slibhb

I didn't say it was a problem, I said it "mattered"

If I wanted to argue that it's a problem, I'd just say that companies investing billions in training frontier models should reap the rewards. And distillation is essentially theft.

Of course it matters. Regardless of whether distillation is legal, there is a difference between training a model with and without distillation. For one thing, the distilled model wouldn't exist without the model it distilled.

Also, companies that use distillation may be competitive but seem unlikely to surpass the companies that are training these models from scratch.

Yeah, only publish stories about coral bleaching, animals going extinct, and forest fires. Censor any good news at all costs!

A member of the faculty (who I won’t name) said to me that the fact that the counterexample was so easy to find just indicated that humans had not spent enough time thinking about the problem, implying that a 60-year-old question of Grothendieck was not actually that interesting to work on. I didn’t tell him that at some point earlier in my career I had spent a week working hard on the problem. In my mind my colleague is just going through the five stages of grief; right now they seem to be in the denial phase.

It seems to me also that the very vocal anti-LLM crowd are in the denial phase of grief.

No, they're very clearly trained to be truthful.

https://claude.ai/share/3c19d0f3-d6e4-47a7-afdd-9a090300c100

So it points out that, according to experts, governments aren't always more efficient but then lists cases when they may be. Seems pretty balanced to me!

Don't know what else you would want. If it neglects to challenge the premise, it's just exhibiting sycophancy.

These models are trained to be truthful. Your disagreement isn't with the model, or the US, or the capitalist world but with economics and social sciences.

If you want Claude to list arguments for socialism, be explicit about that ("List the best arguments in favor socialism). It will gladly comply. You didn't do that, you asked it to assume a premise that runs contrary to the current state of expert knowledge.

but also don't give them AI for things they don't know, perhaps

The study doesn't show that at all. It didn't test actual AI.

They could have tested a cohort of subjects with access to actual ChatGPT. Ask yourself why they didn't.

While it's definitely true that NIMBYism isn't the problem everywhere, I live in Virginia.

I realize I'm not going to convince you but I just want to say...it's good that companies/individuals make a profit from building housing. Building housing is a pro-social activity that should be rewarded, like farming and healthcare.

It doesn't have to be 20 story apartment buildings. Start with duplexes, fourplexes, or townhomes -- none of which have to be ugly. It simply makes no sense to zone these densely populated urban areas for detached, single-family houses.

Yes, this is a real problem. Also, in high-demand areas, building more will lower prices compared to the counterfactual but that doesn't mean prices will actually go down. This makes it a harder sell.

There's no question that underbuilding is the largest factor here. You can look at construction vs. population over time (it's well below historical standards). Or available home vacancy rates. Or notice that the number of households is increasing faster than the number of homes. Or look at the various models, which show a 2 million-4 million shortfall in homes.

That aside, housing prices are reasonable in much of the country. Where population is stagnant or declining, they are generally reasonable.

Downward mobility is almost entirely caused by housing costs, which are a self-inflicted problem. We don't, as a matter of intentional policy, build enough housing.

I live in an expensive area and I'm always shocked by how many of my friends who cannot afford to buy a home (or even a condo) are NIMBYs. These are people in their 20s through 40s who have been earning since college, can't afford to buy, yet get annoyed whenever there's new construction that "alters the character of the neighborhood". Talk about false consciousness. I can at least understand people who own a home feeling this way.

So if say an foss ML project described what they do as "open AI" the company known as OpenAI would have a right to defend the mark. This is saying they could not.

Well if that's all that's at stake here, it seems very reasonable.

No, the opinion doesn't have to do with whether OpenAI is open.

The argument doesn't hinge on whether OpenAI is actually open. Rather it seems to have to do with the name being insufficiently distinguishable from a generic term ("open AI"). I think it's a bizarre ruling given that everyone already knows what OpenAI is.

Give me a break. Apple doesn't sell apples.

When you read a book, you hypothesize a writer and have some kind of inner relationship with the hypothetical writer. But it's not a real relationship. It's in the words, not outside them.

To the extent that people hate tech, it's because:

1. Tech is our "new money". New money is always hated, by every generation

2. Tech represents change and the human animal dislikes change. Most people are okay with everything that exists when they turn 25 and dislike everything that appears after

3. Tech represents capitalism in a pure form. Creative destruction reigns; companies and apps can rise and fall overnight. Ways of life are precarious and the constant churn bothers people

All three really come down to people disliking change. Instead of hating/fearinf change and dynamism, people should look inward. Establish a routine. Exercise. Garden. Read. Cook. It's more within your power than ever before to decide what you like and pursue it at your own pace. But don't expect the external world to meet your desire for stasis.

Your example hinges on whether a bunch of words is true, not on how they came to be written.

When we read personal stories it affects our emotions as we empathize with the author, or otherwise share the feelings that the author is trying to convey. When we find out there is no such author, our empathy and our notion of shared feelings vanishes with the new information even though the words stay the same.

This isn't true for me. If I read an incredibly moving poem and later learned that it was written by someone casting the I Ching and picking words out of a hat, it would not affect how I felt about the poem.

It seems to me that you don't like reading (which is fine). Some people enjoy reading words strung together in a certain way. The value comes from the simple relationship between the text and the reader, not on some kind of social connection.

In my mind, it's pretty simple: I'm a human, LLMs are not. If a human writes a novel, it's inherently worth more.

While I appreciate you laying it out so plainly, I disagree. A novel is a bunch of words and I don't care if they were written by one person, five, an AI, or infinite monkeys on typewriters. What's valuable in a novel (or a poem) is in the words.

Please actually respond to the comment instead of whatever this is. It looks reasonable enough to me and I trust the FDA more than The Guardian or the "Tucson Environmental Justice Task Force".

We look back on the (at the time influential) claim that rock/metal/etc "corrupts the youth" as a quaint moral panic. Modern views about social media are our version of that same silly claim. Future people will look back on all these comments comparing Facebook to crack with the same amused bewilderment as we look back on the PMRC.

Children are not emotionally or intellectually prepared to repel this hostile takeover of their minds.

Then their parents shouldn't let them use the internet.

I find it interesting that so much of how people think about morality involves attributing free will unevenly. I.e. "facebook execs" are using their free will to addict people but those people have no ability to resist. It's so obviously corrosive to think something like "only evil people have free will; good people are just hapless victims".

Following this logic, isn't "your part" undone until you commit suicide?

There will, in fact, continue to be frogs.

One theory of where posts like this come from: https://www.pnas.org/doi/10.1073/pnas.2527316123

If anything these models should be compelled to be public since they have been trained off public data. What an absurd overreach to call this an attack.

It’s clear they are scapegoating national security and China at this point to build an anti-competitive moat.

If all that is required to train these models is public data, why can't Alibaba just use that?

The fact that Alibaba has to resort to scraping Claude suggests there already is a moat...

I don't assume it; it's simply true. Educated/elite Europeans tend to define themselves in opposition to Americans. It's pretty hard to interact with those Europeans (including on this site) and not pick up on that.

There are plenty of Americans who side with the Europeans and also define themselves in opposition to "the kind of American" who has AC/eats fast food/is obese/has no culture. I'm from New England and maybe even a majority of people have that perspective.