HNHacker News
TopNewBestAskShowJobs

velcrovan

3,665 karma · joined August 16, 2012

https://joeldueck.com
submissionscomments
velcrovan··on Show HN: ClaudeStatsBar: your session is 486k deep and nothing told you
What's funny is that, being a claude code user, I don't need to give someone else's python package a 24/7 back door into all my activity, I can just make my own.
velcrovan··on Diesel prices in U.S. top $6 a gallon for first time
How does the planet run out of water?
velcrovan··on Claude Fable 5.1 and Claude Mythos 5.1
I have a pet theory that the Opus prose style/smell we all have grown weary of is due at least in part to the models writing more for themselves and each other than for humans. They're packing lots of signal into fewer words and they don't care if it sounds cringe because it works better as glue in long-running tasks.

I'm also thinking of the 2017 novel "Void Star" where AIs who operate everything have long since left ceased bothering with human languages, and it takes a rare sort of direct matrix-gazing savant to be able to try and horse-whisper them into doing or revealing anything they didn't already plan to do.

velcrovan··on Why your local LLM feels dumber than it is
I had heard of this and actually did set the reasoning to medium ahead of time…
velcrovan··on A Friendly Introduction to Racket
https://joeldueck.com/what-about/camp/
velcrovan··on A Friendly Introduction to Racket
Not only do I produce standalone executables in Racket, I cross-compile GUI apps written on MacOS to produce executables for Windows.
velcrovan··on Why your local LLM feels dumber than it is
That's funny, I downloaded the same model on my 48GB M4 Pro and gave it a problem to solve in an existing codebase, it spun its wheels for twenty minutes and then fell over dead. This was using LMStudio and pi as a harness; I never use pi for anything else, so maybe I'm holding it wrong.
velcrovan··on GPT 5.6 Sol is the best "vision" model OpenAI ever released
When asked to produce a thing, the output has a much better baseline of quality. But when you ask it to evaluate the quality of a thing, how do you evaluate the quality of its response, which is necessarily subjective and not quantifiable? How can a thing which has no experience of friction be said to subjectively evaluate quality?

Either you are yourself already a better judge of the thing’s quality, in which case the response can be of no use to you, or you are a poor judge of the thing’s quality, in which case you will be blind to the flaws in the synthesized opinion handed back to you.

velcrovan··on GPT 5.6 Sol is the best "vision" model OpenAI ever released
So, formulaic output…the opposite of taste
velcrovan··on GPT 5.6 Sol is the best "vision" model OpenAI ever released
Assessing the subjective quality of a thing is in my experience one of the worst ways to use any LLM.
velcrovan··on Working with AI Feels More Like Leadership Than Coding
Yes, ‘Sloptraptions’ advertises itself as an AI-assisted newletter.

Whatever your feelings about AI writing (I make a point of not using AI for writing, myself), it's silly to claim Rao is not the one making the point here. The piece links to the transcript of the chat session that generated it, where after having the LLM ingest the paper, he prompts it:

> ‘I want to make the following critical argument: “Have you met the median manager with a few reports or median faculty with a few grad students? Not your work, not your cognitive debt. Otoh, this sort of framing entirely misses the new kinds of work you are doing, which is often illegible or even invisible. A curious factor affecting AI discourse is that most people researching and commenting about the second-order effects have “individual contributor” mindset. Even a modest amount of supervisory experience demonstrates the difference between thinking vs monitoring/supervising, and between being responsible for getting thinking done versus accountable for it. I use AI a lot and it feels indistinguishable from supervisory work to me, with similar cognitive effects. If you manage too much versus doing your own hands on work, you become… a manager. How good you are depends on your aptitude for management work.’

velcrovan··on Working with AI Feels More Like Leadership Than Coding
I first saw this point made by Venkat Rao more than a year ago in "Prompting is Managing" [1]. It was written in response to the "Your brain on ChatGPT" paper that was then making the rounds, about how LLM use is supposedly making people stupid.

Venkat argued that the study subjects did poorly not because LLMs were dulling their minds, but because the study subjects were freshman students with no management skills being tested on tasks that required delegation, quality gating and exception handling. The study put people in a situation that created role confusion and concluded that the poor outcomes were due to "cognitive debt" induced by LLM use.

[1]: https://contraptions.venkateshrao.com/p/prompting-is-managin...

velcrovan··on Anthropic: Introducing The Conceptual Reasoning Index
I enjoy using the models. I also get that there are shenanigans and that marketing is happening, but as long as the models are this effective I can't much bring myself to care. I suspect most of their users are the same.
velcrovan··on Grok 4.6
This is like saying the person you see in the mirror is categorically a human being because both of you produce similar reflections of light rays
velcrovan··on The title cards in Blade Runner are amazing
It’s a bit like how the end point of the most wildly successful music is that you hear it on a loop in the grocery store
velcrovan··on Harness Engineering for Self-Improvement
https://fundthebasilisk.com
velcrovan··on Twenty Years of Pandoc
I remember noting "djot" when John M. created it but hadn't really looked into it before now. After reading through the design goals and differences from Markdown, I can see we'd have been better off if somehow djot had been able to come first. Gruber has said he thinks Markdown being underspecified was a feature, but as a user trying to mix and match different tools it drove me nuts.
velcrovan··on Everyone is building LLM routers, we deprecated ours
Ironically, my confidence that a human had at least an active part in writing/editing this article went up because of this train wreck of a sentence:

> "A cache-aware model router will take that into account by adding stickiness to the initially chosen model and keeps querying it."

velcrovan··on Delayed Gratification – Proud to Be 'Last to Breaking News'
If there were any upsides to brexit I've never heard what they were. (Though, also: not a brit)
velcrovan··on Delayed Gratification – Proud to Be 'Last to Breaking News'
As someone curious about subscribing: why did you end your subscription?
velcrovan··on DMARC has been public since 2012 but most company domains still don't enforce it
I take the opposite approach, I refuse to whitelist domains. When someone internal complains I send a notice to their contact on the other end (CCing the internal recipient) saying their email is misconfigured and ask them to put me in touch with their IT department to help them fix it. I use a script to do some DNS lookups and write the email for me. I have about a 50% success rate getting them to fix it.
velcrovan··on Fonts In Use – Find out where a font is used
I have this site’s rss feed in my feed reader. I definitely don’t try to keep up with every post but once a week or so I click through a dozen or so recent posts.
velcrovan··on Claude Voice Mode
No mention of using this with CarPlay...
velcrovan··on 10 REM"_(C2SLFF4
And in assembler and machine code before that!

I'm persuaded that just about every language is "homoiconic" and that _bicameral syntax_ is the actually-interesting thing about Lisps : https://parentheticallyspeaking.org/articles/bicameral-not-h...

velcrovan··on Minecraft: Java Edition now uses SDL3
Anyone got any advice for a techy dad with no (zero) Minecraft experience who wants to set up a Minecraft server for the family in 2026? Kids right now are playing on their iPads, and (sometimes) old macbooks/windows PCs.
velcrovan··on Bryan Johnson: I have an autoimmune disease. My stomach is eating itself
"he's rich therefore he has no incentive to scam anyone" is not a good heuristic, though it will help you get a degree from Trump University.
velcrovan··on Moby Dick Workout (2022)
Moby Dick has a bad (high) ratio of “people who try to read it like a thriller novel” to “people who read it for what it is”.
velcrovan··on Codeberg Is Down
Again, Codeberg isn’t obscure. Worrying about front page HN links “confusing” others because you haven’t heard of something is about the silliest thing I have read on this site.
velcrovan··on Codeberg Is Down
Consider that maybe you haven’t heard of all the things that HN readers find interesting.
velcrovan··on Rhombus Language 1.0
> When considering the reasons for a language, look at the negation. Nobody writes a language to be "unapproachable"… Nobody really wants to stick "non-extensible" on their language…

And yet, unapproachable languages and non-extensible languages exist. More specifically, there are languages where approachability and extensibility formed no part of the design goals and it shows.

← PreviousPage 2 of 28Next →