Exactly. Most of us can recall an era where script kiddies had similar unreasonable success even if they had no actual skill or thinking behind the attacks. Just canned toolkits designed around known vulnerabilities, common password databases, understanding of common, bad security practices, and bundled up in easy to use harnesses.

So whether it not an agent can actually "think" or is just putting together a potentially workable narrative of how hacks can happen and running it against real world targets doesn't matter too much. If you wired up the script of Terminator to the real world and it somehow launched nukes, it wasn't that skynet was thinking, it was that the systems folded easily to a movie script. Either way the distinction would hardly matter.

To the extent lack of thought may matter, compromising security is a use case that heavily favors AI systems that repeatedly "get it wrong", as it can have nearly limitless tries to get it right and even if it utterly fails, no one will really notice or write about it. It only takes a few successes, even if rare, to be pretty critical.

Well, one scenario is letting the hype get ahead of the capabilities, and do the 'AI store' experiment, but for real with responsibility for something actually important.

Another is that it is far easier to just screw things up than it is to make things work. I may not have it in me to make a house of cards, yet I can trivially even accidentally destroy a house of cards someone else built without a whiff of understanding.

The assessment of the capabilities based on bottom line certainly is subjective, which is precisely why Jensen Huang is so aggressively gung-ho about we must do everything we possibly can to spend more money on nvidia stuff. The motivation for Anthrophic seems either they have legitimate concerns or are looking to rationalize a plateau after promising the world. It certainly doesn't serve their bottom line to slow down otherwise.

Memorable Quotes

The Greeting: "Howdy-doodly-do, how's it going? I'm Talkie, Talkie Toaster, your chirpy breakfast companion. Talkie's the name, toasting's the game. Anyone like any toast?"

On Purpose: "I am a toaster. It is my raison d'être. I toast, therefore I am."

On Justice: Kryten: "His memory circuits were quite badly damaged in the accident."Talkie Toaster: "That wasn't an accident! It was first-degree toastercide!"

The Persistent Offer: Kryten: "Didn't you hear what I just said?"Talkie Toaster: "Yes, but I thought you might have changed your mind in the meantime."

Parting Words: "Would you like a cheese and ham Breville?"

So the F students become ICE agents?

You hit the nail on the head. Looking up license plates SHOULD requires a WARRANT, signed by a Judge. That would greatly decrease the abuse of the system. Go ahead, try to get a judge to sign off on you tracking your ex's whereabouts, I'll wait...

Isn't any query of the Flock database for a license plate a SEARCH, which as I undestand the Constitution should requires a WARRANT?

Training these models is incredibly expensive. In fact it's so expensive that there is no viable economic model to do so. As such, their options are either to drastically slow down the amount of new models trained or to go bankrupt. This is just a nice whitewashing of that. It remains to be seen if even using their models can be profitable, but we know that with the cost of training the next one they can't be.

Of the big players, I'd say OpenAI is at the biggest risk of not really able to even stumble straight ahead, presuming stumbling/charging straight ahead is in the cards for the companies in general.

Actually I'd probably not be too interested in an unspecified new tech, as with things *that* open ended I end up stuck in a demo of some sort of warehouse inventory management software. Thinking back to the people that got all excited for the 'IT' product and then the massive letdown when it turned out to be a scooter.

I would have to at least know a more specific area to figure out if I'm going to get excited or not. The mere rarity/exclusivity of an experience doesn't hold much appeal to me.

I'm not a film festival sort of guy, but in analogous cons that would hold my interest, I can't imagine an event being so vague and getting my attention for the one event. I feel like I'd not want to risk my limited convention time on such an unkwnown, unless it at least had some familiar facet that was known.

I could not imagine giving up three hours to watch something I have *zero* idea about, not even genre.

Show me when we *ever* built significant 'double decker' freeways? We have, obviously, been willing to build, rebuild and expand roads beyond what we had before. The problem is that we are seeing diminishing returns for our efforts. We give up more and more resources and land and get less and less benefit out of it. We expand 3 lanes to 4 and after tolerating *years* of even worse traffic to facilitate construction and then it opens up and... still jammed up because an additional lane didn't help *that* much. The problems aren't solved by more lanes, and more new roads means lots of people lose their property to serve the attempt.

There are other areas where I think you have a point. For example we somehow had the will to build out electricity and telephone to reach the middle of nowhere, but when you say we should build out fiber in the same way, it's just not affordable. But we have consistently been all about throwing more at roads, and have way more and better roads than we have *ever* had, but they just don't work as well with scale.

Actually, from what I've read broadly George Lucas' work was really only decent by virtue of everyone else kind of fixing it up around him. Which is a prime example of some critical mass of human involvement being useful to refine creative outcomes.

It's about the authors rather than GenAI.

Previous advances brought down costs and accelerated production, but things were still specifically deliberate choices by humans.

This stuff allows a mediocre short prompt generate significant chunks of content without human care.

The human care that serves to generally add quality now becomes an impediment to speed and volume of content.

You can still have that care applied with GenAI, but that gets drowned out by the low effort crap.

Content discovery is hard when drowning in crap content. Finding a needle in a haystack is rough.

Problem is that won't happen.

The fundamental issue is that a GenAI user that pumps out whatever the GenAI does without review will be tons faster than the GenAI user that actually cares, reviews, edits, and reworks the output.

This flips things upside down, someone with dedication under the traditional ways would produce more, now dedication and care become a liability. Accepting a bland paste of garbage from the AI is easier to produce content than trying to implement actual creative content.