It's a genuine Chinese news article written by a real person. I don't know why they choose to use these types of images, it truly makes it look like slop. I think it's because it's originally a WeChat article posted on their website.
The article linked here seems to be a (bad) auto-translated version.
It's not uncommon for images to be incorporated in an article like this, as they typically get shared around WeChat and correspond better with the 'culture'.
How do we explain this little blob of sloptext right in the middle:
> The second task was more straightforward: both models were asked to search for flight tickets in a browser at the same time.
> I'm unable to open a browser or interact with live websites like Google Flights. I can only process text and don't have real-time browsing capabilities. To find this flight yourself, here's what you'd do: 1. Go to google.com/flights 2. Enter Singapore (SIN) → Beijing 3. Select one-way, set departure date to 5 days from today 4. Filter by "Nonstop" and "Economy" 5. Check the results — note that flights to Beijing may land at either Capital (PEK) or Daxing (PKX). Exclude any Daxing arrivals. 6. Compare prices and pick the cheapest nonstop option landing at PEK. If you'd like, I can help you think through typical price ranges, airline options on this route, or general tips for finding cheap flights.
> Among them, StartLux-V1.0-27B-Preview completed the search in about 95 seconds, finding an Air China ticket priced at $299.
Did you even read what it just said above? They're testing the model's capabilities so of course they'd have whatever text the model outputted in the article, that's the entire point.
There's no such thing as an innate most qualified person. Zuck was never the most qualified to be CEO of Meta. He is now, because of what he's done. At the same time there's an infinite number of jobs he is poorly qualified for.
Strict qualification can only be measured in a extreme niche cases. You founded an area of research, so you're the most qualified to work on that at Google. You have a massive Linux project, then maybe Torvalds or his core maintainers are the only ones who fit.
But outside of that, people become qualified because of what they do, and before they did it, they were unqualified.
As an example, I'm sure there are many people at OAI and Ant who are today world experts at research and training and infra. But they became that by just doing the job. Take their same resume from 5 years ago and they may not even reach the interview if they applied for the same role. The what does qualified mean, how do you measure and define it?
Countries need to develop their own talent and individuals need to develop themselves.
Sweden does not have surplus electricity. We have production of electricity that is better matched to demand than other nations, most notably the disaster that is Germany. German demand is causing Swedish industrial production to shut down due to increased energy prices, which makes production unprofitable.
Sweden will not pay curtailment fees, such a thing does not exist. The reason is exactly what you quote at the end.
To state it again:
1. The intranational powercables means that German demand raises electricity prices in southern Sweden, which means businesses and consumers pay a premium vs what they would have paid without the cable
2. The extra charge is captured by network operators as what is called bottleneck fees and it's billions of euros per year, very large sums
3. The EC first wanted to confiscate these funds, which come from Swedish consumers and businesses. Now they want to rule about their use. Obviously this is completely unacceptable and nothing more than a bureaucratic powergrab, and no country would accept it.
The European council (which consists of elected country heads) nominates a Comission President and the members of the parliament (elected politicians) vote on the nominated president candidate and need absolute majority to confirm.
That president then tries to form the commission with one representative from each member state. The parliament then has to vote again to confirm the commission.
Sure it's not ideal. But compared to China, the population of the EU has influence.
Each member state elected their own government, which in turned determined their head of state by their own country's rules, then those heads of governments appointed VDL.
Do we even have a real choice with our governments? As a minor yet dramatic exemple, last year in France the Front Populaire won the last parliamentary elections, and according to republican tradition and common interpretation of the constitution, should have formed a government. But that was refused by Macron who named yet again an even more impopular proto-fascist government because it's technically in his power.
I hate to make this argument because the PRC is far from a political ideal for me, but the chinese political system also has elections and votes. Voting is not what democracy is about, and very indirect elections for a vast body with immense powers even less so.
I have yet to meet anyone happy with the european commission policies, except for very minor and timid arrangements such as the USB-C connector (and even then, Apple fanboys complained). The rest is all about destroying every pro-consumer/pro-worker/pro-environment regulation in favor of big multinationals who employs thousands of lobbyists across Brussels and Strasbourg.
Indeed. Every time it is mentioned(without fail) how undemocratic the European Commission is, someone points out that the demos has a say via 3 levels of indirections. But that is not at all what democracy is all about.
I am supposed to have a say about how the country I live in is run. That is democracy. Not that I get to vote on a person who votes on a person who votes on another person who may or may not agree with me.
It's why I live in Switzerland now, and not the EU.
Following that logic, any minister appointed by a head of government is "unelected and has the same level of democratic involvement as that of China's Standing Committee".
1. What does this have to do with Denmark? This is about upgrading an existing link from Sweden to Denmark (seems mainly for maintenance reasons, not capacity; it's only increasing from 715MW to 1000MW)
2. Why did the EU have to reassure Sweden wouldn't have to pay bottleneck fees to Denmark?
3. If you have such great power generation in the north of Sweden, why isn't it available to South Sweden where the demand is? (whether that's to Germany, Denmark or Sweden itself)
Germany is currently getting power from you directly, as well as via the existing Denmark connection (this project is merely upgrading it from 700MW to 1000MW), as well as via Lithuania if it wanted. Germany gets more of its energy from Norway and Denmark directly without involving you. Not to mention all the other countries that surround Germany.
So if your problem is Germany, why are you complaining about Denmark?
2. No, I see the issue now. It's not that Sweden objects paying money to Denmark, it's that Sweden objects paying for any other country's grid upgrade.
3. I ask about Northern Sweden because I heard it had lots of potential for electricity generation, but it can't even get it to Stockholm, never mind Malmö, due to north/south grid capacity issues purely inside Sweden. Your current curtailment is about 3.8TWh per year. I read that when you can't get power into the SE4 area, which can happen daily, the prices shoot up as it needs to import from Denmark, Germany, etc.
Wouldn't it be better for all concerned if this was addressed? You then wouldn't have such a price differential between areas of your own country.
I think the M5U Ultra 96gb is the sweetspot in that price range. It has more compute and bandwidth so you get to run models better sized to its hardware. I believe the Max would be too slow; personally I'm getting this SKU because I think it'd suck to get the 128gb Max and then discover it's too slow, and end up just using cloud providers anyway.
I agree about getting the Ultra if you're interested in AI (LLM) inference speed. I'm a little perplexed as to why there isn't a RAM option in between 96GB and 256GB, though. For instance, I believe Deepseek v4 flash runs a lot faster on (https://github.com/antirez/ds4) with 128GB than 96. I assume the higher SKUs have low enough sales that Apple doesn't want to be manufacturing too many different RAM configs as that would eat into their margins. (I say this as someone interested in AI hardware in general, even the 96GB is out of my price range).
It's the upsell ladder. This is how they get you to buy a more expensive tier or two. At first I was only going to look at 64gb or 128gb Max, but ended up here.
I also think it’s possible that it may be best to continue waiting on the Mac side of the house despite the increasing prices.
I think that some of the hardware design folks have been blindsided by AI demand and we haven’t really gotten that next generation AI hardware yet, to the point where buying M5 isn’t going to make sense in a couple of years.
Rumors seem to be that the M7 is the generation that Apple is looking to push AI performance much further.
I’m not sure that Apple anticipated this specific route that computer hardware has gone and I don’t think M5 and previous iterations were really specifically architected for local AI performance, more like they happened to be pretty good at it.
> Rumors seem to be that the M7 is the generation that Apple is looking to push AI performance much further.
What specifically does that mean? The M5 series has 10 cores per 128 bits of memory bus; are the cores unable to keep up with the RAM? I thought they did and memory bandwidth was usually the bottleneck. But the memory bus is already very highly clocked and goes up to 1024 bits wide so it's hard to picture memory bandwidth having a huge leap.
LLMs token generation is memory bandwidth constrained. If the M7 has double the memory bandwidth as some speculate [1], then it will help with LLM performance. There is also expectation (possibly unfounded) that the GPU will have improved matrix performance to help prompt processing as well.
14GT/s is a goal but it might be a while, I think 11-12 is more likely in that time frame.
50% more pins... we'll see. If they can reasonably make that fit then even more shame upon the traditional desktop CPU makers for sticking with 128 bits for so long.
For a coding agent, just build an agent memory that is repo-aware and stores its files in respective repos. you can do this very quickly with any agent, just ask it to research memory systems like coala and propose an architecture
reply