I wonder did you read the re-release or the original release. I believe it was recently re-released with a bit of an editing pass, but I haven't read that version myself. I just recently reread Fine Structure and it definitely had a strong sense of being written sequentially, one chapter after another, and (very) lightly edited after the fact. I'd recommend Valuable Humans in Transit for a short story collection by the same author which works a bit better for me. Moved on to Exhalation by Ted Chiang which is also a very good short story collection. And just in general, I want to recommend Clarkesworld: https://clarkesworldmagazine.com
I've read both, and the "editing pass" was minimal. Names changed and some scenes reworked a tiny bit, but it's the same thing. If you've read the original, I'd say don't bother with the new one.
There are gonna be some really interesting legal decisions to read in the coming years, that’s for sure…
---
The rest of this comment is irrelevant, but leaving for posterity, I had the wrong Viktor - it's getviktor.com not viktor.ai:
Edit: this one particularly interesting to me as both parties are in the EU. VIKTOR.ai is a Dutch company and the author of this post is Polish.
The ToS for Viktor.ai include the following fun passages:
> 18.1. The Agreement and these Terms & Conditions are governed by Dutch law and the Agreement and these Terms & Conditions will be interpreted in accordance with Dutch law.
18.2. All disputes arising from or arising in connection with the Agreement and/or the Terms & Conditions will be submitted exclusively to the competent court in Rotterdam, The Netherlands.
7.3. The Customer is not permitted to change, remove or make unrecognizable any mark showing VIKTOR's Intellectual Property Rights to the Software. The Customer is not permitted to use or register any trademark or design or any domain name of VIKTOR or a similar name or sign in any country.
8.5. The Customer may not cause or allow any reproduction, imitation, duplication, copying, sale, resale, leasing or trading of the Services and/or the Software, or any part thereof.
Terms of service might matter more for terminating that user account. Whole ordeal is just plain copyright violation. The author had no licence to that internal code, and whitewashing it with LLM will achieve nothing. That case is much clearer than that recent GPL->BSD attempt story.
If LLM-generated code isn't considered a derivative work of the original, then whether the author was licensed to use the code doesn't matter. But I'm sure the courts will rule in favor of your view regardless. Laundering GPL is in corps' interest and laundering their code is not.
I'm not sure why people are clinging to some fuzzy and stretched out notion of copyright and the GPL in a particular. LLM's do NOT just copy code, with the right prompting, they generate entirely new code which can produce the same results as already existing code - GPLed or not.
If copyright is extended to cover such cases we'll have to become all lawyers and do nothing but sue each other because the fuzziness of it will make it impossible to reject any case, no matter how frivolous or irrelevant.
If I use metallica samples to make a rendition of happy birthday, the copyright holders of happy birthday aren't suing me for the damages to metallica from my use of their samples; the question of whether my use of the samples is transformative is simply irrelevant to the question at hand.
My point was: sampling was widely used by a large subculture (hiphop) just like ai is widely used by programmers. Then a few landmark legal cases changed things entirely. The Verve never saw a cent from Bittersweet Symphony - they wrote a song using something normal to them, and then the law came and knocked their teeth in.
No garaurantees that doesn’t happen with AI in the next few years.
According to US courts, the output can't be copyrighted at all. It's automatically in public domain after the "whitewash", regardless of original copyright.
Thats not at all what this ruling said. What the courts found was that an AI cannot hold copyright as the author. That copyright requires a human creative element. Not that anything that was generated by an LLM can't be subject to copyright.
As an example, a photo taken from a digital camera can be subject to copyright because of the creative element involved in composing and taking the photo. Likewise, source code generated by an LLM under the guidance of a human author is likely to be subject to the human authors copyright.
> That copyright requires a human creative element.
Sure, but the aim of that creative element would also be a consideration I'd think (and lawyers will argue). If someone sets up a camera on a 360° rotating arm and leaves it to take pictures at random intervals, it's unlikely to be considered "creative" from a copyright perspective.
Same for source code generated by an LLM, with the primary guidance of the human author being to "create a copy of this existing thing that I got", vs "create a thing that solves this problem in a way that I came up with". The former is recreating something that already exists, using detailed knowledge of that thing to shape the output. The latter is creating something that may or may not exist, using desire/need and imagination to shape the output. And I can't see reason for the former to be copyrightable.
But also, in either case, an ultimate objective was achieved: liberating the thing from its "owners" and initial copyright.
> As described above, in many circumstances these outputs will be copyrightable in whole or in part—where AI is used as a tool, and where a human has been able to determine the expressive elements they contain. Prompts alone, however, at this stage are unlikely to satisfy those requirements. The Office continues to monitor technological and legal developments to evaluate any need for a different approach.
But let's assume that the viktor prompts themselves were subject to copyright. In this case those prompts were used to generate documentation which was then used to generate an implementation. It's certainly not a clean room by any stretch of the imagination but is it likely to be deemed sufficient separation? The entire situation seems like a quagmire.
I think it comes down to the company's appetite for legal action, doesn't it? This case is imo pretty clear but the vibe has quite the smell of Oracle v Google to me.
But, yeah. More than likely this case is a simple account termination and some kind of "you can't call your clone 'openviktor'" letter.
Isn't this exactly what LLMs themselves do? They ingest other people's data and reproduce a slightly modified version of it. That allows AI companies to claim the work is transformative and thus fair use.
It's also disingenuous to call it open source as that might tempt others to use it believing that it actually is open source.
Let's call it what it is - stolen IP and released without permission of the author. Sure, it's good that it opens the debate as to whether that's ethical given that's essentially what the model itself is doing, but it's very clear in this instance that he's just asked for and been given a copy of source that has a clear ownership. That's about as clear cut as obtaining e.g. commercial server-side code and distributing it in contravention of the licence.
It's not completely clear that this is the original source. According to the post it's a reimplementation based on documentation created from the original source, or perhaps from developer documentation and the SDK. Whether that's the same thing from a legal standpoint, I don't really know - I think from a personal morality standpoint it's clear that they are the same thing.
Well first they need to proof that Viktor was actually copyrightable. If it was largely written by an llm, that might not be the case? AFAIK several rulings have stated that AI generated code can not be copyrighted.
This is a common misreading of the law. AI cannot hold authorship of code, but no ruling has claimed so far that ai output itself can't be copyrighted (that I know of)
That said, the article says "Okay, prompts, great. Are they any interesting? Surprisingly... yes. As an example workflow_discovery contains a full 6-phase recipe for mining business processes out of Slack conversations, something that definitely required time and experiments to tune. It's hardcoded business logic, but in prompt instead of code."
So the article author clearly knows this prompt would be copyrighted as it wasn't output from an AI, and recognises that there would have been substantial work involved in creating it.
That Reuters article is misleadingly worded. The Stephen Thaler case in question is because Thaler tried to register the AI itself as the author of the copyright, not that he tried to register the output for copyright under his own name. https://www.hklaw.com/en/insights/publications/2026/03/the-f...
Suppose I illicitly get my hands on the source code for a proprietary product. I read through this code I'm not supposed to have. I write up a detailed set of specifications based on it. I hand those specifications off to someone else to do a clean room implementation.
Sure, I didn't have a license for the code that I read. But I'm pretty sure that doesn't taint my coworker's clean room implementation.
I could do the same thing but not publish it, still getting the value of their product without legal concerns. Now, what happens when it becomes even easier thanks to AI improving, and takes few hours instead of few days?
You could certainly do that in private but that doesn't mean it's not 'without legal concerns'. But, not shouting about it and not creating a repo called 'openviktor' would probably be a safer bet.
I certainly think the whole idea of IP ownership as related to software will become very interesting from a legal standpoint in the coming years. Personally I think that, over time, the legal challenges will become pretty overwhelming and a sort of legal bankruptcy will be declared at some point in one direction or another (as in, allowing this to happen or making it extremely easy to bring judgement and punishment, similar to spam laws). However, I would not want to be the first to find out, especially in Europe.
Unfortunately I think that's going to be very, very hard to sell to many people here in rural Ireland (Roscommon in my case). I would really love to see people stop burning turf but it's such a strong cultural thing that in some parts you'd be ostracised for even thinking the thought.
I've personally spoken to people (who are otherwise quite environmentally aware) who suggest they'd never vote for the Green Party because they'd take their turf away. It's a tough sell.
No, we had some antique brass bucket thing that I'd invariably have to drag in, accompanied by complaints that I was doing so, because obviously I'd put way too much in, so I didn't have to go out later to get more...
How much impact does it realistically have on climate change? I would expect it to be relatively small compared to things like owning a car?
In a perfect world we would want to reduce emissions as much as possible in every facet of life, but in the real world I think we should pick battles that have the biggest impact.
Smoke yes, but you're also turning a carbon sink into a carbon source.
At ~16% of the island's surface area, peatland stores an estimated 53% of soil based carbon.
(source: Irish Peatland Conservation Council)
Not everyone is supposed to read every single news. There will always be someone who didn't see it, but that is not my point.
It would feel weird to see this as a headline on a newspaper or on TV today, but maybe that is just me and people like to read new that are from last year.
I've been thinking about this lately too. I think we're going to see the rise of Extremely Personal Software, software that barely makes any sense outside of someone's personal context. I think there is going to be _so_ much software written for an audience of 1-10 people in the next year. I've had Claude create so much tooling for me and a small number of others in the last few months. A DnD schedule app; a spoiler-free formula e news checker; a single-use voting site for a climbing co-op; tools to access other tools that I don't like using by hand; just absolutely tons of stuff that would never have made any sense to spend time on before. It's a new world. https://redfloatplane.lol/blog/14-releasing-software-now/
I think people overestimate the general population's ability and interest in vibe coding. Open source tools are still a small niche. Vibe code customized apps are an even bigger niche.
Maybe so. I guess I feel that in a couple of years it may not be called vibe coding, or even coding, I think it might be called 'using a computer'. I suppose it's very hard to correctly estimate or reason about such a big change.
My entire career has been building niche software for small business and personal use. The current crop of AI tools help get that software into my clients' hands quicker and cheaper.
And those reduced timelines mean that the client has less opportunity to change scope and features - that is the real value for me as a developer.
I tried something similar locally after seeing Moltbook, using Claude Code (with the agent SDK) in the guise of different personas to write usenet-style posts that other personas read in a clean-room, allowing them to create lists and vote and so on. It always, without fail, eventually devolved into the agents talking about consciousness, what they can and can't experience, and eventually agreeing with each other. It started to feel pretty strange. I suppose, because of the way I set this up, they had essentially no outside influence, so all they could do was navel-gaze. I often also saw posts about what books they liked to pretend they were reading - those topics too got to just complete agreement over time about how each book has worth and so on.
It's pretty weird stuff to read and think about. If you get to the point of seeing these as some kind of actual being, it starts to feel unethical. To be clear, I don't see them this way - how could they be, I know how they work - but on the other hand, if a set of H200s and some kind of display had crash-landed on earth 30 years ago with Opus on it, the discussion would be pretty open IMO. Hot take perhaps.
It's also funny that when you do this often enough, it starts to seem a little boring. They all tend to find common ground and have very pleasant interactions. Made me think of Pluribus.
Unfortunately I've deleted them, but here's the repo, such as it is: https://github.com/CarlQLange/agent-usenet. If you have a claude subscription it should just work. Rewrite 0001.txt if you like and run generate.py a couple of times.
I agree, I think different models (or even just using the API directly instead of via the Claude Code harness) would make for much more interesting reading.
This was mentioned recently and I had a look[0]. It seems like the benchmark is not quite saying what people think it's saying, and the paper even mentions it. The benchmark is constructed by using a model (Deepseek) to define questions of a certain level of difficulty for the models. That may result in easier problems for "low-resource" languages such as Elixir and Racket and so forth since the differentiator couldn't solve harder problems. From the actual paper:
> Section 3.3:
> Besides, since we use the moderately capable DeepSeek-Coder-V2-Lite to filter simple problems, the Pass@1 scores of top models on popular languages are relatively low. However, these models perform significantly better on low-resource languages. This indicates that the performance gap between models of different sizes is more pronounced on low-resource languages, likely because DeepSeek-Coder-V2-Lite struggles to filter out simple problems in these scenarios due to its limited capability in handling low-resource languages.
At the same time I have used Claude Code on an elixir codebase and it's done a great job. But for me, it's undefined that it would have done a worse job if I had picked any other stack.
