Eddy_Viscosity2

Carpenters can and do make more money when they make things by hand. But the customers base for this is indeed smaller.

It's like complaining about a chef who make all their food from fresh ingredients instead of buying off-the-shelf processed versions. People will pay more for that freshy made food.

I guess there are experimenting with they can enshittify. But even if they undo this current change, all commercials model will eventually be enshittified.

The problem is scale. Human-produced garbage content come in at rate X, but LLM produced garbage can come in orders of magnitude faster because you can spawn off agents to produce garbage content multiplying the output. The only thing limiting this is the economics and since the cost to produce this garbage can be so low, the amount that can be produced while still be economically feasible is massive.

It's the difference between a regular ocean wave and a tsunami. It's the difference between a person bumping into you and a being hitting by a speeding semi-truck. Scale is the problem with LLM garbage content.

No way to enforce it. We already have laws about scams and counterfeit products which Amazon does not enforce, even in cases where it absolutely could.

> these are just different jobs

This is a point I came to make. Different jobs also different skillsets. A good solution taken on by some industries is decoupling promotion and salary rises by having separate engineering and management streams where both have similar opportunities for career progression but you can stick to the actually job that you yourself are best suited for. People who have the people/social/political/organization/high-level-view skills and like exercising them can go into management, and those with focus/deep technical expertise/creative problem solving skills, can stay in engineering. People do usually have bits of both of these, but you know if you are stronger in one or the other.

People do not believe the promises being made by these companies. For good reason.

> You can implement policies that tax more than just than just income.

Umm, no you can't. Otherwise We would have done this already. In a world where the rich would be even richer, this would be even more impossible.

SCOTUS will rule that because the framers didn't specifically and explicitly say that local governments could cancel flock contracts, that therefore they cannot and must continue to both pay them forever and give them access to any location for cameras that they want.

> I'm against having it used everywhere for everything.

The decision-makers of the world have already decided that AI should, and will, be used everywhere for everything. Just like they decided that everyone will be surveilled, and everything will be permeated with ads.

The 0.1 are the ones with super yachts and private aircraft and multiple homes and do space flights for fun. So 100x worse sounds pretty reasonable to me.

Then that's an easy workaround, you put a third party between you and direct ownership and you change the statistics without have any meaningful change to who really owns what.

> College should be for becoming educated. Not for getting a job.

That ship has sailed. College is a now a job gate-keeper and are extorting young people for certificates that employers 'require' to hire them.

I'm all for education, and there are indeed jobs for which degrees should be required (engineering, nursing, etc), but many many jobs just require some generic degree in anything. It's a bullshit requirement and the costs are massive and rising and saddle young people with debt some can never get away from.

Also, your cheese is made of congealed ruminant squeezings.

> take my business to a vendor I can better rely on

This competition phase of AI will be short-lived. It will soon monopolize (or have the equivalent effect of a few big vendors having unwritten and unspoken collusion agreements that have them all behave in lockstep). This is where the profit is, and also the place where they don't care what you want.

Yeah yeah, open weight opens models. I wouldn't count on it. People in the billions still use microsoft despite how hostile they are the customer. Some will find an alternative but most won't.

GAAP are pretty clear that those things don't count, especially the last one.

The journals were captured by the finance-bro mindset long ago. They realized that they are the gate-keepers to science communities career progression/promotions/grants etc. So what does a sociopath do when they own a gate? Charge fees, exploit and extract.

If LMMs fully flood the journal space with crap, then maybe journal publications can, finally, be de-coupled as a metric from performance assessments. The gate and all its perverse incentives then vanish. If anything LLMs might be the very thing that breaks the system and at least allow the possibility of something better to emerge from the ashes.

Why wouldn't it be legal? Reminds me of those 'non-profit' charter schools where the 'school' entity pays huge rent and/or other admin type costs to a very much for-profit entity controlled by the same people. So they set the price at what they want and extract what they can.

the transition from waste gas to wasted gas.

There's the complication that once a set of people have the power to inflict violence on others, they will be abuses. So how many innocent people is ok to cane? Checks and balances could be put in place to lower the innocent caning rate, but these will be expensive and massively opposed by those doing the abuses which leads to them be performative and ineffective. We see this with policing generally. There are also the exempt classes which has enough status to never have to worry about being caned. So then only lower status people are getting caned, and likely with the highest innocent rate as they have the least power to defend their rights. It's a path to dystopia.

Clearly from context they meant every android device not every device.

If it's a complicated but narrow problem then regardless of the prompt its possible that Claude will always converge to the same solution type from its training data. So it will generate different version of the same code that have the different implementations of the same procedure that 'works' but is incorrect in some subtle way.

I likewise have about 4 good hours of thinking work capacity in a day. But even finding that can difficult because you have to navigate all the trivial non-thinking admin tasks, meetings, forms, approvals, and other soul-sucking semi-work which does not progress a project, but yet a project can exist without. Management will always find a way to fill up every last minute with this category of non-thinking semi-work so its a fight to even find a little window to do the hard thinking work.

So far as I've seen, the laws don't apply to AI companies. They do what they want.

Indeed, the question not whether its illegal, but who has more political power to maintain/change/ignore the law. Plumbers/electricians or companies that want to 'disrupt' things with low-wage AI-assisted labor. History tells us that any solution with lower costs will win that battle no matter how shitty they are.

The next generation will grow up where AI was always a thing. I do not think it the existence of that easy button is going to lead to cognitive capability improving (at population levels, individuals will of course vary). That more we can offload to machines, the more we will.

The problem is might see the solution is to not have such laws in the first place if they are not enforced. But another solution is to give the government more and more and more enforcement power which will, eventually, be used against the people instead of for them. Just as bad is all these enforcement people have to continually justify their pay so the bar for what constitutes an infraction starts dropping in order to keep the prosecution rates up.

They send a message asking to see you when you have a moment. In office settings I ask people things when I see them getting coffee and am likewise asked things myself.

I wasn't making a value judgement of it being good or bad. More of a 'well of course' kind of response.

This is just our energy conservation system in practice. It happens every time we experience and easier/better way of doing something. Another version of this headline could have been "Experiencing VIP treatment for just 10 minutes erodes your ability to tolerate the normal shit you have to put up with"

It turns out, many human beings are super ok about even human slavery. So I don't think its going to move the needle much with the likes of AI overlords like Musk or Altman. They aren't into ethics much.