HNHacker News
TopNewBestAskShowJobs

lordgrenville

2,292 karma · joined December 1, 2015

<this user>.github.io
submissionscomments
lordgrenville··on LLMs Are Bad Judges. So Use Our Classifier Instead
True, but they own that:

> Declaration of Interest: [Authors] have financial interests in...Arbitrus.ai. As the title would suggest, the authors are making no effort to obfuscate this fact.

lordgrenville··on Solving a wooden puzzle using Haskell
I think in English one would usually say "position" where the author uses "disposition". (Disposition is closer to "temperament"/"personality".)
lordgrenville··on Show HN: I reverse engineered macOS to allow custom Lock Screen wallpapers
Is this inspired by that viral Sabrina Carpenter lock screen video?
lordgrenville··on Why our website looks like an operating system
More correctly, "our website looks like a desktop environment".
lordgrenville··on A staff engineer's journey with Claude Code
Yes! Reminds me of one of my all-time favourite HN comments https://news.ycombinator.com/item?id=23003595
lordgrenville··on Claude Code IDE integration for Emacs
Well it is already a "graphical" editor. And I would say out of the box it isn't an IDE, but with a couple of common packages added (say a file tree, terminal, code completion, LSP) it looks pretty much identical to any commercial IDE.
lordgrenville··on I watched Gemini CLI hallucinate and delete my files
> I think I'm ready to open my wallet for that Claude subscription for now. I'm happy to pay for an AI that doesn't accidentally delete my files

Why does the author feel confident that Claude won't do this?

lordgrenville··on Show HN: A 'Choose Your Own Adventure' written in Emacs Org Mode
I thought this was going to be a game engine written in Emacs Lisp :) but this is still cool. Congrats!
lordgrenville··on I extracted the safety filters from Apple Intelligence models
I wonder if they used an LLM to generate the list of safety terms.
lordgrenville··on One of the Most Popular Games on the Planet
Here is the report by (the late) Hindenburg Research about dangers to kids on Roblox. Can't vouch for the contents, but I think parents should at least skim it and take precautions. https://hindenburgresearch.com/roblox
lordgrenville··on Read the Obits
Ann Wroe is a treasure
lordgrenville··on Advanced Python Features
And you can use * for the reverse (every parameter from here on needs to be keyword-only).

https://docs.python.org/3.12/reference/compound_stmts.html#f...

lordgrenville··on Deafening Silence from the Cybersecurity Industry
Great comment, articulates something I've been feeling lately but didn't quite have the words for. (Not American, but similar situation in my country.)

Where do we go from here? What kind of action would be effective?

lordgrenville··on Recent AI model progress feels mostly like bullshit
I am hesitant to correct a math Olympian, but don't you mean the median?
lordgrenville··on Alice Hamilton waged a one-woman campaign to get the lead out of everything
My pet theory is the proliferation of digital entertainment options, and gaming / online culture. There are a lot more ways to keep yourself occupied inside, so less reason to go out and get into trouble.
lordgrenville··on Have you ever seen a goth downtown?
Disagree, because I think everyone's conforming. No psychologically healthy human being just acts out their authentic self without any calibration with some type of peer group. Whether that peer group is NYU undergrads, or their fellow misfits in Albany / pump.fun, is just a difference in degree.

Sometimes the peer group can be remote. The only emo kid in Baghdad is a unique snowflake relative to other Baghdadis, but is following a fashion code vetted by people on Tumblr.

lordgrenville··on Shunpo: Minimalist bash tool to make directory navigation a little bit faster
I have something similar (though less elegant). I find the best use case to be hitting ..........<RET> to get to root (in conjunction with zsh auto_cd option), rather than typing `cd /`.
lordgrenville··on I am offering mentoring: In which I help you
So many anonymous writers on Substack who seem to take pride in giving zero biographical information. Then they write geopolitical analysis or whatever with enormous confidence, frequently calling out the experts in the field as idiots.

I guess they'd argue that the author's identity shouldn't matter and that one should just judge the writing on its own merit. The problem is that we're drowning in information. AI makes its trivial to generate authoritative-looking nonsense. Evaluation is costly. So I prefer to read something which the author has associated with their own identity (which means there's a cost to lying), or has provided some credentials to prove why their opinion matters. Of course there are exceptions (Gwern comes to mind), but I'll only trust them if they're recommended by other writers I trust.

lordgrenville··on NYC Congestion Pricing Tracker
The VC-funded startup model is predicated on becoming a large business, not staying a small one.
lordgrenville··on Popeye and Tintin enter the public domain in 2025 along with Faulkner, Hemingway
Great news. The copyright owners of Tintin are very litigious [0].

I have some "Tintin in Vietnam" prints at home that are commonly sold in SE Asia [1]. Those street vendors will be happy to know that they're in the clear now.

[0] See https://news.ycombinator.com/item?id=39625647

[1] eg https://old.reddit.com/r/TheAdventuresofTintin/comments/c0pv...

lordgrenville··on New York Times Tech Guild goes on strike
And median wage in NYC is $74k (according to Google). Sure, Manhattan is different, tech salaries are different, etc. I'm not claiming that these specific workers should/shouldn't be paid more, just that it's really tone-deaf to claim that you can't live on <$100k, when more than half of New Yorkers do.
lordgrenville··on The unnecessary decline of U.S. numerical weather prediction
That's crazy. Why would the number of hours worked per week ever possibly be relevant to a hiring process? At most they might ask "full-time/part-time/other" or something like that.
lordgrenville··on USGS uses machine learning to show large lithium potential in Arkansas
So it turns out that there's no theoretical reason that gradient boosting will always outperform RF (which would violate the "no free lunch" theorem). But it does usually seem to be the case in practice, even with small and noisy data.

I would hazard a guess that with better tuning, XGBoost would still have won. (The paper notes that the authors chose a suboptimal set of hyperparameters out of fear of overfitting - maybe the same logic justifies choosing a suboptimal model type...)

lordgrenville··on Large language models reduce public knowledge sharing on online Q&A platforms
One relevant difference is that a better-quality human answer is correlated with certain "tells": correct formatting and grammar, longer answers, higher reputation. An incorrect LLM answer looks (from the outside) exactly the same as a correct answer.
lordgrenville··on Large language models reduce public knowledge sharing on online Q&A platforms
The answer by humans is a) publicly accessible b) hallucination-free (although it still may not be correct) c) subject to a voting process which gives a good signal of how much we should trust it.

Which makes me think, maybe a good move for Stack Overflow (which does not allow the submission of LLM-generated answers, wisely imo) would be to add an AI agent that would suggest an answer for each question, that people could vote on. That way you can elicit human and machine answers, and still have the verification process.

lordgrenville··on Map with public fruit trees
Someone brought this up in a child comment: what are the taste/health effects of near-permanent car traffic on fruit trees? The fruit of the public citrus trees in my city is pretty sour, but I don't know how much this is because of not being bred for sweetness, vs absorbing carbon monoxide, lead etc from passing cars.
lordgrenville··on What's inside the QR code menu at this cafe?
The author says "I refuse to believe they’re unaware of this. This doesn’t feel like an oversight, it's either a deliberate design decision or they just don't care." Agree that this is an uncharitable way of looking at it.
lordgrenville··on LinkedIn is now using everyone's content to train their AI tool
Not commenting specifically on LI ranking, but Byrne Hobart wrote a terrific rebuttal of the main premise of this book https://www.thediff.co/archive/bullshit-jobs-is-a-terrible-c...
lordgrenville··on Sapling: Source control that's user-friendly and scalable (2022)
It also conflicts with the PowerShell built-in Set-Location (though I imagine that's a pretty small intersection of users...)
lordgrenville··on Greppability is an underrated code metric
related to "grok"?
← PreviousPage 3 of 22Next →