Jarwain

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.

My understanding of the average age of a human being 30s-40s is that the average was skewed downwards heavily due to high infant/child mortality rates

I mean, there are some pretty famous DJs out there

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?

Imo it depends on priorities!

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.

Seems like they want to bring in retail

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

With LLM's, I wonder how far away we are from "a cooking ground for people to experiment with ideas"

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.

There's some prior work on the codebase thing:

[Unison](https://www.unison-lang.org/docs/the-big-idea/) content addresses every definition. Kinda interesting.

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

I'm a big fan of bolt hangers personally. They also mark the spot, provide a nice anchor point, and don't require chopping!

Not much difference between a 12" and a 18" lag bolt for the purpose of "how much trash is visible and impacts terrain".

Surface feels a bit fairer in that sense. Or at least, easier to measure.

If it was that'd be an absurd amount of weed being left behind to make a mark on the map.

The problem isn't that there aren't solutions, the problem is getting everyone on board

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?

My camp, while doing our moop sweep in 2023, found lag bolts from prior years!

2023 was a weird one, because of the heavy rain and so many people not being used to it.

But it also seriously churned the Playa, revealing what was hidden for a whiiile

I lean towards a variant of option 1: you can only generate an article that was previously referenced. But arbitrary phrases, clauses, sentences, paragraphs, can be highlighted and used to form a new article.

Yes this may mean that there are pages for common words like "and"

Yes this may mean that there's a page for letters like "x"

Filtering what ends up becoming a hyperlink becomes a problem that I think can be solved with regex/whitelisting

I think articles should have a backlinks drop down. Might make consistency easier As well as generally just plain text search to pull relevant articles or context when generating a new article.

LLM's are a mediocre map, but they're a great compass, telescope, navigation tools and what have ye

The default model is Opus 1M context, so autocompact doesn't run as frequently, and that just Devours your session budget if you're not careful. There are some env variables you can (ask claude to) set to lower your max context window and autocompact threshold.

Technically aren't most android apps limited to ARM?

I agree, but I do feel like there's a bit of a lag. The shape of this lag has changed over time, and maybe we're in an era where the lag is growing

That's fair!

I feel like a permission list is kind of a superset of a block list and an exclude list. Or they're all different perspectives/solutions to the same kind of problem, that a permission list is the more generalizable solution for.

Or it's a way of framing the problem that doesn't embed the "exclusion" idea in the naming.

And it kinda bridges over to the idea of Access Control Lists a bit better?

I'm not trying to argue about validity here, but rather that these definitions/meanings of the word black are not "primary" definitions but secondary meanings based on that contextual/cultural/colloquial use. Arbitrary in the sense that that "commonality" is arbitrary and cultural, and language could just have easily developed to flip the colloquial definition.

Contrasted against using words where the Primary definition is the one that matters.

Imagine an alien culture encountering the word. Blacklist versus Denylist. The latter requires a lot less context to understand the meaning, because "Deny" has a single pretty consistent definition.

Thank you for sharing the etymology! It's quite interesting, I agree!

I may have been a bit too pithy/I sufficiently clear with that last statement I made.

I meant it in the sense that understanding the word relies on a lot of contextual/colloquial/cultural understanding that's typically gained via time and exposure. At least, more of it than allow/deny requires.

Imagine an alien culture encountering blacklist vs Denylist. The latter requires a lot less context to translate, because Deny is used a lot more consistently.

My argument is mainly one about _clarity_, not hurt feelings.

How is "allowlist" or "denylist" not more clear to, say, someone for whom English is a second language?

Sure blacklist was already an English word, but it's not necessarily _common_, and the distinction between blacklist and whitelist is kinda arbitrary. If you'd like to explain Why the word means what it does I'd love to hear it

Allowlist and denylist are clearer, in that the meaning is in more clear alignment with the words it's made up of.

The old terms just make more sense to those who are old enough to be used to it.

I'd argue evolution is a mechanism that nature uses to engineer genetics, sure