The actuality is, anyone with pre-slop data still has their pre-slop data. And there are endless ways to get more value out of good data.
Bootstrapping better performance by using existing models to down select data for higher density/median quality, or leverage recognizable lower quality data to reinforce doing better. Models critiquing each other, so the baseline AI behavior increases, and in the process, they also create better training data. And a thousand more ways.
Managed intelligently, intelligence wants to compound.
The difference between human and AI idiocracy, is we don't delete our idiots. I am not suggesting we do that. But maybe we shouldn't elect them. Either way, that is one more very steep disadvantage for us.
It says something about how cognitively powerful simple symbols can become, that it took me a couple seconds to figure out why the record player was being represented by a camera icon.
Then I saw the wine glass over "lamp". The notepad over "toaster". And the phone over "watch". I feel a bit gaslit.
So as trade increases, and the need for dollars increases, gold and the dollar increase in value together.
That sounds benign until you realize that is: deflation. It incentivizes everyone to hold their dollars instead of spending them, and parasitically gain from gold's increasing utility use as currency by others.
This further decreases dollars available on the market to use in trade, further increasing their value. Which ... yes this can be an economic death spiral. Or at least, a strongly growth conflicted world.
A world in which you can make 5-10% a year by avoiding economic activity or investment is not a healthy world.
(And you can't increase gold supply to counteract this, because gold supply is also always balanced against the cost of extracting gold. Unless everyone else on the planet except the US government is prohibited from mining gold, there is no play there.)
Can you make the same arguments for holding land and activity that requires land? Or holding any other valuable, limited commodity vs activity that requires that commodity, copper for example?
Investors in those things haven't caused an economic death spiral. It's considered shrewd and good business to buy and hold things that are seeing increasing demand.
> This further decreases dollars available on the market to use in trade, further increasing their value.
... thereby reducing the amount of dollars required to trade a given amount of other things. Where does the spiral come from?
Holding land isn't as straightforward as you describe it. Every parcel of land is different. It takes a lot of knowledge of many kinds to be the one who makes money from real estate, instead of losing outright for overpaying, underselling, or simply being whittled down by taxes and other expense. And its lack of liquidity creates high situation risk, for anyone who doesn't have a large enough csh buffer to weather life's turns, unscathed by any immediate needs.
But yes, sophisticated investors do make money on under utilized land. Parasitically.
And it is an economy damper. Under utilized land is poor economic optimization by definition. Not to mention the undertow it creates, in terms of unproductive scarcity, for people who actually need to utilize land - whether that is for homes or businesses.
The solution, as economist Henry George pointed out, is to tax land, but not the property on it. Tax the exlcusionalry holding of a limited resource, don't tax productive value created by people.
Taxing land and not property isn't as simple as taxing both, but it isn't rocket science either. There are lots of standards and practices, because distinguishing between the two is important for many reasons.
The irony of taxing property, the stuff that makes land productive, is that it is a wealth tax that hurts rich and poor alike. A poor person, unemployed, who spends their time bereft of reliable income improving their own home will now be taxed on the value added - every year for the rest of their lives or until they sell or lose their home. That is a very perverse situation.
So yes, the taxation situation that subsidizes under utilized property, which is also parasitically rewarded by gains in value far exceeding the lower taxes they pay (for having no development to tax), is continually damaging the economy.
Fuck that. The whataboutism term is used to much to sideline an marginalize different view points. I really hate this use of it. We are too engaged in being right about everything.
> Nor is it an argument that companies can’t do better jobs within their own content moderation efforts. But I do think there’s a huge problem in that many people — including many politicians and journalists — seem to expect that these companies not only can, but should, strive for a level of content moderation that is simply impossible to reach.
The three problems I see are:
1. People who imagine content moderation prohibitions would be a utopia.
2. People who imagine content moderation should be perfect (of course by which I mean there own practical, acknowledged imperfect measure. Because even if everyone is pro-practicality, if they are pro-practicality in different ways, we still get an impossible demand.)
3. This major problem/disconnect I just don't ever see discussed:
(This would solve harms in a way that the false dichotomy of (1) and (2) do not.)
a) If a company is actively promoting some content over others, for any reason (a free speech exercise, that allows for many motives here), they should be held to a MUCH higher standard for their active choices, vs. neutral providers, with regard to harms.
b) If a company is selectively financially underwriting content creation, i.e paying for content by any metric (again, a free speech exercise, that allows for many motives), they should be held to be a MUCH higher standard, for their financed/rewarded content, vs. content it sources without financial incentive, with regard to harms.
Host harbor protections should be for content made available on a neutral content producer, consumer search/selection basis.
As soon as a company is injecting their own free speech choices (by preferentially selecting content for users, or paying for selected content), much higher responsibilities should be applied.
A neutral content site can still make money many ways. Advertising still works. Pay for content on an even basis, but providing only organic (user driven) discovery, etc. One such a neutral utility basis, safe harbor protection regarding content (assuming some reasonable means of responding to reports of harmful material), makes sense.
Safe harbors do not make sense for services who use their free speech freedoms to actively direct users to service preferred content, or actively financing service preferred content. Independent of preferred (i.e. the responsibility that is applied, should continue to be neutral itself. The nature of the companies free speech choices should not be the issue.)
Imposed selection, selective production => speech => responsibility.
Almost all the systematic harms by major content/social sites, can be traced to perverse incentives actively pursued by the site. This rule should apply: Active Choices => Responsibility for Choices. Vs. Neutrality => Responsible Safe Harbor.
This isn't a polemic against opinionated or hands-on content moderators. We need them. We need to allow them, so we have those rights to. It is a polemic against de-linking free speech utilization, from free speech responsibility. And especially against de-linking that ethical balance at scale.
So, confirmation? They are catching up quickly!
The actuality is, anyone with pre-slop data still has their pre-slop data. And there are endless ways to get more value out of good data.
Bootstrapping better performance by using existing models to down select data for higher density/median quality, or leverage recognizable lower quality data to reinforce doing better. Models critiquing each other, so the baseline AI behavior increases, and in the process, they also create better training data. And a thousand more ways.
Managed intelligently, intelligence wants to compound.
The difference between human and AI idiocracy, is we don't delete our idiots. I am not suggesting we do that. But maybe we shouldn't elect them. Either way, that is one more very steep disadvantage for us.
reply