drspod @ drspod @lemmy.ml

Posts

13

Comments

779

Joined

4 yr. ago

Well said.

Childish behavior all around, by the looks of it.

You can tell which engineers don't work face to face in an office with a team of other engineers, because they end up sounding like this in text.

Another corporate social media platform, what could go wrong?

Making unreasonable demands is a way of outing conscientious objectors and manufacturing grounds for dismissal so that they can be replaced with subservient true believers.

One of the American politics ones. And we call them communities here, not subs.

I think the way to make it work is to have each instance represent a "world" and you only have stats and equipment within a world. Then when you cross over to another instance you are now subject to that instance's ruleset.

It wouldn't be so much a federated MMO but more like a large variety of games connected geographically (in the virtual sense).

I'm trying to figure out what that looks like. So you mean they're just peer to peer? I'd be interested to see one of those, sounds cool.

Apparently they are still working on a paper; the only statements about the work by the researchers are via xitter.

There is a github repo for their model though: https://github.com/Jiayi-Pan/TinyZero

Thanks, I'm going to have nightmares tonight.

Macromedia Flash Action Script was the first language I saw that could have a RCE vulnerability caused by null pointer dereference.

Thank god HTML5 media killed Flash.

Since nobody has mentioned it yet: Have you never wondered how dumpster fires get started?

Once bitten, twice shy.

They want everyone to update to their new on-by-default AI shit.

Where is INTERCAL, the ur-esoteric language?

The Telegraph is right wing propaganda, not news.

I didn't know they were already doing that. Thanks for the link!

I thought that guardrails were implemented just through the initial prompt that would say something like "You are an AI assistant blah blah don't say any of these things..." but by the sounds of it, DeepSeek has the guardrails literally trained into the net?

This must be the result of the reinforcement learning that they do. I haven't read the paper yet, but I bet this extra reinforcement learning step was initially conceived to add these kind of censorship guardrails rather than making it "more inclined to use chain of thought" which is the way they've advertised it (at least in the articles I've read).

I'm not sure PixelFed is the best choice for private storage of photos. It's social media more than it is cloud storage.

Would Immich fit your requirements? I don't think it encrypts either, but you can self-host it.

Apparently they have barely produced any so they will all be sold out anyway.

This is tabloid trash, not news.