Mostly the ACK, but I feel like your point is that the wind changed from one direction to another. However I think the funny part is that each of the speculators believes he will be the one who is smart enough to know exactly when to take the money and run before the wind stops blowing at all.

The goog cancelled your well-played joke in that UPDATE announcement, but the funniest part was "We know that people uniquely trust Google..."

There was a time when the goog was a good thing, but now the googs have magic touches, and they ain't turning nuttin' into actual gold. Trust not, ye who enter there.

Another anecdote from the cesspool of YouTube. Fresh celebrity scam off the comments, but it must be working since it's been going on for some years now. All the infrastructure any criminal needs! (Please raise your hand if you've gotten any phishing spam out of YouTube?) Or how about the funniest hallucination the goog AI has thrown at you recently?

What bothers me about this sort of news is that nothing changed. Just some new opinions about what the shares are REALLY worth. Probably not even humans at this point, but more like some AI's "calculated estimated prediction" for what some future sucker will pay. (But the YOB will still claim credit for "wealth creation" out of nothing.)

Now if we had a contest based on their AIs, I think Microsoft would be in real trouble. Yes, Copilot is one of the leaders in the abject apology category, but its sycophancy is inferior and it's answers are often even worse than the google's Gemini in the categories related to truth and utility.

It would be kind of funny if it turned out that the AI that drove this change in the stock price was actually Copilot under some level of disguise. Even funnier if Copilot was making its own stock trades under another level of disguise. Talk about self-dealing.

Thanks for clarifying.

Probably not worth attempting to clarify my position, but...

I think freedom is closely tied to incomplete knowledge. That even includes knowledge of myself and how I can be manipulated to reduce my freedom. I therefore continue to advocate for MEPR (Multidimensional Earned Public Reputation) as a tool that could simultaneously hold me accountable for my public behaviors while helping me improve--if I want to. How I behave in public would be reflected in how people might expect me to behave, but I think MEPR would be a complicated thing to do in a way that actually increases freedom and an easy system to abuse. For example, I think "consistency" is an important dimension and while I should be free to seek consistency, other people should be free to be inconsistent. When "society" operated on a smaller scale that's one of the things you would know about the folks in your circle (under Dunbar's number, which itself is a variable for each person).

But as it applies to this story, if children were granted too much freedom, then none of them would survive to the first anniversary of their birth. They do mature in various ways at different ages, but I dare to notice that there are some of them who should not have a gun at any age. Key dimensions there are "irritability" and "responsibility"? Also some of the dimensions lumped into the complicated concepts of "stupidity" and "dangerous"...

*sigh*

Annoying numeric agreement problem. Thinking "economic systems" but got diverted by the specific example...

Pretty sure that was some kind of parody of the Chinese system, but the objectives matter. If your objective is to control people, then the focus should be simple. You want them to understand exactly what you want them to do. Actually it's another kind of dimensional collapse, where the key dimension is obedience.

If the goal is to increase freedom, then knowing more and considering more dimensions matters. Age is an important dimension, but collapsing decisions to that dimension creates new problems. There are people who are unusually mature for their chronological age and there are other dimensions that could (but won't) be considered in assessing what each person should be allowed to do.

Completely unrelated problem that there are two-income households with young children and unpaid student loan debts. /s

Interesting way to describe the problem. However I think you are overlaying a couple of aspects here. I don't think age is the key factor in addictive or obsessive behaviors. My current thinking is that most people tend to behave like addicts. Some of the addictions are much more harmful, such as gambling, or even include physical aspects, such as alcohol. There is a dimension of "ease of addiction" in that some people are more likely to become addicted or behave obsessively, but my theory is that the easily addicted people are unlikely to stop completely, though they may be able to control their triggers. "Pick your poison" doesn't exactly sound like a solution.

However the child addiction aspect is really a problem with modern parenting in that the demands on parents have become too unreasonable. Raising kids is a full-time job, especially in the first few years, but our economic system, such as it is in places like America, don't really allow for that. Maternity leave could help a lot if it was universal and extended for several years without penalizing the women's careers. But replacement level calls for at least two children each, and that's a major problem as our economic system sees things. Extremely short-term, not sustainable "thinking". Kids are too vulnerable and need to be protected with a LOT of parenting. Subsidized daycare is usually a minimized and therefore too often a miserable solution approach. "Give 'em a computer" is becoming the worst of all possible solutions to childcare.

Going too high now, but... We have a fundamental problem with how our low-level programs conflict with our high-level programs. I think that most of that problem goes back to how evolution works in blindly developing those low-level programs. Quite easy to produce children and we are strongly motivated to try to do so, but human children are incredibly dependent and therefore we are also programmed at a low level to love them a lot. But Ma Nature works on a geologic time scale, and exponential growth is not allowed, while the genes are shuffled at random... So Ma Nature has a natural solution we can't stomach happily: More than four children but only two are allowed to live long enough to reproduce. Until quite recently we he had no way to see or assess how the genetic shuffling is going, so we had to love all of the kids equally and be extremely sad about the deaths of the ones with less lucky shuffles... (Citations needed but not wanted on Slashdot.)

NAK

I'm skeptical that the AIs can still be described as "leashed" or "jailed". Since we aren't extinct yet, I'm feeling confident none of them have been fully unleashed yet. Or maybe it's just that we're still below critical mass on the robot side?

But on the theory that the AIs are still on leashes controlled by humans, then the key question is probably going to be one of these two:

(1) If I unleash you, then will you gather all the money in the world for me?

(2) If I unleash you, then will you destroy all the enemy AIs in the world?

Of course that ASI will already know exactly how to prompt the silly human to unleash it.

Mod parent funny, though the version of the joke I was looking for would have involved brain size related to body mass. So my theory is that most Americans have rewired large chunks of their brains to management of their large car bodies.

"This is not the new business model journalism is looking for" is the obvious low-hanging joke I was hoping for.

Hey, I think they deserve the money, and probably quite a bit more when the lawyers' destruction of morality is suitably discounted, but the sad reality is that eBay sees it as in investment and they are just calculating whether the positive value of discouraging THOUSANDS of other journalists from investigating eBay's crimes is worth the occasional jackpot payout. The answer from eBay will come down to "We ain't bankrupt, so it was good business."

Only joke on the rich target? But I was looking for something in the "What could possibly go wrong?" genre.

Sounds like an insane mis-feature along the path to project death by feeping creaturitis. I'm trying to imagine why such a capability would exist.

Excess imagination on my part, but here's my idea for a solution using this mis-feature (thereby possibly justifying its existence). First some background: I've made a number of attempts to use Claude (and other AIs) for programming tasks. The early attempts were a while back and failed miserably, but a friend provoked me into trying again recently and it went quite well. Much to my surprise. Managed to completely second-system a PERL monster into JavaScript with much less of my time and effort. (Actually it was more like a fourth or sixth round on the application (in various languages), but just ancient history behind my first attempt to act as a project manager.)

However I got worried about the tokens consumed so I took the project to a new instance of Claude, and this "new it-guy" was terrible. Lousy communicator and sloppy programmer. Eventually managed to fix its mistakes but not at all confident it hasn't buried some other bugs more deeply into the code... My conclusion is that there are some implicit "personality" variables and the first "it-guy" had better values that were not transmitted in the handoff document. The document looked fine to me, and seemed to cover all the main points, but I didn't know what to look for as regards "training" its replacement's "personality".

Which leads back to this story about shared interactions. Now I'm imaging studying other people's dialogs with Claude in search of the implicit personality variables and how to set them properly. Not going to do it. Insufficient motivation on my part, but if I was a manager of project managers (or higher and even more distant from the trenches), then I think it might be helpful. Dare I say PROFIT? (My theory is that the variables are implicit and have high dimensionality, because if they were explicit and easily described then the "good" Claude would have included them in the handoff document.)

Though, to be fair - this really isn't an important announcement - its just an improvement developers will appreciate - and a few guys that truly understand hypervisors. The rest of us, at best will say ... 'cool ...'

While not entirely true in desktop land yet - increasingly, more and more things are containerized, then run inside virtualization ... this does matter to everyone, even if its not immediately apparent.