Forgot your password?
typodupeerror

Comment Re:really? (Score 1) 58

For those who don't know, Kevin Buzzard is one of the guys behind lean-lang.org

He's not who I'm referring to when I say a group of people don't know how to assess things.

That said, I don't think this will lead to that much in the way of noise... folks who are inspired to try to become some kind of math legend (with just a small amount of assistance from their friend the LLM) will find out pretty quickly how much more complex the environment is than the movies have led them to believe.

Comment Re: Wrong side of common sense (Score 1) 166

Ok but... how did that AI instruction get ingested by the AI...? Did jqwik's developers seek out innocent developers, entice them to use their code and then switch the terms to make their software incompatible with the existing ecosystem... and then take steps to alter that ecosystem to punish it for lack of compliance?

Comment Re: Wrong side of history (Score 1) 166

I'm not sure that's accurate, legally speaking. I think you're thinking like the judge is the prosecutor.

I also don't think you're taking into account the word 'protected' here. What does that mean in the context? Is this an 'official use' system? If so *what are you doing deploying untested code onto an official use system*?

Comment Re: Wrong side of common sense (Score 1) 166

Which laws apply here?

I'm genuinely curious if you can identify any. I'm not saying they aren't going to even... I'm just not aware of what would be considered by the courts here when the 'trojan' will, by definitiion, only be affecting people who are breaking the terms of use for integrating the (free, did we mention it's free?) library into their own codebase.

Comment novel marketing technique (Score 2) 92

Remember when wrestlers used to pretend to have real life beef to attract audience attention?

the staff of these two movies have been having what looks to my eyes to be completely manufactured twitter drama right now, but it's kind of leading to a pretty reasonable conversation about the routes people take into the industry.

my legit guess is that a lot of people who hadn't heard about Obsession did so because of this drama. I cynically expect marketers to try this on again regularly.

Comment I've seen some of Epyllion's analysis (Score 2) 21

It's oriented around the unstated belief that only the triple A games industry represents legitimate business, and tracks a bunch of metrics in a way that is so skewed it almost seems like they have an axe to grind.

Ball's approach is largely top down and only in touch with a series of market sanitized adages, so I'm not sure how he's going to deal with the problem of Microsoft largely being top down and only in touch with a series of market sanitized adages.

Comment Re:Cool (Score 1) 62

Like the fact that they're constructed via linguistic association, meaning that they effectively do all their calculations on the verbal level.
This means that if you ask them to do math, for example, you are relying on them taking the tokens that represent integers and using them with the tokens that represent arithmetic operators, without the system 'knowing' that these tokens are special in any way apart from their connections to other tokens.

But also it means there's no way to control if they just decide to start talking about goblins (or some other random concept) short of wasting valuable context space forcing them to avoid it as an explicit action... but oh... explicit defenses in a virtually infinite testing environment... there sure are a lot of things they haven't seen yet that they'll need to defend against.
They need a lot of testing, but the testing can never be 100% good enough for them to perform like a non-inferrential system.

And that's aside from the fact that a lot of folks, by default, use them like they're guaranteed to be right and don't even try to parse the output any more, because that's what marketing told them to expect. And by extension, they expect incremental quality improvements... because even they have to admit the systems just do random shit sometimes and they expect it to be possible to stop them doing that... and it might not be, because they are constructed via linguistic association.

Of course I don't know you haven't heard of this stuff... but if you're thinking that I'm only concerned about IP rights then maybe we aren't talking about the same things.

Comment Re:We have plenty of games (Score 1) 53

I daresay that already exists even... and it's friendslop.

A group of kid devs who wants to make something dumb to play with their friends are just as likely to create the next big thing as any AAA studio, right now. They can just fire up unity, unreal or godot and start placing assets and scripting, and maybe they'll find something that resonates with a lot of people looking for semi-collaborative experiences

most obvious example: https://store.steampowered.com...

Slashdot Top Deals

The cost of living hasn't affected its popularity.

Working...