Let's be very clear, mass surveillance is exactly what this is.

1. It's being sold under the guise of license plate tracking (also in TFS), but there are Flock cameras installed in places where there is not a road in sight, such as in parks and on playgrounds. See for instance: https://www.reddit.com/r/Flock...
2. They automatically pan and zoom to faces and phones of pedestrians: https://www.youtube.com/watch?... (also, their security is apparently dogshit).

You mean the lawyers? Trying to prevent lawsuits from people who thrive on litigating away their own responsibility?

China censors stuff in the way you think it happens, but American companies are driven primarily by greed. If they could get away without putting in guardrails, they would very, very happily do so.

No, you are the first one to do so; You are very special.

1. The American political system is corrupt as fuck. _That_ is the problem in regulatory capture situations. You're trying to come up with a _workaround_ instead of fixing the core problem.

2. Anti-trust law works when it is applied hard. The EU has had plenty of successful cases where corporate behavior was effectively changed. But the organisations enforcing these matters are absurdly underfunded by decades of corpo-friendly right-wing governments throughout the Western world. Said differently: we haven't really _tried_ to do anti-trust well. Having said that, I am in favor of more structural changes like making corporate tax progressive, i.e. with multiple tax brackets.

3. Providing transparency is great, but the mistake in thinking is that people, even with full information _can_ realistically make different choices. On a geopolitical scale countries buy oil and LNG from heinous regimes because they have nowhere else to go. It is naive and wishful thinking that consumers will somehow force producers into behaving if only they had full information.

4. Counterpoints for nationalization: Your adversary _buys_ the companies that run your essential services, like utilities. Woops, should have nationalized or very, very, very heavily regulated them. Or: the industry is fundamentally unprofitable or profits more when society is doing badly. The US has a very, very wealthy healthcare and defense industry. Those fuckers are laughing all the way to the bank in times of war, prolonged and chronic disease, etc. The opioid crisis was/is a product of capitalism.
Nationalization doesn't mean "Soviet-style". There are many, many purely publicly funded organisations that work fantastically each and every day.

The key thing to realize is that the main strength of capitalism is in speeding up evolution. It _can_ do that, but we don't _need_ it in a lot of markets and 'late-stage' it favors stabilization and hegemony of a dominant 'species'. Again, it needs to be heavily regulated, so as to apply it precisely rather than seeing it as a blanket solution.

This. Too many free-market capitalism anti-government evangelists miss that most of the 'bad'/immoral bits of behavior by producers is completely natural and rational. A cartel is just effective cooperation.

The point is not to get rid of capitalism or free markets, but to acknowledge the strengths and weaknesses and effectively channel those. Given where Western developed countries are today, that means more regulation (and enforcement) and even nationalisation of certain aspects of society.

Yes, because curtailing gambling is unheard of in the US. Just ask the 8 states where online casinos are legal.

Is that how the holographic principle works??

Better would be to create regulation to ensure such devices receive security updates for a reasonable period of time and thereafter must be open sourced for community updates. Additionally, after such legally required support goes away or the device is otherwise in a permanently insecure state, replacing it by a secure device should be subsidized.

Even better would be to enforce that all routers must support running alternative open source firmware.

https://www.smbc-comics.com/co...

Is that better or worse? I was under the impression that most people find the catching of a rocket booster like SpaceX does with those little arms to be more awesome than just landing the booster.

I also thought that outside of having to land on Mars it is the preferred approach because it is more efficient.

Finally: Caught it in a net conjures up the wrong image. If you look at the video the 'net' is much more like the mechazilla arms and not some fishing net they plop the booster into: https://www.youtube.com/watch?...

Otherwise: go ahead and tell me what evidence is required for a philosophical claim like this. You act as if you know.

Do you know how philosophy works?

Bullshit.

"I won't express a strong opinion on the philosophical claim - I do not feel qualified to assess whether this is really analogous to a global workspace"
and
"I feel highly uncertain about what evidence it would take to show models have moral significance or consciousness"

Just admit that you're conjuring something out of nowhere and pretending that somebody agrees with you even though they clearly don't. Otherwise: go ahead and tell me what evidence is required for a philosophical claim like this. You act as if you know.

Furthermore, all of this is completely besides the point that this research has been replicated, which renders this other line of yours false:

You desperately wanting something to be true doesn't make it so. If you're going to quote the commentary, do it right:

"I won't express a strong opinion on the philosophical claim - I do not feel qualified to
assess whether this is really analogous to a global workspace, and this feels like the
least interesting claim to me. There is clearly something significant J-Space is finding
inside models, and this is advancing our understanding of them and ability to make them
safer, which is the important part, whether or not it is analogous to a global workspace.
  This hypothesis did seem to make useful predictions about the technique's
properties, but it's easy to read too much into post-hoc analysis of results like
this.
  I feel highly uncertain about what evidence it would take to show models have
moral significance or consciousness, and this paper didn't move me much on that."

A far more interesting and nuanced take than your naked dismissive statement.

Except the concept of "working memory" is very much agreed upon, with various empirical results showing its properties and limits robustly. Similarly for the distinction between conscious and subconscious processing.

As a GP said: RTFA instead of whoring for karma.

They requested external commentary, which includes reproduction by Neel Nanda: https://www-cdn.anthropic.com/...

"Replication
We created our J-Lens for Qwen 3.6 27B by taking Jacobians to the penultimate layer on
twenty-five prompts from the Pile of length 128 tokens (some experiments used wikitext),
skipping the first four tokens as they had high norm. We note that as this is a different and
weaker model some results should differ. The important question is whether we see broadly
similar phenomena."

In general, this is very significant progress in understanding and measuring what happens inside these models. Linking the J-space to consciousness is tenuous, but you are definitely being too dismissive of it.