vzaliva

I really hate such clickbait titles. When I read it, the assumption was that the Fable model would hide some information in watermarked images by default. This would be very concerning from a privacy point of view.

What the article is actually about is how the authors instructed Fable to do so.

I bet the same is happening with writing, at least with technical and academic. Automated grammar checkers and AI correctors bland the language by enforcing strict grammar. Whereas a slight departure from standard grammar and punctuation is something that could give writing character and personality.

It reaffirms that Ukrainians are:

1. very advanced with drones 2. caring about their soldiers

Disclaimer: I am a proud Ukrainian.

it was very slow, and I had to re-authenticate often. I want something like dropbox daemon, which mounts drive and syncs it. Set up once and forget. Since I have not found anything like this for Proton I am still using Dropbox for some synced data.

In fact I did. I multilingual and ocassinally query ChatGPT and Claude in other languges than English.

"Languages: German and English" – this is odd. That means their dataset is limited. In my understanding, frontier models are trained on multilingual datasets and can combine knowledge no matter what language it was written in.

In my experience, ArXiv has become too slow to publish. My previous paper took weeks to be approved. It was a recommended path for journal submission (posting to a public repository and submitting a link), and after waiting a long time, I just posted it on hal.science instead, and it was up in a day.

Another one of mine has been on hold for 18 days and counting.

I think it is a great service to the community, but clearly they are having a scalability problem.

I expect in the near future all digital cameras to digitally sign the images they take. Even before AI slop, it was useful to avoid manual alterations. AI makes it all too easy, so it makes sense.

However, this will certify only the original image. I think the missing part of this is additional layers of certification which allow some image editing (e.g., rotating, contrast, etc.) yet clearly document that the image was modified and link to the original image ID. Kind of like a signed git log.

I was thinking about making a "personal data search engine". I have almost 30GB of email archives and a bunch of Google Drive and offline files. I am thinking about downloading it all from the cloud and storing it on a large RAID array, with an indexing and search interface (and possibly a local MCP server).

Would Hister be suitable for this? Can it index mbox files? Would it handle this amount of data? Does it have a search API so I can build an MCP server?

I’ve finally cancelled my Cursor subscription. It was an interesting product, but the fact that most interesting features only work with per-token pricing, not Cursor and Codex subscriptions, is a bummer. As to their Composer models, they were good for simpler coding tasks, but after the company purchase, they will probably end up merged with Grok, and I do not want to pay nor contribute to Elon Musk’s businesses.

Imagine all licence plates being small screens and periodically showing numbers derived from a private key issued by the DMV. As they change, vehicles could not be tracked without knowing the private key, yet the DMV can verify the ownership because they know one. I am not a crypto expert, and the actual scheme may be a bit more sophisticated, but you’ve got the idea.

You nailed it: there is a conflict between surface area (wanting more for solar charging) and efficiency (you want it to be aerodynamic and light, hence less area for solar charging).

I also think an extendable canopy (for a car, not a bike) should be a better approach. I wonder if anyone tried it.

I do not know any EU-only, but ProtonMail is in Switzerland.

The intention is good. I like the reviews idea. Since the author has hit the limitations of ATProto, could ActivityPub be an answer?