Signal has a central, proprietary server. It's between impractical and impossible to run your own Signal server like you can with Matrix, Revolt, or Delta Chat for example. BlueSky has a similar approach (compare to Mastodon).
Also Signal requires a phone number to sign-in. It's not exactly private. AFAIK the proprietary server can glean your IP, your phone number, who you talk to, and when you talk to them. This type of metadata is valuable information.
The WhatsApp co-founder gave Signal $105M in 2018. Signal costs ~$50M/year to run. It's also funded by wealthy donors such as Jack Dorsey (Twitter, BlueSky, Square). BTW Jack is now pushing Signal to integrate Bitcoin.
When evaluating the "ethics" of a chat platform, we should factor-in the metadata, soft power, and eventual leverage that centralized (controlled by a few) platforms like BlueSky and Signal afford to wealthy folks who are bankrolling it.
Critical portions of the server are not FOSS. Also the core software forces you to join their servers.
Also we have no proof that they are running the server software published on GitHub. This concern is exacerbated by the fact they didn't publish server code updates for many months.
I mean sure, but also, the client app source code lets us know that unencrypted data is not sent to the server. So at best they could perhaps be collecting some additional metadata, but I don't think it's a whole lot
> AFAIK the proprietary server can glean your IP, your phone number, who you talk to, and when you talk to them. This type of metadata is valuable information.
To the best of my knowledge, so can matrix.org or whatever servers you connect and federate to. This is required to route messages between users. What is your point?
When a company is writing all of their technology and business using, lets say JavaScript. And then they come here and post about them switching to some NIH home-brewed language instead because they couldn't understand how functions or arrays work. That is not a problem with the people commenting on their stupidity.
Yeah but if everyone was saying "I don't understand how functions or arrays work in Javascript", that's a pretty solid indicator that functions and arrays are badly designed in Javascript and are unnecessarily hard to understand.
I think in some cases things are just fundamentally difficult and them being hard-to-understand is intrinsic. For example formal verification in Lean is hard to understand but I don't think Lean is badly designed.
But it's hard to see why package management is one of those things. There are soooo many ways Nix could be easier to use and understand. The language itself is unnecessarily esoteric in my experience - compared to something like Starlark for example.
Just the latest in the line of "my totalizing world view will solve all your software problems" to which the answer of "this doesn't do what I want" is always "you're holding it wrong."
There is an Appendix E, it just has no content besides the title. There's also a reference with only the text "More details on prompt p′ information can be found in Appendix". I'm thinking this isn't a final draft, maybe?
If not this exact paper, This kind of memetic attack likely exists out in the wild. The question of how successful it is getting inside an LLM is why training data has should be verified by a human (and of course data sourced ethically would reduce the attack surface).
Friendly reminder that China publishes twice as many AI papers as the US[1], and twice as many science and engineering papers as the US.
China leads the world in the most cited papers[2]. The US's share of the top 1% highly cited articles (HCA) has declined significantly since 2016 (1.91 to 1.66%), and the same has doubled in China since 2011 (0.66 to 1.28%)[3].
China also leads the world in the number of generative AI patents[4].
