know of is a small portable solar system and an instant pot. The instant pot is very energy efficient so you don’t need a very big solar system depending on where you live. If you store flour, you can also get a bread maker, which is pretty efficient.

The reason I like this strategy is that it’s cheap, and you don’t have to think about it very often. You can just trash the food after some number of years. You also don’t have to eat food close to its expiration date normally.

The only hiccup is fat. Cooking oil goes rancid, so you probably want to replace that every year or so, but canola/soybean oil is very cheap.

Of course none of this is very healthy, so you’ll want to include some vitamins…

They have video surveillance signs. Audio is a different category in a lot of states. In California, it's illegal to record audio without consent of all of the parties, and I believe you can get a civil remedy of $5000 for this.

When the AI does something good, the human takes credit. When it does something bad, blame the AI. Take as old as time.

This is really obnoxious. I've used this feature for a couple of decades so that my email is not beholden to Google in case I decide to stop using them someday. I guess that day is coming sooner than I had thought.

However, there is another factor at play, which is that people who are more image conscious drive nicer cars at the same level of income/wealth, and in my experience, people who are more image conscious also tend to act more like jerks.

Yes! The Claudisms do seem to have this slightly uncanny clickbaity feel to them.

The real problem is the climate, the fuel, and structure hardening. In some big fires, it was power lines, but in others it was a cigarette or arson, or a camp fire, or lightning, or a car parked on dry grass, or people using power tools that made a spark.

Trying to pin down liability just based on the ignition source just makes a deep pockets lottery. If the power company lights a massive fire, you get a payout. If someone throws a cigarette out their window, they go bankrupt and you get nothing.

Eventually if you have a dry forest with a lot of dry brush and you have enough hot windy days, it's going to catch on fire. If the power lines are perfect, maybe you'll delay it a few years, but the same thing is going to happen. Either you deal with the problem from a fuel standpoint, or you make peoples' houses harder to burn, or you're going to end up with a bunch of burned down houses. There really isn't some magical solution where you can just blame it all on PG&E.

I really like the GE Sun Filled soft white bulbs (not the daylight version). They’re pricey and only 60w equivalent but they’re the closest to incandescent I’ve found.

To a degree you might say that they live too much in the space of the abstract and not enough in the concrete, or too much in the general and not enough in the specific.

I think that's a fair criticism a lot of the time. Most of the time we have precedent that can guide our actions well. Trying to reason everything out from first principles can be wasteful navel gazing when we've collectively seen the movie a million times.

Focusing on reality and specifics though is what causes people to say there's no global warming because December is cold.

It is sometimes possible to see a trajectory that has never happened before though and that's when you need the autists.

Honestly I think American scientists not speaking in Fahrenheit units to an American audience is an incredibly simple and stupid, but very significant cause of the problem.

If I tell you the Earth is going to get 2 degrees C warmer, Americans are going to think, "2 degrees, not that much, barely noticeable." If you tell them it's 4 degrees F, they'll think, that's not a ton, but it's definitely noticeable.

There was also the Qualcomm pdQ and the Kyocera 6035 I think that combined a Palm and a phone/cellular modem. It’s surprising to me that it took as long as it did for smartphones to really get popular.

Some of it was just that the typing UI was pretty bad until the blackberry had the mini keyboard and then Apple came out with a usable touch screen/software combo that was good enough.

Also for a long time web browsers on phones/PDAs existed and worked, but the UI was lousy so they were very cumbersome to use. Apple had a few big innovations in UI that helped, multitouch, and double tap to zoom to a column, but the 2006 era windows mobile phones also had fairly decent browsers. Microsoft really blew it on phones.

This is my frustration with AI coaches, specifically Garmin. They want me to have a 10k time goal or something when I don’t ever run competitive races and don’t directly care about times. I just want to get healthy.

My guess is most of those are passengers from Milwaukee connecting in Chicago. You could take the train, but then you'd be in downtown Chicago, and have to take Uber or the El to O'Hare. That's a big hassle and takes a while.