Unfortunately, the safety net necessary to support oneself while they do the hardest thing isn't guaranteed.

We need to make engineers who work in surveillance or advertising ashamed enough to avoid putting that work on their resume.

I think that's a pretty big difference.

I began pursuing a biology degree on the side maybe 3 years ago so I can do that kind of work. Several of my professors are involved with projects that have recently lost funding due to NIH cuts and can't retain their engineering support. It hasn't been encouraging.

The more dollars there are, the more deeply in debt we are. If these were interpersonal debts where we all owe the dollars to each other such that they go away when whatever promise is eventually kept, that would be a tight knit society. But instead we're all indebted to the banks, so instead we have a lot of collateral at risk, and a lot of uncertainty about whether it's a stable arrangement.

If there isn't enough money to satisfy the asking prices set by the owners of these abstractions, then we can always go deeper into debt until there is. Or we could have a debt jubilee and let the prices re-settle to something more in tune with reality.

It could be that whatever lackluster expertise you can squeeze out of an LLM is good enough to discourage investment in the real thing since unlike NP-complete problems, expertise isn't generalizable.

At the end of the day what matters is which team is better, not which model. If Anthropic continues to feel like the good guy, relatively speaking, then people are gonna chose to spend more time getting to know its products and less time with OpenAPI's and on average Anthropic's will be the more capable teams.

I think vibes are gonna matter more and more going forward. The potential for bad behavior on the part of an AI company is severe. We're gonna have to tolerate whoever we enable in this space, so I propose that we make their marketing teams work as hard as possible to show us which will supply better vibes.

These are things you can infringe upon, but they all have dynamics that depart pretty wildly from the laws governing property.

I think we can agree that data is at least not on the uncontroversial end of that spectrum.

I guess I just don't see a meaningful difference between:

"____ cannot be property"

And

"At some other place or time ____ might be property but as a participant in the consensus for this place and time I am proposing that we not allow ____ to be property"

Its like rights. They only exist if you fight for them. Controversial notions of property are only legitimate if we let them be... so let's interfere with that legitimacy (and if we must, enforcement).

There are laws about what happens to me if I break into your house and steal your property. I can therefore find you case precedent indicating that a TV is property because people have been charged with violating those laws when they steal a TV.

But I can't present to you the absence of such a thing. We have trademark, copyright, and patent law, but as far as I'm aware there's no crosstalk with things that talk about property, things like armed robbery.

I wouldn't classify debt as an uncontroversial kind of property. In medieval Europe, Christians were prohibited from owning debt by their religions (Jews weren't, so they ended up being the lenders, which is probably why the stereotypes exist today).

I'd argue that the fungibility/resale of debt is a bad idea because it takes on weird properties when too much of it accumulates in one place.

But regarding the particular implementation as codified in US law (and I think elsewhere also), property rights do not extend to data.

One of them refers to tangible things, was first codified more than 5000 years ago, and is almost entirely uncontroversial.

The other was popular in 1700's France re: their system of privileges, and the people found it so onerous that they embarked on a campaign of executing nobility until it seemed like the concept was good and dead.

We can use the word however we like, it's just a word, but if we conduct ourselves as if they're the same sort of thing, which France was doing at that time, we're in for the same sort of pain.

So what I'm saying is that its a bad idea for us to let data be property.