My impression is that BandCamp is not inherently against using new technology to generate new sounds. This decision seems to be not about "what is art" but about "what is good for BandCamp right now".
As a small platform like BandCamp you do not want to be flooded by cheaply generated AI copies of existing songs / genres - this would alienate both artists and customers and could endanger the whole platform. You also don't want to expose yourself to internal complaints and external copyright claims because someone uploaded hundreds of "Popular work X in the style of popular band Y" songs.
The AI slop avalanche will pass by and probably leave behind some cool stuff. In the meantime, it seems like a sensible option for BandCamp to step aside and evaluate their position in a year or two.
Hmm, a few years ago, such a feature would have been called "AI" in marketing materials, no? Aren't all current GenAI tools in some way "just some base patterns selected by sliders attached to options adjust stuff" - only the 'base patterns' are some weights in a neural network.
I have volunteered in various roles for ACM conferences and thus have some insight into ACM's path towards Open Access over the past years.
Just a few things to consider:
- ACM is not a for-profit publisher like Springer or Elsevier. Any profits made from their/our publishing activities subsidize e.g., outreach activities, travel stipends for developing countries, and potential losses from e.g. conferences.
- In my experience, ACM is one of the very few publishers in computer science where you can generally trust the published papers.
- Keeping a long-term digital library is not just "putting PDFs on a server" but involves a lot of additional costs. The ACM HQ is rather lean IMHO, but there are multiple people involved in developing the Digital Library, handling cases of copyright infringement and plagiarism, supporting volunteers, etc. Also, the ACM DL contains a rising number of video recordings of conference talks, etc. Additionally, there are several contractors to be paid. For example, authors no longer generate their own PDFs but submit the LaTeX/Word manuscripts to a central service (TAPS), developed and operated for ACM by an Indian company, Aptara.
- In the past, subscriptions to the ACM Digital Library were a major, stable source of income for ACM. ACM has to be careful to not get into financial trouble by giving away their crown jewels without generating sufficiently stable alternative income sources.
I don’t see what the benefit of most of that is, and why publication fees are a good way to pay for it. Take video recordings: why should we pay for them to develop a video hosting platform when perfectly good ones exist? In fact, conferences that I am familiar with put the recordings on YouTube, and this works great.
> ACM has to be careful to not get into financial trouble by giving away their crown jewels without generating sufficiently stable alternative income sources.
The attitude that the work of science belongs to these publishers is what grates me the most. Yes, ACM is not as bad as Elsevier, but this attitude is still fundamentally wrong. They are in the position they are mostly by historical accident, able to extract rents because it requires a lot of coordination to switch.
Why do I call it rent extraction despite the ACM doing stuff? Suppose the ACM charged separately for using their video platform. Would anyone pay for that?
> For example, authors no longer generate their own PDFs but submit the LaTeX/Word manuscripts to a central service (TAPS), developed and operated for ACM by an Indian company, Aptara
And what good is that? Why should we pay for a separate company to run pdflatex for us? The system exists primarily to check that we’ve put ACM branding in the paper.
Sometimes people also say that the real service is long term storage of pdfs, but let me preempt that right now: there are government sponsored long term storage facilities like Zenodo that are likely to outlast ACM. Second, commercial storage paid for indefinitely using an annuity would cost less than $1 in present value for hosting the pdfs of a conference, about 0.0001% of ACM publication fees.
It doesn't. arXiv is exclusively a pre-print service. The ACM digital library is for peer-reviewed, published papers. All of the peer-review happens through the ACM, as well as the physical conferences where people present and publish their papers.
Yes, and that peer review happens through the ACM. It serves an organizing function. The conferences themselves are also in-person events, and most of the important research papers come out of those conferences.
Small addendum: at least in Germany, early telephone wires (up until the 1950s?) were wrapped in paper, drenched in oil. The bundles were then enclosed in a lead-copper alloy to protect them from moisture.
Yes it is! And yes it was a typo! Thank you for finding the link to it. I was trying to find it before but couldn’t find it in official sources, DB deleted it from the official app. It did not stop at Troisdorf though.
It is operated by National Express, but I guess the routing comes from some other company (likely InfraGo)
Claude Code user¹ says Claude Code wrote continuously incorrect code for the last hour.
I asked it to write Python code to retrieve a list of Kanbord boards using the official API. I gave it a link to the API docs. First, it wrote a wrong JSONRPC call. Then it invented a Python API call that does not exist. In a new try, I I mentioned that there is an official Python package that it could use (which is prominently described in the API docs). Claude proceeded to search the web and then used the wrong API call. Only after prompting it again, it used the correct API call - but still used an inelegant approach.
I still find some value in using Claude Code but I'm much happier writing code myself and rather teach kids and colleagues how to do stuff correctly than a machine.
FWIW, many of the researchers on the paper did not study in the U.S. but immigrated after their PhD studies.
I checked the first, middle, and last author: Lars Mescheder got his PhD in Germany, Bruno Lecouat got his PhD in France, Vladlen Koltun got his PhD in Israel.
(Edit: or maybe they did not actually immigrate but work remote and/or in Europe)
As a small platform like BandCamp you do not want to be flooded by cheaply generated AI copies of existing songs / genres - this would alienate both artists and customers and could endanger the whole platform. You also don't want to expose yourself to internal complaints and external copyright claims because someone uploaded hundreds of "Popular work X in the style of popular band Y" songs.
The AI slop avalanche will pass by and probably leave behind some cool stuff. In the meantime, it seems like a sensible option for BandCamp to step aside and evaluate their position in a year or two.
reply