Slightly OT, but after using DeepSeek V4 Pro for the last few weeks, I’ve found that it’s basically on par with Opus…except when it comes to driving Blender. This isn’t even a visual issue (DS isn’t multimodal), for whatever reason Opus just understands the Blender API a lot better.

There always seem to be pockets where closed frontier models perform slightly better.

This is why I keep a z.ai subscription as a backup.

is GLM genuinely comparable to claude models? haven't had a chance to test it yet.

I’m not going to pretend it’s on par, it’s clearly a step behind when it comes to thinking and planning.

However, it’s a very good “doer” - if I give it a list of well-scoped tasks, I can usually rely on it to execute them all correctly.

I've been working on Mixreel, a video/motion graphics editor with integrated support for 3D visualization. I'm working with some business clients to produce instructional videos for construction, industrial fabrications, etc.

https://mixreel.ai

I was apartment hunting in Melbourne in 2015 and I was appalled at the quality of most inner-city apartments. Tiny shoeboxes, no sunlight, paper thin walls.

At the time I didn't think they should have been allowed to be built. But looking back, they probably did keep a lid on rents. A bad roof over your head is better than no roof.

I’ve never delved fully into IP law, but wouldn’t these be considered derivative works? They’re basically just reimplementing exactly the same functionality with slightly different names?

This would be different from the “API reimplementation” (see Google vs Oracle) because in that case, they’re not reusing implementation details, just the external contract.

Because copyrights do not protect ideas. Thankfully. We are free to express ideas, as long as we do so in our own words. How that principle is applied in actual law, and how that principle is a applied to software is ridiculously complicated. But that is the heart of the principle at play here. The law draws a line between ideas (which cannot be copyrighted), and particular expressions of those ideas (e.g. the original source code), which are protected. However, it is an almost fractally complicated line which, in many place, relies on concepts of "fairness", and, because our legal system uses a system of legal precedence, depends on interpretation of a huge body of prior legal decisions.

Not being a trained lawyer, or a Supreme Court justice, I cannot express a sensible position as to which side of the line this particular case falls. There are, however, enormously important legal precedents that pretty much all professional software developers use to guide their behaviour with respect to handling of copyrighted material (IBM vs. Ahmdall, and Google v. Oracle, particularly) that seem to suggest to us non-lawyers that this sort of reimplementation is legal. (Seek the advice of a real lawyer if it matters).

Taking a step back, it seems fairly clear that wherever you set the bar, it should be possible to automate a system that reads code, generates some sort of intermediate representation at the acceptable level of abstraction and then regenerates code that passes an extensive set of integration tests … every day.

At that point our current understanding of open source protections … fails?

Depends whether you sit on the MIT half of open source, or the GPL side of open source, I suppose.

there's usually a test for originality, and it involves asking things (from the jury) like, is it transformative enough?

so if someone tells the LLM to write it in WASM and also make it much faster and use it in a different commercial sector... then maybe

since 2023 the standard is much higher (arguably it was placed too low in 1993)

> If they're doing it against the terms of service (and publicly so), I can't pin that one on Anthropic.

Anthropic claim that superintelligence is coming, that unaligned AI is an existential threat to humanity, and they are the only ones responsible enough to control it.

If that's your world view, why would you be willing to accept someone's word that they'll only Do Good Things with it? And not just "someone", someone with access to the world's most powerful nuclear arsenal? A contract is meaningless if the world gets obliterated in nuclear war.

Anybody who works with the military has to deal with that moral dilemma. Many people believe that the military has some legitimate use. They have to figure out for themselves how do deal with the the possibility that it can also be used illegitimately.

So I don't blame Anthropic for getting into bed with the military, and getting out when it got bad for them. A lot of military suppliers are facing a similar dilemma, I suspect. The army runs on its stomach, and I do not envy the people delivering pizzas to the Pentagon, knowing what room those pizzas are consumed in.

“Roughly the same” != “the same”. Changing one or two words words in a contract can make a huge difference.

As a complete outsider, I genuinely believe that Dario et al are well-intentioned. But I also believe they are a terrible combination of arrogant and naive - loudly beating the drum that they created an unstoppable superintelligence that could destroy the world, and thinking that they are the only ones who can control it.

I mean if you sign a contract with the Department of War, what on Earth did you think was going to happen?

Not this, because this is completely unprecedented? In fact, the Pentagon already signed an Anthropic contract with safe terms 6 months ago, that initial negotiation was when Anthropic would have made a decision to part ways. It was totally absurd for the govt to turn around and threaten to change the deal, just a ridiculous and unprecedented level of incompetence.

> was totally absurd for the govt to turn around and threaten to change the deal, just a ridiculous and unprecedented level of incompetence.

I think in this case it's safe to assume malice rather than incompetence. It's a lot like the parable of the frog and the scorpion.

Government always has the option to cancel contracts for convenience, they knew what they signed up for or else they were clueless and shouldn’t be playing with DoD

The keyword is "cancel", not threaten seizure with the DPA and destruction with a baseless supply chain risk designation.

If they made a completely private nuclear reactor and ended up with a pile of weapons grade plutonium, what do you think the department of war would do? It was completely obvious it would happen, as it will be not surprising when laws are passed and all involved will have choose between quit or quit and go to jail. There are western countries in which you’d just end up in a ditch, dead, so they should think themselves lucky for doing the ai superintelligence thing in the US.

The US government clearly doesn't take seriously the claim that AI is more dangerous than (or even as dangerous as) nukes, because if they did they wouldn't allow anyone except the military to develop or use them, they wouldn't allow their export or for them to be made available for use by foreigners like me, they wouldn't allow their own civilians to use them, they would probably be having a repeat of the cases in the cold war where they tried to argue certain inventions were "born secret" and could not be published even if they were developed by people who were not sworn to secrecy.

I don't think the US has ever done/threatened anything like this to a US company so it's not surprising that Anthropic were caught off guard.

The less conspiratorial view is that the Australian government was the first to actually legislate it. Much easier for everyone else to follow in someone else's footsteps rather than being the first to take the plunge.

I think MoltenVK probably is the translation layer they're using.

it would be cool to use cosmickrisp instead at some point... in fact i saw someone run minecraft on macos using cosmickrisp -> zink to have modern opengl features that macos does not have