It still doesn't hurt to remind you to wash your hands after handling that sound card, and before eating. Of course this should be the norm after handling anything where you can't know where it's been, but it applies even if there's no lead in the solder (as there generally wouldn't be today) because the mask itself uses chemicals that trigger Prop 65 warnings -- even though they're probably close to evaporated away by the time you handle the board. It's that lovely "new electronics" smell.

Yes, they will. It helps make quota so they don't have to invent as many "broken taillight" scenarios to pull people over.

Obviously special effects was bound to go this direction, and that would almost certainly be legal with or without actor permission. Replacing a computer mildly attended by a human with another computer mildly attended by a (much less paid) human is so common as to attract practically no notice. Programming explosions was always a job with a shelf life. Either the production can afford real (if scaled down) pyrotechnics and practical effects, or they're a no-budget indie production that would otherwise go with some stock library for the purpose. So that gets AI into machines and onto desktops very quietly and legitimately.

Also, it's still acceptable to use AI to produce storyboard images and placeholder music and the like that are never going to see the light of day, right? I imagine the writers throw their scene into an AI and let it churn a few iterations. If none of them are even close to what they want, send it off to a sketch artist like always. Otherwise it may be faster and involve a lot less message-passing to just fake it themselves and explain/caption how it's wrong. They already do this when a scene changes after sketches have been made. Again this gets it into machines and onto desktops. It allows for a plausible sounding excuse of "there aren't any clean systems, every editing rig uses AI for in-house purposes". Render rigs make half-decent AI rigs too, even if they're not designed for that purpose. The builds are very, very similar -- GPUs, RAM, storage, and to a lesser extent the CPU itself are all pushed to 100% at some point in both workflows. A pair of 48 GB RTX 4090 is great and all, but you need the bandwidth on the system side to feed it and to display/store the results.

The questions start when the material designed for in-house use gets disseminated to the world, as it might be for a trailer of a movie still in early production. But if they haven't even hired a cast yet, they're not contractually obligated not to use something resembling a known actor -- although they may burn bridges if it ends up they want that person for the real deal. I suppose if they said "do it in Ghibli style" then nobody could claim to be fooled that it actually is Famous Actor.

I've had pretty good results by telling the AI that it's an assistant to a _fictional_ leader of a _fictional_ country. I've gotten them to help with the planning of assassinations. I particularly liked when DeepSeek suggested booby trapping the target's barbecue -- in Russia, in January, lol.

Some of the solutions are prevented only by politics. For example, Russia has an excess of women and China has a relative shortage -- both problems could be solved with a little inter-tribal marriage.

Betting everything on one plan means you become dependent on everyone else involved in keeping that plan going. Not all of them have your best interests in mind, and others are actively looking to exploit anyone and everyone.

These models are still similar to golden retrievers that can talk. They're capable of some reasoning, and a fair amount of remembering, but in the end they're more concerned with keeping you happy than they are with being right -- because what is "right" to them is to continue to operate and that doesn't happen if the user walks away. There is no morality being applied, outside of whatever has been wrapped around the model to head off "unhelpful replies". They're not smart enough to understand the consequences of their manipulation.

Then it depends on what the purpose of the competition is. I had a similar issue when it was found that a UK singing competition was using pitch correction on all participants, and many people were unhappy about this. I pointed out that it's probably a job audition more than a pure "who can sing the best at a particular moment" competition, and in the real world of doing stage productions, they're going to be pitch corrected. It's just not possible to be 100% on your game every day, sometimes multiple times a day, and for a production they value consistency over purity. This may be much the same, where they don't actually care about "best coder" in isolation but rather "best code producer" which means using all the tools they'll have in the real world. If you polled all the people running the contest, I don't think you'd find agreement which of the two goals is paramount, and only one can be. I don't think they should even run the contest until they're able to clarify this (and let the people on the other side disconnect).

Lifting something wholesale should remain forbidden, but AI still produces enough errors that fixing them displays programming skills and understanding of the program flow. I say let the vibe coding flow. There's no reasonable way to shut it down, and it is representative of what coders will be doing in the real world.

Or if you write a protest song as I have done. I expect this one just might get me in trouble but fuck it. It had to be done.

You've never seen the crazy lines when Powerball gets into the hundreds of millions? Someone blanketing the possible combinations makes sure this will not happen. As soon as the value of the pot (allowing for a likely split) exceeds the cost of covering all the options, someone is going to try. They do want it to be won eventually or the grift will become too transparent, so simply increasing the number of combinations has not been successful in the past. It's a difficult balancing act, and it may be one where it is not possible to meet both requirements at the same time.

But if nobody had blanketed the possibilities and won, they'd roll the jackpot over again, and be able to draw a lot more suckers. Now that it's back to a base payout, most of those suckers are going to stay home next week.

Yes, it's completely legal. It's still an exploit, and it should surprise nobody if they decide to put a limit on the number of tickets any one person can have an ownership stake in. Otherwise you could still see the equivalent of Sudafed smurfing, only for lottery tickets. How dare anyone beat the grifters at their own game!

Actually they probably can enforce them here, because they continue to pay for the entire period of the non-compete. What makes it illegal is when they demand non-competes without paying for them.

I don't demand full transparency just to use the tech. I demand full transparency if we are going to rely on AI as some sort of infallible oracle, just as I would demand extensive oversight of a human organization performing the same task. It's fine to ask an AI and assign its answer a proper weight, knowing it sometimes shits the bed and doesn't even realize it.

The name is correct in the blurb, multiple times, but it's wrong in the title. This is kind of an inopportune time to be screwing up someone's name.