Elektrine lite

← Feed

robryk

robryk@social.wuatek.is

(I moved here from <a href="https://qoto.org/@robryk.">https://qoto.org/@robryk.</a>)<br/><br/>I enjoy things around information theory (and data compression), complexity theory (and cryptography), read hard scifi, currently work on weird ML (we&#39;ll see how it goes), am somewhat literal minded and have approximate knowledge of random things. I like when statements have truth values, and when things can be described simply (which is not exactly the same as shortly) and yet have interesting properties.<br/><br/>I live in the largest city of Switzerland (and yet have cow and sheep pastures and a swimmable lake within a few hundred meters of my place :)). I speak Polish, English, German, and can understand simple Swiss German and French.<br/><br/>If in doubt, please err on the side of being

Posts

  • View post

    @MicroSFF@mastodon.art Isn&#39;t that rather a blessed blade though?

  • View post

    @soatok I&#39;m confused what are the consequences of first vulnerability. It seems to allow an attacker who can control the public key the victim will believe their counterparty has to cause the victim to establish session secrets ~controlled[1] by the attacker. If the attacker is in such a position, they are essentially able to impersonate the victim&#39;s counterparty anyway. How does being able to force session secrets to the fixed value give them more? (Or am I misunderstanding something m...

  • View post

    @emily @ifixcoinops Another approach that mitigates the issue is sharing infrastructure with friends, but not ones you live with. Then the potential for abuse is split, and potential for abuse of infrastructure you use by people you live with is even lower than for single-user instances, because of lack of physical access. Which abuse avenues via home automation do you think are most important? Ones where abuser uses it to spy, or ones where they command it to directly abuse and/or gaslight,...

  • View post

    @mjd Do you have a reason to expect whether replacing conversation with directly amending the language model&#39;s output (in a similar way to one that people use to guarantee that the output is e.g. valid json) and using that amended output for context would be better/worse?