PSA: Please note that the names are hallucinated author lists part of the hallucinated citations, and not names of offending authors.

AFAIK the submissions are still blinded and we don't know who the authors are. We will, surely, soon -- since ICLR maintains all submissions in public record for posterity, even if "withdrawn". They are unblinded after the review period finishes.

Code search over all of Gitlab (even if available) wouldn't help much when many of the interesting repos might be on Github. To be truly useful, it would need to index repos across many different forges. But there's a tension in presenting that to users if you're afraid that they might exit your ecosystem to go to another forge.

This seems to be the app: https://www.sancharsaathi.gov.in/

Looks like it's quire popular/established already, with over 10 million downloads. Basically a "portal" for basic digital safety/hygiene related services.

Quoting Perplexity regarding what facilities the app offers:

1. Chakshu: Report suspicious calls, SMS, or WhatsApp for scams like impersonation, fake investments, or KYC frauds.

2. Block Lost/Stolen Phones: Trace and block devices across all telecom networks using IMEI; track if reactivated.

3. Check Connections in Your Name: View and disconnect unauthorized numbers linked to your ID.

4. Verify Device Genuineness: Confirm if a phone (new or used) is authentic before purchase.

How does an app inspect other app's storage data (like whatsapp). I thought Android security model blocked that. Does it have root access?

It probably just asks you to enter the associated WhatsApp number

Every single Indian SIM holder got dozens of SMS from the regulator to push the app installations. When your marketing campaign is “Notify every Indian SIM holder”, 10M should be expected. Look at the reviews.

Oh thats why india scams the rest of the world, we just dont have their apps to report it properly

Interesting question. I don't know much about WebGPU, but I'd posit (heh!) that the GPU on the client devices doesn't matter too much since folks will likely be working over the network anyways (cloud-based IDE, coding agent connected to cloud-hosted LLM, etc) and we also have innovations like Modal which allow serverless lambdas for GPUs.

As long as silicon is scarce it would make sense to hoard it and rent it out (pricing as a means of managing scarcity); if we end up in a scenario where silicon is plentiful, then everyone would have powerful local GPUs, using local AI models, etc.

I guess in my mind I was thinking use cases other than AI. Like statistical or hierarchical scientific models, simulations or ETL work. I also don't know if some of the econometricians I know with a less technical background would even know how to get setup with AWS, and I feel more boardly there's enough folks doing data work in a none tech field who know how to use Python or R or Matlab to do their modelling but likely isn't comfortable with cloud infrastructure, but might have an apple laptop with apple silicon that could improve their iteration loop. Folks in AI are probably more comfortable with a cloud solution.

There are aspects of data science which is iterative and you're repeatedly running similar computations with different inputs, I think there's some value in shaving off time between iterations.

In my case I have a temporal geospatial dataset with 20+ million properties for each month over several years each with various attributes, it's in a nonprofessional setting and the main motivator for most of my decisions is "because I can and I think it would be fun and I have a decent enough GPU". While I could probably chuck it on a cluster, I'd like to avoid if I can help it and an optimisation done on my local machine would still pay off if I did end up setting up a cluster. There's quite a bit of ETL preprocessing work before I load it into the database, I think are portions that might be doable on the GPU. But it's more so the computations I'd like to do on the dataset before generating visualisations in which I think I could reduce the iteration wait time for processing for plots, ideally to the point I can make iterations more interactive. There's enough linear operations you could get some wins with a GPU implementation.

I am keen to see how far I'll get, but worst case scenario I learn a lot, and I'm sure those learnings will be transferrable to other GPU experiments.

TBC, I too did not really mean "AI" (as in LLMs and the like) which is often hosted/served with a very convenient interface. I do include more bespoke statistical / mathematical models -- be it hierarchical, Bayesian, whatever.

Since AWS/etc are quite complicated, there are now a swarm of startups trying to make it easier to take baby steps into the cloud (eg. Modal, Runpod, etc) and make it very easy for the user to get a small slice of that GPU pie. These services have drastically simpler server-side APIs and usage patterns, including "serverless" GPUs from Modal, where you can just "submit jobs" from a Python API without really having to manage containers. On the client side, you have LLM coding agents that are the next evolution in UI frontends -- and they're beginning to make it much much easier to write bespoke code to interact with these backends. To make it abundantly clear what target audience I'm referring to: I imagine they are still mostly using sklearn (or equivalents in other languages) and gradient boosting with Jupyter notebooks, still somewhat mystified by modern deep learning and stuff. Or maybe those who are more mathematically sophisticated but not software engg sophisticated (eg: statisticians / econometricians)

To inspire you with a really concrete example: since Modal has a well documented API, it should be quite straight-forward ("soon", if not already) for any data scientist to use one of the CLI coding agents and

1. Implement a decent GPU-friendly version of whatever model they want to try (as long as it's not far from the training distribution i.e. not some esoteric idea which is nothing like prior art)

2. Whip up a quick system to interface with some service provider, wrap that model up and submit a job, and fetch (and even interpret) results.

----

In case you haven't tried one of these new-fangled coding agents, I strongly encourage you to try one out (even if it's just something on the free tier eg. gemini-cli). In case you have and they aren't quite good enough to solve your problem, tough luck for now... I anticipate their usability will improve substantially every few months.

Opencode (from SST; not the thing that got rebranded as Crush) seems to be just that. I've had a very good experience with it for the last couple of days; having previously used gemini-cli quite a bit. Opencode also has/hosts a couple of "free" models options right now, which are quite decent IMO.

https://github.com/sst/opencode

There are many many similar alternatives, so here's a random sampling: Crush, Aider, Amp Code, Emacs with gptel/acp-shell, Editor Code Assistant (which aims for an editor-agnostic backend that plugs into different editors)

Finally... there is quite a lot of scope for co-designing the affordances / primitives supported by the coding agent and the LLM backing it (especially in LLM post-training). So factorizing these two into completely independent pieces currently seems unlikely to give the most powerful capabilities.

> created by monkeys

I don't particularly care for either Zig or Github, but...

they do precisely cite the technical issues. That snippet links to a Github discussion comment https://github.com/actions/runner/issues/3792#issuecomment-3...

(reproduced below)

"The bug in this "safe sleep" script is obvious from looking at it: if the process is not scheduled for the one-second interval in which the loop would return (due to $SECONDS having the correct value), then it simply spins forever. That can easily happen on a CI machine under extreme load. When this happens, it's pretty bad: it completely breaks a runner until manual intervention. On Zig's CI runner machines, we observed multiple of these processes which had been running for hundreds of hours, silently taking down two runner services for weeks."

"I don't understand how we got here. Even ignoring the pretty clear bug, what makes this Bash script "safer" than calling into the POSIX standard sleep utility? It doesn't seem to solve any problem; meanwhile, it's less portable and needlessly eats CPU time by busy-waiting."

"The sloppy coding which is evident here, as well as the inaction on core Actions bugs (in line with the decay in quality of almost every part of GitHub's product), is forcing the Zig project to strongly consider moving away from GitHub Actions entirely. With this bug, and many others (severe workflow scheduling issues resulting in dozens of timeouts; logs randomly becoming inaccessible; random job cancellations without details; perpetually "pending" jobs), we can no longer trust that Actions can be used to implement reliable CI infrastructure. I personally would seriously encourage other projects, particularly any using self-hosted runners, to look carefully at the stability of Actions and ask themselves whether it is a solution worth sticking with long-term when compared with alternatives."

----

I agree that the writing in the blog post is more colorful than precise, but sanitizing every bit of expression dulls the internet. Humans invented language for a reason.

Then blast the product, not the people who built it.

They are blasting the product tbf. The people part is a small part of it. And apparently at least distracting the HN Community from their point.

Which is exactly why to cut it out. If you put salt in my cup of tea, I’m gonna notice and it’s gonna ruin the drink.

Microsoft poured salt into your cup years ago, you just did not notice.

If you put more salt into this rather thinly-stretched metaphorical cup when telling me what Microsoft did you are not going to endear yourself to me. Why muddy your message?

You cannot divorce a product from the people who built it. The product reflects their priorities and internal group well-being. A different group of people would have built a different product.

If you've worked in a large company, you know that the product reflects the priorities of the company so much more than the people who work there. Leadership states the priority and the employees do what they're told.

Leadership is part of the group of people who built the product, therefore different leadership would have also built a different product.

With that said, it's also not correct to claim that line folk have no influence at all. I don't believe that you can blame any individual since they may have stood up against something bad being put in the product, but they're still part of a collective group of people that built a bad product.

There's no stupid product, only stupid people.

The product was made by people. Or by AI which was made and controlled by people.

The product isn't some result of a series of "oopsies". The worst aspects of bad and/or user-hostile software products are that way because the people working at these companies want them to be that way.

Unless you want to call them just that incompetent. I assume they'd complain about that label too.

In short: No it's not "the product", the people building it are the problem. Somehow everyone working in big tech wants all the praise all the time, individually, but never take even the slightest bit of responsibility fro the constant enshittification they drive forward..

If you only knew...

I must be missing something huge here, or maybe it's the wine -- how is the code in PR 3157 (referenced in a later comment) a proper fix?

https://github.com/actions/runner/pull/3157/files

Is : doing something unusual in GH actions?

The original loop is:

while (time() != timeout) {;}

The fixed loop is:

while (time() < timeout) {;}

I see. I did not realize SECONDS was a built in bash variable.

It is still not a proper fix. It is still busy-looping 100% CPU.

Given that Github Actions is quite popular, probably wasting large amount of energy.

But probably good at generating billable Actions minutes.

One can only hope that not many people use sleeps to handle their CI race conditions, as that itself is also not a proper fix.

Clearly the job for a microservice. Accept number of seconds to wait as url, return content after that many seconds. Then just use curl in runner.

Brb founding a SaaS startup. I’ll call it cloudsleep dot io of course. After our Series B I’ll buy the .com.

Only task to do before lining up investors is how can I weave AI into our product?

Describe the task you're waiting for as text, and let an LLM pick the number of seconds for each request. More expensive the better model you clearly need for this. There, your AI pitch.

Excellent, welcome aboard, Chief Product Officer!

Retries won’t work in that case. Would be better to have two endpoints: get the time in x seconds and wait until time passed. That way retrying the wait endpoint will work fine and if time hasn’t elapsed it can just curl itself with the same arguments.

If you have curl (but not sleep) sure, but if not maybe you can use bash's wacky /dev/tcp. The microservice could listen on ports 1 through 64k to let you specify how many seconds to sleep.

Yeah, definitely not a proper fix.

Maybe a more serious fix is something like "read -t $N". If you think stdin might not be usable (like maybe it will close prematurely) this option won't work, but maybe you can open an anonymous FD and read from it instead.

    while wait && [[ $SECONDS -lt $1 ]]; do
      read -t $(($1 - SECONDS))
    done <><(:)

although if you're not too concerned about finishing early if a signal interrupts, probably

    { wait; read -t $1; } <><(:)

would be fine. You want the wait because otherwise bash won't reap the zombie from the process substitution until after the read times out.

Interestingly, it does reap on a blocked read without -t, so potentially the behaviour on -t would be considered a bug rather than as-designed.

There's also a loadable sleep builtin supplied with bash which calls into the internal fsleep() so should be reliable and without forking.

It's just all hacks. It should use `sleep`, period.

Sleeping is an OS scheduling task. Use the OS syscall that does that.

As is suggested on the Github issue that Microsoft has been ignoring for half a year.

Me neither. I am over 40 and did not know this. Feels good to learn something today, in an unexpected place!

> I agree that the writing in the blog post is more colorful than precise, but sanitizing every bit of expression dulls the internet. Humans invented language for a reason.

Where do you draw the line, then? Is a racist screed acceptable to you as long as the following paragraph references technical issues correctly?

The language in the blog post is insulting. Imagine how you would feel if you were the person who wrote this code, and now you are being called a monkey in front of thousands of people on the internet. Certainly you've made mistakes in your code before...or are you saying you've always written flawless code without errors?

These codes of conduct always seemed a bit superfluous to me, but after reading comments like these I can totally see why they are necessary.

Would you perhaps have preferred if they referred to it as "unprofessional" or "sloppy" instead alluding of monkeys?

To me all those mean the same thing, except the latter is more flavorful and makes my eyes less likely to glaze over.

> Imagine how you would feel if you were the person who wrote this code, and now you are being called a monkey in front of thousands of people on the internet.

Er.. so? Why should anyone be allowed into a position of responsibility where their code impacts millions of people if they can't handle the tiniest bit of strong feedback? It was, after all, a pretty egregious bug.

> Certainly you've made mistakes in your code before...or are you saying you've always written flawless code without errors?

I've definitely made mistakes, and also accept that my output might have on occasion been "monkey-esque". I don't see what's insulting about that; we are all human/animal.

> To me all those mean the same thing, except the latter is more flavorful and makes my eyes less likely to glaze over.

And to many others, the difference is that one is informative, the other is likely to turn them off of the author and project forever.

I noticed that you never answered my question. If this is acceptable to you, where do you draw the line? If you can answer that question, maybe you'll be able to see the flaw in your argument.

> the other is likely to turn them off of the author and project forever

Which is absolutely fine. It's their project, their website. If they can't be colorful on their own website, where else can they be! If it turns off some people, I'm sure the author is aware of the risk and happy with that risk.

I, for one, find this kind of colorful language refreshing. Everyone trying to be politically correct makes the internet a dull place.

not being an asshole != political correctness

Surely you have your own line on what is or is not acceptable discourse. What is it?

> Surely you have your own line on what is or is not acceptable discourse. What is it?

I do but I decline to share it here. I'm not going to shift this thread from what the author is doing on their website to my personal beliefs and boundaries!

All I am saying is it is their project, their blog. They can be however much rude they want to be on their website. It's their website, their lines and their boundaries. Where I set my boundaries has no bearing on what Andrew should write on their website.

If Andrew alienates people by his writing, it's his decision, his action, his consequences that he has to deal with. How does it matter where I draw the line?

> All I am saying is it is their project, their blog. They can be however much rude they want to be on their website. It's their website, their lines and their boundaries.

That's funny, because if that is true he violated his own code of conduct: https://ziglang.org/code-of-conduct/#safe-constructive-only

> I do but I decline to share it here

The point is that everyone has different lines for what they consider to be "acceptable" or not. That is exactly the reason why codes of conduct exist - it's an attempt to find a common denomiator so that it can help foster a community where people can feel included without feeling like they are being attacked or insulted.

> That's funny, because if that is true he violated his own code of conduct

Yes, he did. It is funny. I don't know why we need to talk endlessly about it. If you are bothered so much by this violation, file an official report on their issue tracker.

> The point is that everyone has different lines for what they consider to be "acceptable" or not. That is exactly the reason why codes of conduct exist

When I said I decline to share my lines and boundaries here, I meant just that. I didn't mean that I need a lecture on CoC from you. I know what CoCs are and why they exist. Thank you very much. I am not morality police. Neither are you.

My morality applies to myself. Andrew's morality applies to himself. But yeah... CoC may apply to him too. So you've got a good point. I don't know if the CoC applies to their website. If you know more and if it does, a violation of CoC should be reported on their issue tracker. If this is such an important topic for you, please do report the violation to them. That'd be fair.

> Yes, he did. It is funny.

Yeah, it’s hilarious! Calling someone a monkey is such a clever and thought provoking insult!

> I don't know why we need to talk endlessly about it

If you are confused by this, why are you continuing to respond?

> When I said I decline to share my lines and boundaries here, I meant just that. I didn't mean that I need a lecture on CoC from you. I know what CoCs are and why they exist.

I really don’t think you know why CoC’s exist, because you are chastising people when they point out a legitimate violation (e.g. being the "morality police").

> But yeah... CoC applies to him too. So you've got a good point

Thanks for finally admitting this, I guess? Not sure why you needed to add all the extra argumentation about it, but at least you got there eventually.

> If this is such an important topic for you, please do report the violation to them

No thank you. I’m not actually offended by what he said, I just find it weird when people rush to his defense on this.

> If you are confused by this, why are you continuing to respond?

I'm not confused by anything. That was a rhetorical question. I continue to respond because there are other things that I care about and I have things to say about that. I don't care about what style or tone or words Andrew choses on their website. But I care about people trying to be morality police and discouraging someone blogging on their own website from writing rudely and writing politically incorrectly. So that's why I continue to respond.

> Thanks for finally admitting this, I guess? Not sure why you needed to add all the extra argumentation about it, but at least you got there eventually.

Credit where credit is due. If you make good points I agree with, I'll certainly say that.

> Not sure why you needed to add all the extra argumentation about it, but at least you got there eventually.

Because there are other points of yours I don't agree with.

Must a person always 100% agree or 100% disagree? Can a person not 10% agree and 90% disagree? The latter is happening here.

> But I care about people trying to be morality police and discouraging someone blogging on their own website from writing rudely and writing politically incorrectly

This appears to be a strawman. You already admitted he violated the CoC - so he is in the wrong here.

I'm not sure what else there is to disagree with - that's been my assertion from the beginning.

If he wants to write childish stuff on his own website that is not covered by the CoC, that's his choice. I'm also free to express my opinion on that, but I never implied that he shouldn't be able to write whatever he wanted on his own personal blog.

> You already admitted he violated the CoC - so he is in the wrong here

I didn't say that. This is what I said -

"But yeah... CoC may apply to him too. So you've got a good point. I don't know if the CoC applies to their website. If you know more and if it does, a violation of CoC should be reported on their issue tracker."

Emphasis: "may", "I don't know if", "If you know more".

You did say that. You performed a stealth edit and modified your comment, but fortunately I quoted what you originally said in my previous comment:

> But yeah... CoC applies to him too. So you've got a good point

Since you’ve just proven yourself to not be arguing in good faith, this will be my last response to you.

It's not a stealth edit. It's an open edit. HN allows edits for 2 hours for good reason. I misspoke first when I thought the CoC applies to him. Obviously I don't know for sure since I hadn't read the CoC. So I corrected myself to be less sure.

But you chose to reply to my outdated message although at the time you were replying my message said that I wasn't sure whether the CoC applies or not.

read the CoC carefully, it says which spaces are governed by it. the website does not seem to be. that's deliberate, the CoC only applies to "working" spaces.

If it was a company, you'd say the CoC is meant for lackeys, not for C-suites.

No, it's meant for interactions inside the company, not towards random giant corporations outside of it.

Made by monkeys and losers (because everyone else has left) does not target the company, but its employees.

> If that is how you feel, why are you spending multiple comments defending the language used?

I don't care much about the language used. I neither intend to defend it nor criticize it. But I do care about people trying to be morality police. That's why I am spending multiple comments here.

> It feels like there’s a reason you refuse to define your line in the sand.

Yes, the reason is that my line applies to myself. My line doesn't apply to you. It doesn't apply to Andrew. So my line, which is a personal and private matter for me, isn't something I want to share here. It is irrelevant when talking about the words Andrew chose on his website. That's the reason. It's a simple reason. Don't overthink it!

Not everyone is build robust enough for being called a monkey or loser straight into the person's face.

My first thought when reading the article is that the author would never attempt to call the person a monkey or loser to their face "in real life".

One of the unfortunate side effects of the internet is that it really brings out the asshole in a lot of people.

Maybe they would be happier working a job that requires zero accountability.

The website is not covered by the CoC, so they are not in abeyance

Ziglang.org disconnected from the Zig Code of Conduct is wild.

I really want to pull on that thread more, to see how far you’ll go to defend an indefensible position, but I won’t.

it is what it is. i didnt write it, i was surprised when i read it too.

> This document contains the rules that govern these spaces only:

> The ziglang organization on Codeberg

> #zig IRC channel on Libera.chat

> Zig project development Zulip chat

> Is a racist screed acceptable to you as long as the following paragraph references technical issues correctly?

I'm not the morality police. Nobody should be. I'd still take the article on its technical merits. As a random example, if Satoshi's paper called people using the banking system cattle, I'd still continue reading it.

> Imagine how you would feel if you were the person who wrote this code, and now you are being called a monkey in front of thousands of people on the internet

It would be absolutely fine, nobody is named specifically. He wasn't like Josh Examplemann working on Actions is a piece of shit that botches any feature he touches. Nobody is going to remember a blog post and forever hold anyone that worked on Actions to an unhirable status. And personally, I think it would be good for people to feel some shame for having implemented a feature in such a terrible way. It's not like they were told by their managers to commit these the way that they did. Calling into the sleep binary wouldn't even be more work.

Whoever is behind the new React Start Menu in Windows

along with whoever is responsible for the Chrome Web Environment Integrity

along with whoever is behind the design of OSX Tahoe

along with anyone who is working on Windows Copilot that screenshots your screen

should be ashamed of themselves. The more articles that do that, the better. They are not doing good.

I would appreciate a simpler conceptual explanation for someone not steeped in the Haskell / functional programming world :-)

I gave a talk on this topic at Zurihac this year, called A History of Effect systems: https://www.youtube.com/watch?v=RsTuy1jXQ6Y

It was given to a Haskell audience, but not everyone knew Haskell, so I hope it's generally accessible. It describes how Oleg's work fits into the overall history of Haskell effect systems.

Wow, this is so cool that you did this and this showed up right now. There was an online book (soon to be published) by Xavier Leroy on the history of control mechanisms in programming languages with the last few chapters on Effect systems.

Thanks! You might also like the introduction I wrote to the Bluefin docs, which has some similar content: https://hackage-content.haskell.org/package/bluefin/docs/Blu...

Yes, Xavier's book was really interesting. It's here: https://xavierleroy.org/control-structures/

It's missing some subtleties of capability-based effect systems (which Bluefin is). Maybe I should write to him.

Uhhh... the above comment has a bunch of loose assertions that are not quite true, but with a enough truthiness that makes them hard to refute. So I'll point to my other comment for a more nuanced comparison of Markov models with tiny LLMs: https://news.ycombinator.com/item?id=45996794

To add to this, the system offering text generation, i.e. the loop that builds the response one token at a time generated by a LLM (and at the same time feeds the LLM the text generated so far) is a Markov Model, where the transition matrix is replaced by the LLM, and the state space is the space of all texts.

The Markov property means that the next token is determined purely by the current token. Well, if it were a Hidden Markov Model, the next state would actually be determined by the current state, and the respective tokens would be a lossy representation of the states.

The problem with HMMs is that the sequence model (Markov transition matrix) accounts for much less context than even Tiny LLMs. One natural way to improve this is to allow the model to have more hidden states, representing more context -- called "clones" because these different hidden states would all be producing the same token while actually carrying different underlying contexts that might be relevant for future tokens. We are thus taking a non-Markov model (like a transformer) and re-framing its representation to be Markov. There have been sequence models with this idea aka Cloned HMMs (CHMMs) [1] or Clone-Structured Cognitive Graphs (CSCGs) [2]. The latter name is inspired by some related work in neuroscience, to which these were applied, which showed how these graphical models map nicely to "cognitive schemas" and are particularly effective in discovering interpretable models of spatial structure.

I did some unpublished work a couple of years ago (while at Google DeepMind) studying how CHMMs scale to simple ~GB sized language data sets like Tiny Stories [3]. As a subjective opinion, while they're not as good as small transformers, they do generate text that is surprisingly good compared with naive expectations of Markov models. The challenge is that learning algorithms that we typically use for HMMs (eg. Expectation Maximization) are somewhat hard to optimize & scale for contemporary AI hardware (GPU/TPU), and a transformer model trained by gradient descent with lots of compute works pretty well, and also scales well to larger datasets and model sizes.

I later switched to working on other things, but I still sometimes wonder whether it might be possible to cook up better learning algorithms attacking the problem of disambiguating contexts during the learning phase. The advantage with an explicit/structured graphical model like a CHMM is that it is very interpretable, and allows for extremely flexible queries at inference time -- unlike transformers (or other sequence models) which are trained as "policies" for generating token streams.

When I say that transformers don't allow flexible querying I'm glossing over in-context learning capabilities, since we still lack a clear/complete understanding and what kinds of pre-training and fine-tuning one needs to elicit them (which are frontier research questions at the moment, and it requires a more nuanced discussion than a quick HN comment).

It turns out, funnily, that these properties of CHMMs actually proved very useful [4] in understanding the conceptual underpinnings of in-context learning behavior using simple Markov sequence models instead of "high-powered" transformers. Some recent work from OpenAI [5] on sparse+interpretable transformer models seems to suggest that in-context learning in transformer LLMs might work analogously, by learning schema circuits. So the fact that we can learn similar schema circuits with CHMMs makes me believe that what we have is a learning challenge and it's not actually a fundamental representational incapacity (as is loosely claimed sometimes). In the spirit of full disclosure, I worked on [4]; if you want a rapid summary of all the ideas in this post, including a quick introduction to CHMMs, I would recommend the following video presentation / slides [6].

[1]: https://arxiv.org/abs/1905.00507

[2]: https://www.nature.com/articles/s41467-021-22559-5

[3]: https://arxiv.org/abs/2305.07759

[4]: https://arxiv.org/abs/2307.01201

[5]: https://openai.com/index/understanding-neural-networks-throu...

[6]: https://slideslive.com/39010747/schemalearning-and-rebinding...

Just a thought... would it make sense to maintain a govt/central registry of copyright owners, and have an "official" means of contacting them, on which they have an SLA to respond (say 3 months) which might be part of the ground rules for maintaining rights.

From a macro societal perspective, would this evolve "copyright" into a more balanced (value generating) deal for all of society?

I have no idea how accurate this comment from last week is, or if it applies beyond games, but the model is interesting:

> Japan has a scheme for orphaned games where if you can prove you did due diligence in searching for a rightsholder and couldn't find one, you can go ahead with rereleasing the game and the royalty payments get held in escrow by the government in case the rightsholder comes forward. I wish the US had something similar for cases like these.

Source:

https://news.ycombinator.com/item?id=45877983#45878084

At least with books, it's mostly individual authors who are most opposed to orphan works legislation. Disney isn't going to forget about whatever legal hoops are needed to maintain copyright. Individual authors (or their estates) may well do so.

hearing about Japan doing things intelligently gives me such a warm fuzzy feeling.

(previous example that I remember: https://www.japantimes.co.jp/business/2025/05/27/companies/j... )

Its not practical. Lots of things are copyright by default. HN comments are covered by copyright. Every photo you take is covered by copyright, so are letters contracts, kids drawings as well as professional artists,.....

What would work is an orphan works exemption, whereby if a work is not available and its not possible to trace the copyright holders you could use it.

The other problem is the term of copyright is far too long. it is ridiculous that something written during the reign of Queen Victoria could remain in copyright into the 21st century in the UK and EU. US law is slightly saner (in avoiding bringing out of copyright works back into copyright) but not much.

The whole basis of licensing law is everything is forbidden unless you have a written license permitting it from the copyright owner.

This is contrary to most (all?) other parts of the law where everything is allowed that isn't forbidden.

So it's the right of authors to ignore email requests to discuss a re-publication if they so wish.

It is, but I think what the comment you're replying to is saying is that the world might be better if things didn't work that way.

Given experience with OSS contributors burnouts we can see how world would not be better.

Oh how naive.

We have that in place for open source software. No one is contacting the authors on GitHub they just grab and use it.

Second thing is big bad guys will see if someone copyright is just a person that doesn’t have means to fight for themselves in court - you still have to sue them and still have to get initial cost of lawyers.

Last but not least there is a lot of content that you don’t want to be easily tied to the owner because history is showing us how that can be used to hunt down people having “wrong ideas”.