carver

I agree that if a sub agent could work then handoff is the wrong skill. These docs on the site agree too:

> What it buys is portability, not compression. That makes the skill narrower than it sounds. You need a file only when the work has to travel — to a new harness, a new directory, a colleague, or a side task you want to fork off. If nothing is travelling, you do not need a handoff: staying in the session, /clear, a subagent and /compact cover the ordinary end-of-phase case, and /compact covers it more often than this skill does.

https://www.aihero.dev/skills-handoff

(Emphasis mine)

What a great seminar, that was. I really appreciated his advice on writing recommendation letters, too: the expectation is shifted wildly towards effusive. If you are plainly complimentary, it can come off as a secret warning that you don't think they are worth hiring.

But there were also great AI papers, and meta advice on reading them efficiently. (I don't remember any crimes against ferrets, but presumably the reading list changed over time)

I appreciated that class, and it's only grown on me over time. Another line that really stuck with me was something like "forsan et haec olim meminisse iuvabit" (Which I remembered as "Perhaps we will look back on even this with fondness") It's so easy to undervalue amazing things when they are happening to you. I was really convinced that I was appreciating it, even more than many around me. But I still look back and think I could have soaked it in, even more.

If you are interested in variants of minesweeper, this one is awesome:

https://magnushoff.com/articles/minesweeper/

(scroll to the bottom)

Ahah! I see it now, thanks.

Fascinating. Thanks for the clue. It's the most complete blank for me of all the tests. I looked up the reference image, but still cannot see it in figure 2G. I can't even guess at where the eyes/nose/mouth are in the clouds.

It seems that Aphantasia does not globally bin into two groups, since I don't fit in either.

By my rough count of Figure 2 tests, where Derek is at 0 to Loren at 6 (ignoring F), I have about 3.5 atypical responses.

My experience with Figure 2:

A) I can flip between cone and weird triangle, saw the cone first

B) I see it as if someone placed identical cat stickers on the drawing. I can intellectually understand the perspective, how the upper-right one is supposed to be bigger, but don't experience it that way.

C) I see that there is an implicit rectangle (to me it looks slightly wider than tall). But the color doesn't "spread" to the middle, it's just like 2A -- a boundary in the surrounding shapes implicitly extends into the empty space to form a rectangle shape.

D) It takes minimal, but non-zero effort to see the vase

E) It's trivial to flip between the two orientations of the cube

F) skipped

G) I don't understand what I'm looking for here. I see clouds, sky, and a silhouette with a tree. Is there a face in it somewhere? I can see the smiley face on https://en.wikipedia.org/wiki/Pareidolia

Is this the "uh, thing"?

https://www.theatlantic.com/politics/archive/2015/01/the-blo...

I don't buy that the reversion to income tax should happen on high percentage gains if the absolute gains are low. I think the tax could be made even more progressive. Something like:

The first $500k of capital gains get current cap gains treatment. After that, all gains are taxed as regular additional income.

Based on context, you seem to be referring to gas when you say 12 million "ether gwei" (the block limit being 12 million gas, lately). But gas is its own unit, not measured in ether (or in wei, the smallest atomic unit of ether).

The reason the math still works is that 12 million gas priced at 1 gwei/gas would cost 12 million gwei, which is .012 ether. (Since one ether is defined as a billion billion wei) ... Maybe this was all what you meant, but it was hard for me to decipher.

Yeah, digital-only NFTs will likely be a fad. I'll be interested again when the trend moves toward deeds. A title for a physical work could have staying power. Even potentially an iou from an artist for a one of a kind (physical) piece.

> Hypothesis is pretty smart to find edge cases that you usually don't think of.

This is worth reemphasizing.

For certain kinds of hard problems, I've found it really valuable for building confidence in an implementation.

If you happen to have a reference implementation, with a small surface API, it can be a very fast way to write tests: tell hypothesis to input anything it wants and validate that both libraries produce the same result. It's not magic, you will probably still have to give hints about how to build interesting inputs, but still a very quick way to get a lot of coverage. (Also be sure to crank up the number of examples to build more confidence, 100 is often not enough to really explore an input space).

Tail call optimization is appealing. Sometimes recursion is the cleanest implementation, but call depth is not known up front.

RIP Pat Winston - he infuriated me the most when he said something so obviously true and which I wanted to be false. One example: "No one will ever read your paper in entirety, so write for the skimmers." As much as I loved his technical lectures, I especially remembered his peripheral aphorisms.

But he said one thing that I immediately agreed with (conceptually) and have recalled at many hard times in my life: "Perhaps we will look back on even this with fondness." (But in Latin, because Pat Winston)

`if re.match(p1, line) as m`, maybe?

Was this a traceable ring signature[1], or something different?

[1] https://en.wikipedia.org/wiki/Ring_signature#Applications_an...

It's even worse than that: politicians created over-reaching laws and enforcement, to demographically target political opponents and take away their right to vote.

> "The Nixon campaign in 1968, and the Nixon White House after that, had two enemies: the antiwar left and black people. You understand what I'm saying? We knew we couldn't make it illegal to be either against the war or black, but by getting the public to associate the hippies with marijuana and blacks with heroin, and then criminalizing both heavily, we could disrupt those communities. We could arrest their leaders, raid their homes, break up their meetings, and vilify them night after night on the evening news. Did we know we were lying about the drugs? Of course we did." – John Ehrlichman, to Dan Baum for Harper's Magazine in 1994, about President Richard Nixon's war on drugs, declared in 1971.

Meta source: https://en.m.wikipedia.org/wiki/War_on_drugs#20th_century

The link said triple blind, technically -- ”Masking: Triple (Participant, Investigator, Outcomes Assessor)”

> You can only do that if the transaction is atomic, which requires that the tickets and money are on the same blockchain and exchanged in the same transaction.

Yes, they should be on the same chain, so that the transaction is atomic. This is trivial to do in Ethereum (and other blockchains, presumably).

The most natural follow-up to currencies on a blockchain are assets, like a ticket to a concert, shares in a company, or a deed to a car.

Once more assets are represented this way, many kinds of applications would be valuable. One simple and powerful example is escrow.

Enabling one-shot, instant escrow with strangers on the Internet has real-world benefits. For example, the current safest way to exchange Burning Man tickets is to meet someone in person, inspect the tickets and cash, and trade them by hand. If they were a blockchain asset, they could be traded nearly instantly and safely across the world. With a stable dollar asset, the official escrow service could even make it more difficult to sell above face value.

Agreed, in most situations.

One counter example:

If your processes in sum tend to A) access many disk locations, at large total disk space B) hold a lot of underused data in ram

This isn't totally impossible. Maybe an Ethereum node with a script doing a bunch of data reads, running side by side with hundreds of Chrome tabs, few of which are regularly accessed. (Totally hypothetical, of course...)

Swapping some rarely used ram out so the OS can buffer disk into ram seems like a reasonable approach (although maybe even more reasonable is: close some tabs).

Your point still stands that more ram is strictly as good or better performance in this scenario, but you might be able to get an equivalent performance boost much more cheaply with some swap space for the underused ram. Also, upgrading past 32 GB ram starts to veer from expensive to impossible on a laptop.

> And worse, its energy consumption is not a function of its transaction volume, but of its price in fiat.

It's also a function of time, dropping in half every ~4 years. (Rounding transaction fees to zero)

Which isn't to argue against proof of stake, just to clarify the current situation.

> Also, Ethereum is in a quagmire regarding scaling.

"Quagmire" is a bit provocative, but yes scaling is a serious problem that needs to be solved.

> In fact, I do believe it's cheaper to spend bitcoin now than it is ether ($0.25 in gas per transaction).

This is no longer true. Since the recent fork/upgrade, it's been consistently cheap. Specifically, an ether transfer with an $0.001 fee rarely takes more than a few minutes to confirm.

That's terrible news. I was undergrad there in mid 2000s, and perceived that some, but not all, of these issues had improved. Being a man, and less attentive to the issues at the time, it doesn't surprise me that I missed it. It could also be that the undergrad experience is better than the research staff experience.

Is there anything on the list that has improved?

Could you explain more about where your feelings come from?

>There is no way you could push down the market 10 percent today.

GDAX fell about 8% yesterday, from $96.82 to $89.07 in one hour (May 4, 9-10am) with about $2 million worth of trades. It's a little early to call this a post-volatility era.

It depends on the context.

Ethereum is a platform that uses a blockchain to build and record consensus. Conceptually, the blockchain is a technology underpinning each cryptocurrency. Across the many cryptocurrencies in use, there are many blockchains keeping track of who owns what.

It is never correct to refer to a capitalized "Blockchain," or refer to it singularly without articles (a or the), imho. We don't refer to other data structures that way, like a priority queue.

You've nailed it. In this case, the call stack depth would halt you before TheDAO runs out of funds, though. So you end up with some multiple, say 30x, where you can put in 1 ether and drain 30 ether. There is about a month lockup period during which your seed cash and your multiplied cash is frozen.

What made this particularly devastating is that there was a second bug which allowed the attacker to transfer their token that represented that first ether back to the DAO. So instead of 30x return, they could eventually drain the whole DAO (although they stopped after the discussion of a fork began).

I'm largely basing my understanding off this article: http://vessenes.com/deconstructing-thedao-attack-a-brief-cod...

While I prefer non-upgradable contracts for the same reason, there are mechanisms to reach some middle ground, like:

1) a significant delay between the announcement of an upgrade and the actual upgrade (enforced by the blockchain)

2) a mechanism to opt-out or cash out before the upgrade happens

The US government freezes bank accounts all the time.

You might be looking for the Kelly criterion. It provides a quantitative answer to what percent of your bankroll you should bet, given the characteristics of a game. Roughly, it tries to avoid having you run out of money before you start winning.

So an answer to the paradox is that some people, like yourself, have intuition about the Kelly criterion: that they should limit the portion of their bankroll that they bet.

One of my favorite techniques for enforcing two-way connections is the "Jeopardy technique". Here's an example for a beginner in python:

  Side A: This library is the de-facto standard for http in python.
  Side B: What is the "requests" library?

This card is easily reversible in Anki, so you only have to write it once and get quizzed both ways.

Edit: formatting