jkhdigital

I can’t stop thinking about this exact notion. The main reason we don’t always use stuff like TLA+ to spec out our software is because it’s tedious AF for anything smaller than, like, mission-critical enterprise-grade systems and we can generally trust the humans to get the details right eventually through carrot-and-stick incentive systems. LLM agents have none of the attentional and motivational constraints of humans so there’s no reason not to do things the right way.

That last line is exactly what I was thinking. Find an expressive language and then progressively formalize your workflows in DSLs that enforce correctness by design, not through layers and layers of natural language “skills” and deadweight agentic watchdogs.

Still basically relies on feeding context through natural language instructions which can be ignored or poorly followed?

The answer is not more natural language guardrails, it is in (progressive) formal specification of workflows and acceptance criteria. The task cannot be marked as complete if it is only accessible through an API that rejects changes lacking proof that acceptance criteria were met.

7 years ago, my graduate distributed systems professor required everyone to complete his projects in Elixir because it was trending on HN. It was my first functional language and after getting over the initial hump I fell in love with it.

Now, I’m teaching undergraduate courses of my own and, while I do not have the flexibility to change the languages used in my current offerings, if I ever start teaching a systems programming course I will absolutely require the students to use Pony.

I often hear a different version of that quote: You can be right, or you can be happy.

Love this example. I started putting my Kindle highlights in the SRS—no prompt or anything, just the quoted text verbatim—and the effortless periodic review essentially burned the quotes in my memory for easy recall in moments when they were appropriate.

Yes, all behaviors should be implemented like definitions in category theory: X behaves like a Y over the category of Zs, and you have to recursively unpack the definition of Y and Z through about 4-5 more layers before you have a concrete implementation.

The tectonics.js blog has some really incredible write-ups on how to do proper simulation of plate tectonics: https://davidson16807.github.io/tectonics.js/blog/news.html

I believe sailors from the USS Ronald Reagan were deployed to the disaster area to help, but that was not my ship.

I can tell this is much more than just “Tabletop Simulator on a tablet”, although at $500 you’re likely to get a lot of attention from the Twilight Imperium and Gloomhaven crowd. I know more than a few childless people in my local gaming circle who would drop a half-large on accessories that simplify game execution.

But clearly this product isn’t about making existing board games easier to set-up/play/clean-up. I think the marketing dept has a lot of heavy lifting to do, convincing buyers that this isn’t just Juicero for existing board games.

You’re right, you don’t know! When I worked on a nuclear aircraft carrier, I learned that the people working down in the reactor spaces got significantly less radiation exposure than the people working on the flight deck. My ship was ported in Japan when the Fukushima disaster happened, and we had to abruptly go out to sea because we couldn’t let the (minimal) fallout contaminate the reactor spaces and make it impossible to monitor the reactor itself properly.

This is about assisted suicide. You can argue all day long about how you have the right to end your own life, but the real issue is whether you have the right to grant another person immunity from charges of homicide for facilitating your death. That is an entirely different beast.

I would argue that banning it is not an “easy solution” but in fact the hardest solution.

There’s no bright line. Suffering is a subjective experience. Lack of prohibition is tantamount to encouragement.

How many “legitimate” assisted suicides are worth one questionable case? There are no foolproof guardrails, and you’re inviting moral hazard.

Hard to take seriously a paper that begins with a statement that the “uncritical adoption” of combustion engines was a “blunder” by scientists.

About 20 years ago I failed out of the undergrad CS program at UIUC because I thought I was smart enough to skip most lectures. I did manage to get an A in the C++ Data Structures course because the lectures were recorded and I just binged them all the night before each test.

Anyways, now I’m a full-time lecturer teaching undergraduate CS courses (long story) and I’m actually shaping curriculum. As soon as I read this article I thought “I need to tell my data structures students to read this” because it echoes a lot of what I’ve been saying in class.

Case in point: right after two lectures covering the ArrayList versus LinkedList implementations of the Java List interface, I spent an entire lecture on JUnit and live-coded a performance test suite that produced actual data to back up our discussions of big-O complexity. The best part of all? They learned about JIT compilation in the JVM firsthand because it completely blew apart the expected test results.

I think the plastic food displays are due to high uncertainty avoidance, so patrons can see exactly what their meal looks like before ordering. Yes you could use real food but the hassle of periodically filling the display case with freshly cooked dishes would be silly.

Japan’s fairly simple urban zoning scheme seems to work quite well. There’s essentially a sliding scale from light residential to heavy industrial, but “lighter” uses are always allowed in “heavier” zones.

The current war was initiated via a sneak attack by Hamas militants from Gaza who committed rather unspeakable atrocities against Israeli civilians. Not justifying their actions but as an uninvolved third party I’m going to at least acknowledge the history of the situation before judging their actions.

Your Panglossian description of the purpose of publicly-funded research shows very little understanding of reality.

Intolerant theocracy? And you’re using this phrase to refer to Israel?

This is the salient point that is glossed over without so much as blinking.

To get 1J of biofuel-based fuel, how many J of fossil fuel is burned? If it’s like 0.9J then you really have to ask how much carbon is released when turning carbon sinks into farmland, because it will take many many crop cycles to recoup that from biofuels.

Except that industrial-scale agriculture in general requires quite a bit of exogenous petrochemical inputs—not just the fuel to power the machines that do the work, but also the fertilizer.

My PhD research is in steganography. I was familiar with those papers because I’d read them before.

I’m going to assume the savings were accumulated after you got sober? ;-)

Just wanted to say hello to a fellow class of 2014 “graduate”. I failed out of CS @ UIUC in 2003 because I just skipped class and got high most days. Now, after 11+ years of sobriety, I have most of a PhD and I’m teaching CS to undergrads. It’s amazing how much better life turns out when you’re not actively burning everything down in the fires of addiction!

Not trying to be argumentative, but mirror writing is neither encryption nor steganography. The concept is not there.

Can you give a specific example? The term “anamorphic encryption” was used in a paper from EUROCRYPT 2022 and I don’t see any prior formal study of the subject. A follow-up paper from PoPETS 2023 establishes that most of the widely-used cryptosystems introduced over the past few decades are indeed capable of anamorphism but nobody used the term “anamorphic encryption” until the last few years.

To say that they have been used for “hundreds” of years seems like a sweeping generalization in need of better context.

I ran into this a bit while working on my PhD research that used LLMs for steganography. The output had to be deterministic to reverse the encoding, and it was—as long as you used the same hardware. Encoding a message on a PC and then trying to decode on a phone broke everything.

No the hardest problem is teaching CS undergrads. I just started this year (no background in academia, just 75% of a PhD and well-rounded life experience) and I’ve basically torn up the entire curriculum they handed to me and started vibe-teaching.