Selkirk

How many decisions per hour of work do you make? The answer is a lot. Most of them aren't worth recording. So, the questions that need to be answered in my mind are "What makes this decision worth recording?" and the related "How is having this ADR going to impact what agents actually do in the future?"

I guess the benefit would be that agents could catch future or related tasks via the codebase "greps" and could theoretically design toward the future or keep them fresh? Is that worth the obvious downside? Keeping old tasks seems like a bad idea. How do folks address stale tasks without this tight coupling? A stale task description isn't a problem for people, but agents are so literal.

Interesting observation. Plans are abstractions. The missing model capability is the ability to recognize work at different levels of abstraction, and understand which level of abstraction is called for and to be able to work at multiple levels of abstraction at the same time. The LLM architecture elevator has a bias for down.

Theoretically the value of systems like Superpowers should be regulating bias like this.

I really like this writeup. I find the metaphor satisfying, though there's a reason metaphor receded into the bushes as a practice in XP. A great metaphor is great to have, but they don't always exist. Fortunately with organizational systems, things will tend to rhyme. For example, a role that "Verifies that all processes have been followed before anything merges" seems like it could fit many different systems. We're at this fun stage where people are just trying stuff and I love it. I hope we get more info from this experiment.

From a user experience standpoint this sounds like a clear win. Congratulations.

With AI safety top of mind so much lately, I can't help but notice the announcement does not address this.

With "Chat" mode, there was a user expectation that session had only limited capability to produce unexpected side effects, read sensitive files, etc.

With "Cowork" mode, it seems like more powerful capabilities have been on by default, requiring deep settings and safety understanding to disable if desired.

Merging modes feels like it is removing a simple and easy to understand risk management tool. How does the combined mode help users understand, manage and feel confident about what risks they are accepting?

The dashboard for the Andon Market shows a bank balance of around $7k remaining out of the $100k seed. Rent is due at the end of the month. Last EOM brought a $11k-ish drop. Sales do not appear to be on track to cover the rent that's due in two weeks. What then?

The linked article is intentionally misleading by omission because they left out "in mice" in the university driven article and they certainly know the relevance and consequences of leaving it out.

I bought the "right kind of Dell," for this purpose but ddpm is unusable for me as it slams CPU periodically.

I'm missing some data to evaluate this article. If you engage a set of humans on these seven tensions, what range of bias do you get?

Multi vitamin use has been studied A LOT with respect to all cause mortality, where it consistently has no statistical effect (or a slightly negative effect). So, to me, "biological aging" seems a way to hack for effects to advertise. Does the average consumer understand that reduced "biological aging" does NOT mean you will live longer? Because if it did, they would be saying THAT. Centrum (study sponsor) is already creating advertising based on this study.

Partially funded by entity related to manufacturer of daily multivitamins. Study was in people over 60 average age 70 if I recall. I didn't care enough to look, but the question I'd ask is how were the people who died during the study accounted for in the "biological aging testing?"

I have a colleague that recently self-published a book. I can easily tell which parts were LLM driven and which parts represent his own voice. Just like you can tell who's in the next stall in the bathroom at work after hearing just a grunt and a fart. And THAT is a sentence an LLM would not write.

So ... Test Driven Development?

1. Planner (Write a failing test or tests) 2. Executor (Generate a solution) 3. Verifier (Until the tests no longer fail) 4. Repeat

Sounds like "Old Masters and Young Geniuses"

https://press.princeton.edu/books/paperback/9780691133805/ol...

That book used artwork valuation as a performance measure and analyzed it over top artist's lifetimes finding two patterns. The "Young Genius" where an artist has a vision and realizes some innovation and their most valuable works center around that with value tapering off over their life. Picasso. (Who had two peaks but still fit the pattern.) Contrast to the "Old Master." This is someone who keeps refining their craft and their most valuable works and innovations are their late life works. Cézanne.

Looks like game theory interaction strategies presented with a bit of color:

Stupid = Spite (lose / lose); Helpless = Altruism (lose / win); Bandit = Selfishness (win / lose); Intelligent = Cooperation (win / win)

The OP "Laws":

Everyone underestimates the number of spiteful individuals in circulation.

The probability that a certain person be spiteful is independent of any other characteristic of that person.

... and so on through ...

A spiteful person is more dangerous than a selfish person.

Not sure about other chemicals, but carbon seems particularly hopeless. Plants produce CO2 as they respirate and consume it as they photosynthesize. So rather than think in terms of number of plants, think in terms of the weight of the plants (not the dirt). Carbon sequestration is going to have to add net weight to the plants. A human produces 2.3 pounds of CO2 per day. So, how many pounds per day will your crop of carbon scrubbing plants have to increase by in order to sequester that carbon? How much green material by weight do you need to achieve that rate of growth?