z.ai is a fairly well known AI lab out of China and their GLM models are probably the most popular outside of Anthropic or OpenAI’s. I don’t think it’s presumptuous for them to not introduce themselves in a post on their own blog, I think you’re just a bit out of the loop here.

And honestly that's for a reason. GLM5.3 on max has in my experience far less hallucinations than any other open weights model and it feels it has some intuition to bring in the right information when it's in principle out of context but relevant to the topic. Like its goal is more to bring value and assist you than just solving the given task with the least token spent.

I don't get the outrage. Do you post this kind of stuff on every topic on hackernews that you are not knowledgeable about?

Maybe my post sounded harsher than I intended, and yeah, it's probably on me that I'm not familiar with GLM. Actually the other major Chinese LLM Kimi does ring a bell, maybe it's because three-letter acronyms are a dime a dozen and annoy me because I'm confronted with them regularly at work too (people at my company seem to love acronyms), but that's obviously on me too...

> Maybe my post sounded harsher than I intended

Appreciate the clarification. For me it was the "F" in "WTF" that tipped me. Other than that, it's more than fair for you to not know what GLM is. Things are moving so fast that I would be surprised if anyone can keep track of it all. Cheers, have a grand day!

It didn't read as harsh. Only unaware and you broadcasted that you don't have the decency to do basic searches.

It's presumptuous for them to assume that a reader of their blog is familiar with their product?

Also I feel like the obvious way to read the very first sentence is that GLM is a language model

> As we develop GLM, the model sometimes exhibits capabilities that surprise us

It's only the top open-weights LLM in the world,

https://artificialanalysis.ai/#intelligence-category-tabs

Ziphu, aka Z.ai, is the company that makes GLM (a very competitive Chinese LLM).

Why would you be reading their corporate blog posts if you don't even know who they are?!

A ai model family similar to Codex, Gemini or Claude.

Where GLM-5.3-Flash is the newest "small / fast" model.

Except there is no ai model family called Codex or Claude. There is Gemini, I give you that.

>As we develop GLM, the model sometimes exhibits capabilities that surprise us, and even unsettle us.

Come on now

Also, why would they introduce themselves on their own blog?

If you had bothered to at least skim TFA, this is addressed in the third paragraph:

> The idea is not to turn modern cargo ships back into sailing vessels. Instead, these systems are designed to work alongside conventional engines, using the wind to reduce fuel consumption whenever conditions allow.

News articles to this effect have been popping up occasionally for the past 20 years at least. So the economics don't seem to stack up. At least not yet; but if oil gets a bit dearer...

They really are being built and used commercially. Probably incentives play a part in this, but their numbers are slowly increasing, so it's not totally vaporware (of course we don't need new articles every ~6 month talking about how amazing this technique is) [0]

[0] https://en.wikipedia.org/wiki/Rotor_ship#Modern_vessels

Some of these sails claim 10% fuel reduction, which is ~5% operational cost, or ~150k per trip to spend on some tubes of metal or sheets of cloth.

Is that 10% best case or average case? Also, what's the capex on the the sails, and what's the additional opex (training, maintenance)? Not trying to pooh-pooh the idea, but as said, I've seen articles like this regularly for decades, and it hasn't become the norm, so I'd figure there's a reason for that.

I've seen a few articles about Norsepower (mentioned in the top article) validating the shipboard measurement equipment meant to confirm how much is actively being saved by their sails. I haven't seen any hard numbers on what the actual savings are, but they've had at least one ship (Chinook Oldendorff) with sails since at least 2024. So hopefully the public will get some concrete numbers on how much is saved.

Given the increasing number of companies that are mentioned as upcoming partners for actual installations, it might be that they've finally been able to convince shippers that wind power is viable for cost savings.

Sorry, typo, and unclear. I meant to say fuel is ~50% operational cost.

The issue is that when conditions do not allow they actively increase fuel consumption through windage, maneuverability loss, air draft increases and extra weight. They also increase costs through increased maintenance needs and crew training and limits on which docks can load and unload them.

We have had hybrid sail/combustion engine cargo ships for 200+ years, today they are reduced to novelty or historical status. We have been experimenting with goofy hi tech rotary sails for more than a century, and no one has really done more than make a few demonstrators for publicity.

You just have to take a look at the Rube Goldberg-esque abomination that is the AWS web interface for a few minutes to realize that it's not that simple. Granted, most executives won't do that...

Or encounter their CLI and it's lovecraftian combination of positional arguments, named arguments, piping, json documents, etc, distributed largely by random lot, as far as I can tell.

Especially with Amazon, who are well known for squeezing every last bit of profit from their employees, contractors etc., it doesn't really sound surprising. "Offsite backups?! Sure, you could have had that if you had found the right page in the AWS console and if you would have paid 50% extra!"

Oh sorry, you picked "AWS Backup" but you actually needed to use "Backup AWS" to solve that problem. Perhaps hop on a call with our sales engineering and cost magnification teams to guide you to a better, more solutioned, tomorrow?

I sounded all legit until the comma after "solutioned". I suspect that might be a fake.

Perhaps, cheapness is always an factor but I'm potentially reading it as the customers perhaps not wanting data to move outside of the country and with AWS only have one datacenter in said country produced this result.

but the general assumption is your data isn't going to get lost if you use Amazon. so once the trust is gone it's gone.

It all depends on how *off* off-site really is. You know, *off* isn’t a binary, it is a spectrum.

I think that's overly negative. Roundabouts have their place, mostly for low traffic intersections, but I agree that they tend to be overused, also here in Europe. Two examples where roundabouts are not the right solution:

- when there is too much traffic: if there is a constant stream of cars in the roundabout so you have to wait for a long time before being able to enter it, or force your way in (risking a fender bender), that's

- when you actually want traffic to flow. My small city has built a bypass road which is however IMHO completely botched: it starts with a roundabout, and the first exit leads to the route going through the center, only the second one goes to the bypass, so people are more likely to drive through the town. And it has three more roundabouts along the route, further making the option unattractive (in addition to a relatively low speed limit of 60 km/h and several tight curves).

> - when there is too much traffic: if there is a constant stream of cars in the roundabout so you have to wait for a long time before being able to enter it, or force your way in (risking a fender bender), that's

I think it's fine with lots of traffic in a roundabout, but it has to be somewhat balanced.

For example, we have a roundabout that sits on a relatively popular road, with one exit/entry that leads to a bigger highway. During rush hour, there is heavy traffic, from from/to every one of the three points, so traffic naturally divides, as cars turn off.

However, when there is no rush hour, the highway exit/entry isn't used, so cars mostly just pass through the roundabout, and everything still works, as the two entries/exits are still balanced.

Issues would arise if it's only one entry/exit being used, as then it'd always block the two other entries from being used properly, but seems that never happens, so seems to me it's proper and works well, regardless of the traffic being heavy.

Well, it doesn't have to be mutually exclusive, does it? An old building doesn't have to be run down, and a new one can be both expensive and shoddily built.

If you think this book is too big, I'd like to remind you that there are one-volume paperback editions of The Lord Of The Rings (I have this one: https://www.worthpoint.com/worthopedia/r-tolkien-lord-rings-...) with 1100+ pages.

I think War and Peace is the longest book I've ever read. Most editions have 1300-1400 pages.

One contributing factor might be that AI companies are raising money via (amongst other methods) also issuing bonds, which might compete with government bonds.

Civilian targets heavily involved in handling "dual-use" items that the Russian army isn't providing to its soldiers, so they have to buy it by themselves (https://factually.co/fact-checks/business/wildberries-milita...).

...this after Russia has been targeting civilian targets in Ukraine from the first days of the invasion.

Sadly, the success rate is significantly better than 0.1%. as a broad strategy, it clearly works. I wish it didn't, but there is a reason why all these companies do it.