> The GPUs were designed for graphics [...] However, because they are designed to handle everything from video game textures to scientific simulations, they carry “architectural baggage.” [...] A TPU, on the other hand, strips away all that baggage. It has no hardware for rasterization or texture mapping.
With simulations becoming key to training models doesn't this seem like a huge problem for Google?
China also has 58 nuclear reactors and are currently building 30 more. [0] They are doing everything right to get rid of coal and also become energy independent.
It is! Their biggest solar farms are also in the Inner Mongolia where the irradiance is twice what you get in Germany. That said the sheer scale is crazy!
Why dwell on the past? Currently per capita electricity consumption is higher in China than in Germany (6.5 MWh vs 6 MWh).
However, it is true that even in light of this current situation China is building out solar a bit faster (on a per capita basis, even if adjusted for consumption) than Germany. In Germany it‘s about 1 GW added each month, which adjusted for population and energy consumption is about a factor of 1.5 compared to Chinas 25 GW per month.
Wind is lagging behind in Germany but, to be honest, looking at numbers from 2024 compared to China it’s about the same factor 1.5 difference when adjusted for population (3 GW compared to 87 GW).
Germany should be and could be as fast as China – but there aren’t humongous differences between the two countries.
How fast has Germany's PV capacity expanded in recent years? In another subthread I wanted some estimate of how many hectares in Germany had PV on them, but the pages I visisted on the subject were outdated.
Germany has a space problem. There aren't large swaths of land available to put solar panels on. Added to that you have to realize just how far north Germany is, and consequently how... Bad solar is because of that.
Seriously, dropping down the same amount of panels gets you significantly less electricity in Germany then where China is building them, much closer to the equator.
Overall, Germany is in a shit place for renewable energy
For many years I have been writing about how bad Germany is for solar panels because of, mostly, how far north it is. Specifically, the capacity factor for utility-scale solar in Germany has historically been about 10%, which is mostly because of the unfavorable latitude and cloudiness. You can see this quantified at https://solargis.com/resources/free-maps-and-gis-data?locali.... The capacity factor for utility-scale solar in China has actually also historically been about 10%, and I have no idea why that is. You are certainly correct that it ought to be much better, because China gets so much more sun. You can see that quantified at https://solargis.com/resources/free-maps-and-gis-data?locali....
However, Germany does not have a space problem. Germany is 357'114km² with "photovoltaic electricity potential" of about 3.0kWh/kWp/day according to Solargis (see above link), which would be a capacity factor of 12.5%. I'm not quite sure how they calculate that, but multiplying by the country's area, the solar constant of 1000W/m², and a fudge factor of 0.8, it works out to something on the order of 30–40 terawatts, electric. That's roughly 50% to 100% more than the entire world's marketed energy consumption, which is about 18 terawatts, about a third of it electric. Germany produced 488.5 TWh in 02024 (https://en.wikipedia.org/wiki/Electricity_sector_in_Germany) which is 55.73GW.
Therefore, roughly 0.2% of Germany's land area would suffice to produce all of its current electrical consumption with solar energy, about 700km². This would also require something like 450 gigawatts (peak) of solar panels, which would cost about €45 billion at today's prices, roughly 4 days of Germany's GDP.
It is absolutely true that, if you put those same solar panels in the Mojave Desert, mounted with single-axis trackers, they would produce two or three times as much power. (California's average utility-scale solar capacity factor was over 29% last I checked.) So, yes, solar is much more expensive in Germany. But if you check out https://www.solarserver.de/photovoltaik-preis-pv-modul-preis... you will see that, in March 02023, solar modules cost three times as much as they do now. So solar generation in California then cost what solar generation in Germany costs now. (Except that, because of Biden's anti-renewable-energy tariffs, actual California prices were and are much higher than you would expect from Solarserver.)
To put this somewhat in perspective ("why is Germany not installing 700km² of PV panels right now?"):
Grid operators project costs of around €250 billion [1] (within 2045) for grid expansion alone, and NIMBYism is a big problem everywhere in Germany.
Battery storage to allow going full PV (1 week) would be at least twice that [2].
There is also plenty of "anti-green" sentiment in Germany generally and the current government coalition is in a somewhat precarious position, discouraging long-term investments like that.
Yes, I have always thought cloudscape design is a great framework to build dashboards like this. Feel free to check out the source code for the whole project as an example, everything is open-source!
Were participants given time to customize their Cursor settings? In my experience tool/convention mismatch kills Cursor's productivity - once it gets going with a wrong library or doesn't use project's functions I will almost always reject code and re-prompt. But, especially for large projects, having a well-crafted repo prompt mitigates most of these issues.
> Up until this point, both students had nearly perfect homework grades while failing every in-class quiz.
From a student's perspective: I think it was the same with SO. While LLMs make c&p even easier, they also have the upside of lowering the bar on more complex topics/projects. Nowadays, the average person doesn't touch assembly, but we still had a course where we used it and learned its principles. Software engineering courses will follow suit.
StackOverflow users at least tried to fight against it. The disdain for anything that looked like "homework questions" was one of the reasons it got a bad reputation among some people.
The problem isn't that the question is "for homework".
The problem is the lack of analysis that goes into producing a useful question for others that fits in with the rest of the site.
True, proper analysis of the homework rarely yields such questions, and even less so in 2025. But the point was always to have a question that's about a clear, specific problem, not about a task that was assigned. Because the latter can only possibly help people who were assigned the same task.
What is a realistic path to this? Being from Europe I visited US last year and was horrified at the quality of your food. You see a lot of documentaries/youtube videos/etc... discussing the problem, but how do you even go about this?
We have good food, but you won't find it on "The Easy Path."
The Easy Path is that gentle encouragement to hit up Chipotle for lunch, because it's "right there."
The Easy Path says dinner's hard and you've had a long day, so get something simple, like take-out or microwave.
The Easy Path is entropy. The Easy Path is self-care over struggle. The Easy Path is simple carbs shown on prominent display in store shelves. The Easy Path is advertising.
Hitting the gym isn't on The Easy Path, but forgetting to cancel your gym membership is.
These days, big food companies love "The Easy Path" because it's so easy to commoditize, it's the "Path that Americans are Expected to Take." For financial stewards, being on The Easy Path turns lack of willpower into your ally.
On the other hand, getting good food in the US requires passing the marshmallow test: you have to meal prep, or you have to shop around the sides, or you have to get something on the salad menu. You have to say no to advertising. You have to expend willpower, the most limited of resources to the average American. You have to Go Hungry or Suffer, or have An Upset Stomach. You frequently have to spend more money or time.
Semaglutides are not currently on The Easy Path. Maybe they will be someday. I personally doubt that, because putting GLP-1 on The Easy Path would require big food companies to rethink their entire portfolio.
But you're not wrong in that they could be Easy Path-ajdacent. The dialectic would shift: food companies would shift around to be Organic and Nutritious and Less Calories and find other ways to stay on The Easy Path. Sugar and fat's addictiveness is highly Easy Path-enabling, and that's a pretty big vacuum to fill.
I think you're making this sound harder than it is.
If you count calories and stick to a budget, you will lose weight, even if those calories come from deep-fried fast food. Sure, it's good to eat more whole fruits and vegetables, and you should, but weight loss doesn't require some kind of Edenic perfection. Stick to a calorie budget and you will lose weight, the end.
We can add some second and third order provisos, sure. The next tip would be to go low carb. And to keep a spreadsheet with calorie numbers for everything you eat. Track what you're doing.
But basically, if you eat 1500 kcal/day for nine months, you will be much thinner. We don't have to make it harder than that. It works. Perfection is not required.
If we're talking about this as a public health issue, then I agree with you. You can't really expect much of people as a herd or mass.
(Hell, look at the state of elections.)
If we're talking about this as individual, rational people, though, then it's different. You can absolutely maintain a reasonable weight if you just attend to it.
2.
There's actually some knowledge embedded in the "count to 1500", which, if you're not in the habit of thinking about, may be surprising. Specifically, the kcal amounts themselves.
Say you just go with the flow of society, and you eat "normally" without thinking about what you're doing:
You wake up, and you have something marketed as "a meal" for breakfast. A breakfast sandwich can easily be 550 kcal. If you add a venti latte to this, especially one with sugar, then you could easily add another 250 for that. Now you're at 800 kcal just for breakfast.
Then consider lunch. Even a "small" meal from a healthy salad place, like Sweetgreen (which is expensive), is going to be like 900 kcal. Say that's what you eat. Now you're at 1700 kcal.
The afternoon comes, and you have two chocolate-chip cookies. Two cookies isn't excessive, right? Just a little treat. But each one is probably 120 kcal. So that's 240 kcal. Now you're at 1940 kcal.
Finally you have dinner. Some microwaved thing, relatively small. It's probably like 600 kcal. So now we're at over 2,500 kcal for the day.
Everything you did in the course of that day was relatively normal. Probably only the venti latte at breakfast was obviously excessive. But now you're substantially over your calorie budget.
Now, 2,500 kcal is still salvageable. Every mile you walk on flat ground is about 100 kcal. If you live in the city, you could easily have walked a mile and a half during your commute to work, and another mile when you stepped out for lunch, and a mile and a half on your way back, giving you four miles. You're almost at breakeven. Just need a little more exercise, and you'll maintain your current weight.
(In the burbs, though, you probably drove to all these places and now it's on you to go to a gym, which is a pain in the butt.)
Anyway, my point is, if you weren't attending to all this, how would you know? You'd probably just be doing all these habits without thinking about them. Most parts of this routine seem pretty reasonable. But you'd still be getting fat. Because you're going with the flow instead of counting.
So, there's a little (easy) knowledge involved, but mostly it's not an issue of intellectual ability, it's a matter of attention.
>The Easy Path says dinner's hard and you've had a long day, so get something simple, like take-out or microwave. The Easy Path is entropy. The Easy Path is self-care over struggle.
If you see this sort of food as "self-care" then that's where the war has been lost.
>or you have to shop around the sides
Sometimes I wonder if I'm the only person in the world who enjoys this. Discovering, for example, the versatility and cost-effectiveness of skim milk powder was a real game changer for me. Similarly for dried legumes and fruits.
The Easy Path is a meal service like Factor that delivers healthy food directly to your door step.
The Easy Path is signing up for a fitness class on a regular schedule and baking it into your morning routine.
The Easy Path is not buying extra snacks - just don't have them laying around the house for you to eat when you're bored.
The Easy Path is the path of least resistance. However, you have some agency over the environment you create for yourself, so that path of least resistance is to some degree under your control.
Get ground beef in the supermarket, it’s cheap and takes 7 minutes to cook. If that’s all you eat you can’t be fat and out of shape. You also won’t be hungry.
At some point blaming society isn’t going to cut it.
At some point "have willpower" isn't going to cut it. That point was decades ago.
Blaming society sounds fatalistic, but yes we absolutely can change society. There are mechanisms to do it, and the first step to utilizing any of them is people getting pissed off about the state of things and talking to other people about how pissed off they are about it.
I’ve done plenty of grinding it out in my life with no energy or time to spare.
Even in that place you can do better than ground beef.
Get an instant pot and invest in some rudimentary cooking skills. I’ve spent a whole year making variations on the same dish in 15 minutes, using that cooking as end-of-day stress relief. Shopping for the same handful of things once a week.
It wasn’t fine dining but it was healthy, cheap, and a few steps up from ground beef. Come now.
I love fatty meat and salt. I can’t think of many things better than that. It’s also the only staple us and our ancestors have eaten for millions of years.
I think the instant pot is too imprecise a tool for cooking by recipe, it’s for speed and convenience. I’ll improvise a curry with what’s on hand, sometimes the result is fine and sometimes it’s great :)
It’s not the only thing. You can eat any meat, eggs, or many dairy products. Probably most fruit is fine too, although I don’t eat carbs myself.
Yes, all the processed grains and seed oils and artificial sweeteners and chemicals should never have been invented. But I’m not too fussed about that. What I can do is come here and write these comments about the diet that changed my life for the better. If only one person who reads them tries it out, I will already have succeeded.
The US is a huge country (9,833,520 square kilometers!), so I find it curious that such generalization can be made about the food available here, or even the eating habits of 334,914,895 humans. I could say that I visited Amsterdam 2 years ago and I was shocked with the quality of the food.
But I would never do that, since I mostly ate at the Red Light District, and I couldn't possibly generalize the country eating habits with stores in a major tourist area.
Maybe I'm too European to understand why not, but seems to me that regulations around food and what companies are allowed to put into it is really helpful in avoiding companies from just stuffing whatever down people's throat.
There's plenty of high quality food, you just have to know where to look. For example, come to the Bay Area and check out Whole Foods and any number of high-end restaurants.
That's overkill. There are very marginal health benefits to eating organic watercress vs. eating whatever dark leafy green is currently on sale at the discount supermarket.
The problem isn't that people go to supermarkets and they can't find any healthful ingredients to cook with. The problem is that they go to supermarkets and pass those over in favor of convenience foods that have been optimized for "craveability" [1].
GLP-1 drugs can alter this behavior by reducing food cravings. Someone who's no longer craving the most craveable food can make more objective by-the-numbers buying decisions the next time they go grocery shopping.
was in greece for three months. the quality of the BASIC fruits and vegetables at the regular local market down the street from where I was staying was on par with wholefoods. It was surreal how cheap it was to eat HEALTHY.
Almost any ethnic grocery store will do since they cater to immigrant communities that are likely to be lower income. Here in SoCal some are cheaper than others but they're all way cheaper than Ralphs/Vons/Trader Joes/Costco/etc (I don't shop at WalMart so I'm not sure how they compare)
There are also native stores that are increasingly entering into low cost produce like Grocery Outlet and then there are the usual like Food4Less but they tend to eventually move upmarket.
Why is the quality of food in Europe so much worse than southeast Asia?
Because you guys are way, way, way fatter than e.g. the Japanese.
Back on topic -- we have excellent food in the US, but regulations allow for highly processed crap to be sold too. Pretty sure most of the crappy processed foods are easily available in Europe, too.
eu and canada have stringent laws on advertising to children, and laws on nutrition and additives to their products
even things like bread in the US have an insane amount of sugar
but the other reason is suburban/car culture & zoning, which means more hypermarkets and shopping not every few days for fresh food but every week or two for more processed food that lasts longer because going to the store is a bigger PITA than a quick 5 min walk to the neighborhood stores; which also means mroe walking rather than driving, another area of calorie burning and lean muscle maintenance which maintains high metab
The first thing that would help is actually having a realistic discourse about food, and not the idiotic - "You shouldn't be eating processed food, its not good for you".
Like most of the food that we eat is not really that bad. Its not optimal for sure, especially for sedentary lifestyles, but a lot of the health problems are not directly tied into the actual food, rather the over-consumption of it, and passing down of bad genetics (for example, children of obese people are more likely to be obese).
European obesity tripled in the last 40 years as well, despite higher quality of food.
In the Netherlands until recently you could get it for ~$10/mo (now ~$20). We have a whole website naming prices in different pharmacies around the country.
What about the awareness of what self-harm/depression is in the first place? It's easier to say you're depressed if you understand what depression is, and it's easy enough to Google that. Same with self-harm, I wouldn't ask my classmate but I could easily Google it. Similarly, mental health campaigns could contribute, accelerating how fast kids learn about bad mental health. We could've had the same percentage of depressed young people, they just wouldn't know how to express it/act on it.
I think that better diagnosis does account for some of the diagnosis count (like many mental health issues the solution used to just be "if you can't figure out how to act normal on your own you're going to a home/asylum, we may smack you around a bit first to see if we can get the message through to you").
But talking with any educators for that age group it's not just that the number of people diagnosed as depressed/anxiety but the level of depression/anxiety in the individuals diagnosed escalating that alarms them the most.
With simulations becoming key to training models doesn't this seem like a huge problem for Google?