grayclhn

Sounds like a fun party.

I’ll just point out that the management consultants don’t seem to be the problem in this story…

I prefer the law of evens — every thing that can be counted is an even number, or one away from an even number. Ergo, each year should be two days, and however many rotations of the earth it takes to get there is the only answer that makes sense.

> I for one am just a tiny bit chuffed about that because I'm certain that more and better progress on LLMs (and beyond) would have happened had Lisp been used instead of Python as the primary vehicle for AI exploration.

Yeah… clearly “not using lisp enough” was the major bottleneck in AI development over the last 40 years. /s

(While we’re patting ourselves on the backs for no reason, I’m certain that modern AI is as big of an argument for worse is better as C and Unix ever were.)

If you take an initial draft of a paragraph from someone else’s work and reframe it into your own wording, that is considered plagiarism unless you properly attribute the original source. You should follow the same guidelines for the llm. I agree that you should also make sure you understand and believe what you wrote. None of this is very slippery — understand the ideas you present and don’t pretend to have come up with ideas that originated elsewhere.

This is a little silly — if I quote a large passage from an encyclopedia (or whatever), I cite the source. If I use an LLM to write some of a paper or report, I’m not writing that part (even if I edit it or rephrase it or whatever) so I’d better credit the LLM in whatever way is most consistent with how its output is being presented in the paper.

Too late to edit - should have looked at comment history and not profile I guess, because like an idiot I spent time replying to a clearly ai-written comment… Well GP isn’t going to learn anything from my (gracious! beautiful! dare I say, heroic?) attempt to find a common shared experience in our otherwise separate lives, but hopefully other people got something out of it. I hope to god there’s a lesson there for someone.

Just skimmed your resume and, assuming it’s up to date, it looks like you mostly work independently. If that’s the case, I can totally see why my argument doesn’t resonate.

“The author doesn’t understand the material” isn’t an ad hominem — most of my work involves working with other people, usually less experienced, to help them deliver reasonably complex econometric or ML models. Sometimes I’m hands on, usually not. Knowing if it’s “the author not understanding the material” is important because my job is to help them understand it so that they can deliver something that 1) works and 2) does what it’s supposed to. AI written docs are at best a mixed blessing.

Or maybe it’s me not understanding! Either way, if something looks wrong, I want to discuss it with people who know the subject matter and can talk me through the parts that look off so that I can learn from it.

I can indeed use AI to do research on topics I learn about on HN. That’s a different activity than what I come to HN for, or when I’m reviewing a proposal for something at work, or if I’m interested in an article and want to learn more about some of the details that didn’t make the final draft from either the author or from other people with experience in the area.

“Why didn’t the author consider [this thing that seems similar]” is almost never something you can ask an llm, especially for articles where the author is putting in so little of their own effort that they don’t edit out obvious AI tells.

Oh for sure. My only point is that IME (and I’d assume many other commenters) the quality of discussion and engagement around an imperfect ai-written doc is considerably worse than the same doc written by a human. Work is where it’s most obvious to me… because many of these are design docs for systems I care deeply about us delivering, but I’ve noticed the same thing elsewhere.

I think classifying it as an allergy or a status thing is a little too glib. I’ve read and reviewed, conservatively, hundreds of AI generated documents for work, and “written”/commissioned a bunch too. My biggest issue is that it’s impossible to engage with and give feedback on an AI written document, because it’s impossible to know whether misconceptions or gaps are because the author doesn’t understand the material deeply enough or the author does but the AI doesn’t and the author’s not proofreading carefully enough. Or if a surprising idea is raised — is it the authors insight, can they elaborate on it, where did it come from, etc?

Hackernews isn’t work, obviously, but “it’s impossible to engage deeper with the material because the author doesn’t really exist” is sort of a problem for a discussion site. If the human coauthor puts in enough work, they can make sure the doc really reflects their views and their understanding, but in my experience that’s much less common.

IME a bad abstraction results in the same thing, just with a lot of wasted effort coming up with the abstraction first, and a lot more resistance to fixing it because people are too emotionally invested. I’d rather have something clearly chosen for expedience and that no one likes.

Algebra is the class where I learned I shouldn’t try to figure out how to prove theorems named after people during tests.

And I think you’re underestimating the jump from discrete math and linear algebra to abstract algebra… I think I attended each of those classes and opened their textbooks a total of 3 times each and did fine - once for each exam. But fml abstract algebra and measure theory were rough.

matplotlib has an xkcd style for a different sort of handmade aesthetic. And as reluctant as I am to bring this up as a comment to this post, “plot xyz and make the graph look like it was published in the financial times (but without ripping off their visual brand)” is a remarkably effective prompt after a little tweaking and I imagine something similar would work for other styles.

We sure as fuck developed specialized parts of our hands…

It does. Getting your stomach good at that (and figuring out what food you can personally tolerate best) is a nontrivial part of training for longer ultramarathons.

I was expecting something a bit more comprehensive and damning after the link… but if one of the most objectionable and controversial claims in the book is about adverbs modifying “unique” I think I’ll continue to recommend elements of style.

Climbing gym :)

Wow, it's almost like ADHD is independent of mental health.

When I jump, do I become a satellite?

If the MVP isn't good enough to get useful feedback from customers, you need to either 1) narrow the set of people you consider as early customers or 2) it's not an MVP yet.

IME “bad Bayesian analysis” is when people use plausibly defensible priors to manufacture outcomes… which happens all the fucking time.

That has not been my experience of dating at all fwiw

Dude, if I trusted myself I’d be so fucked.

Some of the things you listed are bad practices, some are bad outcomes. Change the bad practices (no version control, no beta environment, etc) and use them to prioritize and change the bad outcomes over time.

At least half of the stuff you listed will probably never change. Congrats! Being the senior person means becoming comfortable with people making objectively worse decisions than you would, and putting the structure and architecture in place so that it still works anyway. As a bonus, most of those “objectively worse” decisions can be really good and better suited for the team than your decisions would have been ;).

The secret is — when you’re a senior IC, if you think that it’s important to spend 2 days a week coding, you’re expected to tell people to fuck off so you can spend 2 days a week coding.

Not sure that list of example injuries supports the “risks of serious injury or death” claim, but ymmv. If “contusions and abrasions” needs to be spelled out, it seems like someone’s padding the list. :)

Switching AirPods over from phone to watch and back is pretty nice. That’s the only example I can think of offhand though.

I’d love to not have to charge my watch on a 10 day backpacking trip. I’m old enough that most of my formative camping trips were before cell phones, and taking a charger and battery on a hike seems crazy to me. :)

A lot of these questions (not just yours) indicate that bad management is rare… that hasn’t been my experience. :)

I’m sure the impact to the parent commentor could have been mitigated, but absolutely sure that it almost always isn’t.