Hacker News: jghn

New comment by jghn in "Leaded gas was a known poison the day it was invented (2016)"

jghn — Sat, 11 Jul 2026 21:30:27 +0000

Even then, it's overblown. The fibers embedded in the cement siding that was common for a while as an example, that's just never going to be friable to a degree that matters. Not to mention the issue is really more of a long term exposure thing. You'd need to be quite an unlucky person to take one whiff of some asbestos fibers and get mesothelioma decades later.

New comment by jghn in "Leaded gas was a known poison the day it was invented (2016)"

jghn — Sat, 11 Jul 2026 19:14:36 +0000

Asbestos gets a bit of a bad rap. There's an enormous difference from e.g. asbestos tiles and siding in a house vs loose clumps of insulation at mass scale in large factories and such.

But at some point people decided that any asbestos was an immediate ticket to mesothelioma and had to eradicate it altogether.

New comment by jghn in "Apple Silicon Exec Explains Mac Mini AI Demand and On-Device Future"

jghn — Sat, 11 Jul 2026 16:53:56 +0000

Interesting. For me it seems to be much more common that even basic requests start spinning up sub-agents. But I also routinely ask it to do this so maybe it's stored this as a preference somewhere deep in its bowels.

And yes, I agree. I find the experience better less for speed and more for context management. But it's far from necessary.

New comment by jghn in "Apple Silicon Exec Explains Mac Mini AI Demand and On-Device Future"

jghn — Sat, 11 Jul 2026 15:51:43 +0000

What about when claude code starts spawning sub-agents? it's pretty good about doing that now even if a user doesn't request it. Does this could as parallel inference?

New comment by jghn in "AI changes the economics of software rewrites"

jghn — Thu, 09 Jul 2026 17:56:04 +0000

> What's the point of the rewrite if it doesn't fix the underlying issues, though?

Depends on what you mean by underlying issues. If you're in a regulated environment, it may be such a mountain of red tape to change behavior that it's not worth it, even if you know it's not ideal.

But if the underlying issues are tech debt, bad design, and other things invisible to the outside world, that's different.

New comment by jghn in "GPT-5.6"

jghn — Thu, 09 Jul 2026 17:48:39 +0000

I have found Claude Code to be so much better than other common harnesses that it's kept me solely in the Anthropic ecosystem.

New comment by jghn in "Postgres rewritten in Rust, now passing 100% of the Postgres regression tests"

jghn — Thu, 09 Jul 2026 12:40:39 +0000

Perhaps before embarking on one of these rewrites the first step should be a heavy round of mutation testing and property based testing. Contribute any new testing code from this back to the original project. And *then* embark on the rewrite.

New comment by jghn in "Introduction to Genomics for Engineers"

jghn — Mon, 06 Jul 2026 12:51:14 +0000

All one needs to do is look at the Claude Science thread here last week and note how many comments were surprised that it appeared to be a statistical/analysis tool.

New comment by jghn in "What ORMs have taught me: just learn SQL (2014)"

jghn — Sat, 04 Jul 2026 17:33:35 +0000

> built one for Java, in the early 90s

So was your ORM for Oak? Java didn't hit the public sphere until 1995 IIRC

New comment by jghn in "What ORMs have taught me: just learn SQL (2014)"

jghn — Sat, 04 Jul 2026 16:47:03 +0000

The second issue is, if these queries are just "select * from table where id in ...", WTF bother with a library to abstract that away in the first place? It's trivially easy to handle this as SQL

New comment by jghn in "What ORMs have taught me: just learn SQL (2014)"

jghn — Sat, 04 Jul 2026 16:43:19 +0000

The problem is that "ORM" does a lot of heavy lifting as a term and can mean different things to different people. Like yes, obviously, one needs some sort of SQL -> data structure transition on the boundary (using "object" overfits to OOP!). But that can be extremely light weight. Let people write SQL, have a thin layer to pull the results back out into the appropriate data structures, and move on.

New comment by jghn in "What ORMs have taught me: just learn SQL (2014)"

jghn — Sat, 04 Jul 2026 16:15:58 +0000

To make matters worse, most of the time I've successfully argued a project to just use SQL instead of an ORM, what has happened is that people over time built a home rolled ORM in the development language.

It's like people can't just let go.

New comment by jghn in "CarPlay Is Additive"

jghn — Fri, 03 Jul 2026 17:38:05 +0000

I have a mid-aughts VW and have never had any of these issues. So I suppose you can count mid-aughts VWs.

New comment by jghn in "Superpowers 6"

jghn — Fri, 03 Jul 2026 04:02:08 +0000

What I like about superpowers is that for my workflow, I spend most of my time brainstorming with the claude session. Using its brainstorm mode helps to keep it from shifting to just writing code. That's basically never what I want until I actually want it. Once we've locked in a design, whatever, I don't care.

But when I don't trigger brainstorm mode, even using the built in plan mode, it's just never as in depth of a brainstorm partner for me.

New comment by jghn in "CarPlay Is Additive"

jghn — Fri, 03 Jul 2026 03:43:08 +0000

I'm in this camp: I will not buy a car without CarPlay. And I put so few miles on my car that while I'd like a new one, if the vendors make this impossible then no one gets my money.

New comment by jghn in "Ask HN: Who is quitting? (July 2026)"

jghn — Thu, 02 Jul 2026 12:58:23 +0000

I started my career in the mid-90s and 4GLs making all of us obsolete was already a thing by then!

New comment by jghn in "Senior SWE-Bench: open-source benchmark that assesses agents as senior engineers"

jghn — Thu, 02 Jul 2026 12:28:29 +0000

I'd take this a step further, but that step also curls back to the other side a small bit.

The real skill is being able to both pull the necessary information from these sources as well as being able to intuit gaps in that knowledge based on their understanding of the business and their domain expertise & wisdom. Sometimes you can't get a perfect picture, sometimes the people who should know aren't able to tell you what they really need. You still need to do the right thing.

A benchmark like this can potentially do the second part. But I don't think any model would be good at it, for now.

New comment by jghn in "Bring back crappy forums"

jghn — Thu, 02 Jul 2026 12:23:27 +0000

This. I only dealt with a non-threaded take on usenet for about 1 week in the early 90s.

New comment by jghn in "Meta caps internal AI token spending"

jghn — Thu, 02 Jul 2026 01:12:27 +0000

> Who could possibly have predicted that happening?

Charles Goodhart :-)

New comment by jghn in "Most rewrites serve the engineer, not the business"

jghn — Wed, 01 Jul 2026 21:16:08 +0000

Iff the engineer's incentives are aligned with the businesses. Which is far from always true.