I have dedicated checkpoint and commit skills. Checkpoint has it run my review skill, then run my commit skill. The commit skill just asks it to suggest commit messages for my review. Having these skills in place has resulted in claude preferring to use the commit skill instead of committing on its own, and suggesting commit messages to me before actually doing the commit.
Wait I'm sorry but if the policies are to cover catastrophic damage, but if catastrophe actually strikes and the insurance company becomes insolvent, What's the actual point or purpose of insurance then?
What do we care more about? The present? The current generation? Or future generations? Our children, grandchildren, great grandchildren? The perpetuation of human civilization? The perpetuation of human civilization with a set of values that that supports the growth and happiness of all?
That's the first question, right? Alignment. Sustainable alignment.
From there, it's all about sustaining those shared values, and minimizing risk.
Build bonds: keep your friends close and your enemies closer. Or establish trust. It works out better in the long run; people don't usually Like to stab each other in the back
Trampling starts cycles of revenge, and that's no fun for anyone involved either
Exploitation prioritizes short term gain and screws you in the long run, more often than not.
Your two examples are extremely different.
Walking away with $9 if you're unlucky and walking away with $0 if you're unlucky are two incredibly different outcomes.
Being successful from winning the lottery, versus being successful from working hard enough such that you qualify for a lottery with significantly better odds.
Sure luck is Involved, the outcome is determined by luck, but work is putting a hand on the scales. And in that sense, your prior comment of it being 100% luck is exaggerated
There's always someone to support you when you need it, someone who will agree when you're right, who will call you out when what you say or the way you phrase something is inaccurate or wrong, but will redirect to focus and support on where you're right or what points have value instead of outright shutting you down.
I'm not saying that AI now does this, nor that billionaires all have this experience either. It's seems like billionaires only get the upsides of this kind of this behavior without enough of the pushback. Or any of it. Which might be your point.
But too many people just don't get anything like this. Helpful critical feedback or positive encouragement and support from relatively knowledgeable sources is an unfortunately scarce resource.
I think there's a broad spectrum of people, some of whom are role playing, some who think there are no consequences, some who have strong distinctions between the animate and inanimate, and some who just do what they think makes sense
You say this like "oh it's just acts on a different kind of opiod receptor" means that it's not meaningfully different.
Your comment also seems to imply that this treatment involves consuming ibogaine habitually or regularly
The protocol for ibogaine treatment, according to the article and the experiments being done, do not have this requirement.
Like other treatments involving psychedelics and hallucinogens, the protocol here is that a one-off treatment, a single dose, results in meaningful improvements in both addictive behaviors and PTSD symptoms a month later and potentially longer
This is not the same as something like methadone or naltrexone, which _are_ what you describe: replacing a more harmful opiate with a less harmful one.
A [Code Property Graph](https://en.wikipedia.org/wiki/Code_property_graph) takes a codebase and turns it into three graph representations: it's AST, a Control Flow Graph, and Program Dependence Graphs. These graphs are overlaid and shoved into a single property graph. It's a structure mainly used by some static analysis tools like [Joern](https://joern.io/)
---
This has been a topic of a lot of interest and research for me. I've been experimenting with figuring out a system inspired by these ideas, among others, to apply the same idea (shoving multiple graph representations together) to a broader set of information
This penalizes honest mistakes, or moop from prior years resurfacing, or wind blowing trash into camp, or any number of things that are outside of a given camp's control
The moop map, and community holding itself accountable, seems to be a decently functioning system.
Not to mention the administrative overhead, at the org level and at the camp level.
Frankly being a camp of 100+ people, not just taking dues but also handling this Deposit, and distributing the cost fairly?
Running a camp is enough of a pain in the ass without adding on this kind of thing.
Monetary incentive systems like what you're suggesting are just a way of enforcing culture. If culture spreads organically, why bother with the overhead of bringing money into the picture?
Honestly, that contrast is what draws me in. In the same way ultralight hiking forces you to think about and let go of extraneous weight, going to Burning Man and doing the whole camp thing and seeing the city work showcases the "dead weight" of "making things happen".
They aren't referring to the regulatory requirement, but the response, I think?
Like if people can put in this much time and effort in a remote desert environment to meet regulatory requirements, and document their efforts so thoroughly, why can't corpos?