"So, for at least the next few years, this definition should provide a good outline of what constitutes artificial intelligence, and it avoids the philosophical issues that dominate attempts to define the meaning of either artificial or intelligence."

Still valid I think... 🧪

But it was rather inconsidera te of you, not being able to concentrate fully on providing entertainment for Euler!

(Though....bonus pets is one of the few good things coming out of the proliferation of videoconferencing 😁)

Thank you for the great talk!

I do not think directly. You could create a personalized feed with a tool like skyfeed.app, but that works best for positive filters (i.e. to globally/selectively follow Tags or similar), not "My feed, but without stuff I do not care about" approaches.

Thank you for this link, though I am especially interested in one aspect the guide mentions in a single bullet point "Are the statistical design and analyses appropriate?"

I am versed in the general review process, I am just hoping to become better in reviewing wrt statistics (and eventually ML).

To help with the first problem (reviews) while working on the second (workflow):
Are there training programms (or guides) to help become a better reviewer (And thus also author and collaborator) with regard to statistics?
I'm not sure I completly trust the sufficiency of my statistical experience...

That book sound so interesting!

Do you know of a (legal) ebook edition? Sadly I find only, rather expensive, print editions...

No. The theory, which I would put under test, is that they reduce harm. Domination would be the easiest to show an effect, but it is by far not the only way how such a reduction could show up/fail to show up.

Anyways, I think we will not get agreement on this. But thanks for the discussion :-)

Different legal systems do. (Also I clearly argue for a hypothetical, as sadly I do not remember any such rules to be in force. Other than maybe the "novel food" (and related) regulatory approach of the EU, which basically says "You cannot feed it to people unless you show us lack of harm")

Don't we need to study the influence of a gun control measure (as in what happens when we drop it) to argue they lack justification? (i.e. what I wished for?)

This, I think, is where we definitly have to (agree to) dissagree. I can see too many ways to game such an imbalanced playing field (in the lack of regulation direction).

(I agree that the enforcement aspect against individual citizen can/will be problematic...)

No it's not on you.
But it is on us to not overstate the degree to which the results of our papers generalize. So, I would say, your paper sucessfully restricts the degree to which the relationship between gun control and gun related deaths can be a straightforward relationship.

Sure, but I would forcefully argue then, that at least _all_ policies wrt gun control are held to the same standard. Getting rid of requirements to centrally track ownership? That is an "inverse" gun control measure, and should be treated as such then...

That is not to say that people arguing for specific measures could ignore your, or similar, analysis. But I expect there are mechanistic models that would expain your results, while still allowing for the overall argument...

I believe there are more assumptions in the analysis and confounders that could overpower any correlation.

How long until the "marginal gun" is well controlled? How effective was enforcement/How well were incentives taken up? Were there social shifts in that time? etc.

That goal, we do definitly agree on :-)

Are you arguing that the risk that needs to be justified with regard to gun control _is_ police violence? Or that they enable the "legally justified" increase in police violence?

How?