

Interesting - I can sort of intuit why it might help. Feeding the model bad data and instructing training it to identify it as such would be advantageous compared to being entirely unaware of it.
Interesting - I can sort of intuit why it might help. Feeding the model bad data and instructing training it to identify it as such would be advantageous compared to being entirely unaware of it.
How do you mean? Most authoritarian regimes are maintained with fear and violence. Even in democratic states that don’t primarily rule through fear the government generally has a monopoly on using violence to enforce the law.
I have.
It includes “compliance with EU regulations” which in this case is soon going to involve redirecting and tracking visitors to sites such as thepiratebay.
Fully expect this to be a move to enable them to enforce this via blocking DNS providers that don’t comply with censorship lists, instead directing people to use this.
I don’t need an EU DNS, I already have OpenDNS.
Having seen a couple of vids from folks on the ground it seemed akin to a proper French riot. Folks lobbing bricks, makeshift barricades, vehicles burning etc.
Yeah folks may feel justified an upset but it isn’t exactly surprising when authorities crack down to restore order in such a situation.
The EU is already trying to block and censor ips via DNS, so I don’t trust this initiative at all, nor, frankly, do I trust the EU as an organization. It should stick to foreign policy, not trying to overrule our national governments to force legislation onto us that we don’t want.
This is obviously subjective depending on what you want to achieve with your llm, but “Bad” data in that it showcases the opposite of what is desirable output. Think bunk conspiracies, hostility, deception, racism, religious extremism etc.