Hacker News: austinjp

New comment by austinjp in "Claude for Legal"

austinjp — Sat, 16 May 2026 22:52:21 +0000

Find people in her field who are already hopping on the wave. Better still, gently persuade her to find those people. She'll probably know where to look. In fact, I'd be surprised if her professional newsletters aren't already starting to surface those folks.

New comment by austinjp in "Show HN: Explore color palettes inspired by 3000 master painter artworks"

austinjp — Wed, 06 May 2026 01:27:41 +0000

So... ummm... the website and all the submitter's comments here seem very Claude generated, no?

Spam filters are going to have to get a lot more sophisticated. "Slop" filters, even.

New comment by austinjp in "Palantir employees are starting to wonder if they're the bad guys"

austinjp — Thu, 23 Apr 2026 22:35:14 +0000

I'm unsure of how this information is being presented. But it's entirely possible for the majority of people on Earth to avoid all those things. And it's entirely possible for many people who are (perhaps unwittingly) funding U.S. defense companies to stop doing so.

New comment by austinjp in "Newly created Polymarket accounts win big on well-timed Iran ceasefire bets"

austinjp — Thu, 09 Apr 2026 13:14:37 +0000

> It's already illegal to threaten journalists [in] America

Not necessarily, it depends on the pressure and the intent.

> In America we generally make bad things illegal, not activities that could become motivation for bad things.

I didn't mention making anything illegal. I suggested constraining Polymarket and similar.

> Should we ban [X]

I didn't mention banning anything.

New comment by austinjp in "Newly created Polymarket accounts win big on well-timed Iran ceasefire bets"

austinjp — Thu, 09 Apr 2026 09:18:11 +0000

Polymarket gamblers have pressured at least one journalist regarding reporting of missile strikes. This requires regulation or, as others here have suggested, non-anonymity, maybe other measures too.

https://www.theguardian.com/world/2026/mar/18/polymarket-gam...

New comment by austinjp in "Ask HN: What are you building that's not AI related?"

austinjp — Thu, 09 Apr 2026 08:43:30 +0000

A crossword puzzle generator, just for fun. Grid generator in Python because it's easy to hack around and grid generation doesn't need to be particularly fast. Go for populating the grid with words because with large grids there are combinatorial explosions, and Go's speed is beneficial.

Not source-available yet because it's a bunch of hacks (particularly the Python) but maybe one day.

New comment by austinjp in "System Card: Claude Mythos Preview [pdf]"

austinjp — Wed, 08 Apr 2026 08:28:23 +0000

It very much depends on the crime. The truly awful stuff is committed by intelligent people.

New comment by austinjp in "System Card: Claude Mythos Preview [pdf]"

austinjp — Wed, 08 Apr 2026 08:19:38 +0000

And there's always a product provider who's willing to add that flag, despite all the warnings.

New comment by austinjp in "The future of code search is not regex – 100x faster than ripgrep"

austinjp — Thu, 02 Apr 2026 09:23:57 +0000

To save people the digging, here's the git repo:

https://github.com/dmtrKovalenko/fff.nvim

"FFF stands for freakin fast fuzzy file finder (pick 3) and it is an opinionated fuzzy file picker for your AI agent and Neovim. Just for file search, but we do the file search really fff well.

FFF is a tool for grepping, fuzzy file matching, globbing, and multigrepping with a strong focus on performance and useful search results. For humans - provides an unbelievable typo-resistant experience, for AI agents - implements the fastest file search with additional free memory suggesting the best search results based on various factors like frecency, git status, file size, definition matches, and more."

New comment by austinjp in "Ask HN: Who wants to be hired? (April 2026)"

austinjp — Thu, 02 Apr 2026 01:17:15 +0000

Kudos for listing the things you're *not* willing to work on, and for those things in particular.

New comment by austinjp in "I'm betting on ATProto"

austinjp — Tue, 31 Mar 2026 08:59:36 +0000

Erm.... you're not exactly proving them wrong are you.

New comment by austinjp in "Ask HN: Are you too getting addicted to the dev workflow of coding with agents?"

austinjp — Tue, 31 Mar 2026 08:18:24 +0000

I feel this sneaking up on me. I've only recently allowed Claude to actually edit some files directly, rather than just show me suggested edits. It could certainly be addictive to just hit enter while code magically appears, thinking "oh yeah, I totally would have done it like that".

New comment by austinjp in "Ask HN: Are you too getting addicted to the dev workflow of coding with agents?"

austinjp — Tue, 31 Mar 2026 08:08:09 +0000

> rapidly leading to task paralysis with the sheer scale of the plan.

Yikes. I feel seen.

New comment by austinjp in "Alzheimer's disease mortality among taxi and ambulance drivers (2024)"

austinjp — Sun, 29 Mar 2026 10:58:07 +0000

So let's say there's a causative link (see other comments here for why this may not be the case), it would take a lifetime of daily complex spatial navigation for several hours every day to significantly reduce Alzheimer's disease risk, and it's still not guaranteed. If there's a linear dose response (a big if) it would still require hours per week for decades for a more modest impact.

That seems unattainable for anyone at all.

Man, Alzheimer's disease sucks. We need more investment and more research into this horrible illness.

Personally I'm curious about the impact of super-early diagnosis, decades before symptoms, and interventions that maximally slow progress.

New comment by austinjp in "False claims in a widely-cited paper"

austinjp — Thu, 26 Mar 2026 01:46:35 +0000

Oh I'm sure the grifters will find ways in. The other disciplines may have provided a "moat" for the past few decades, but it won't last forever.

New comment by austinjp in "Show HN: Email.md – Markdown to responsive, email-safe HTML"

austinjp — Wed, 25 Mar 2026 10:42:01 +0000

I'm not so sure. It's definitely the de facto standard, but I suspect minimal HTML is better. Just enough tags to add structure and meaning (H1-H6, p, a, em, section for structure including nesting, maybe more). LLMs were trained on a lot of HTML, they're good at processing it. HTML requires more tokens than markdown but I believe it's worth it. I'll find out in a few weeks as I experiment with both.

New comment by austinjp in "Show HN: Email.md – Markdown to responsive, email-safe HTML"

austinjp — Wed, 25 Mar 2026 10:32:37 +0000

Yep both are widely used. I forget which markdowns extensions they originated from. The Pandoc website probably has the details.

New comment by austinjp in "Show HN: Email.md – Markdown to responsive, email-safe HTML"

austinjp — Wed, 25 Mar 2026 10:31:22 +0000

It uses MJML under the hood.

https://mjml.io/

New comment by austinjp in "Ask HN: Is using AI tooling for a PhD literature review dishonest?"

austinjp — Tue, 24 Mar 2026 09:21:13 +0000

Whether you can or can't in reality is moot, unfortunately. The literature search in biomedical fields should indeed be theoretically reproducible. I don't know about other fields, but it would seem odd to me if a search was not reproducible, that would lead to a very arbitrary literature selection.

As for the experiments, yes, in experimental fields. But in all (most?) fields, including non-experimental, the whole process should be well documented so it could be reproduced end-to-end if possible. If it's not reproducible there should be good, well explained reasons why not.

Note that reproduciblity does not necessarily mean the exact same answer will definitely emerge, just that the methods can be followed closely.

New comment by austinjp in "Ask HN: Is using AI tooling for a PhD literature review dishonest?"

austinjp — Tue, 24 Mar 2026 09:16:17 +0000

Quite probably there are differences between fields. In biomedical literature reviews the search terms and databases are detailed, and (in systematic reviews) a PRISMA flowchart [0] provided. The theory being that other researchers could repeat the searches and the in/out decisions and get the same stack of papers to review.

[0] https://www.prisma-statement.org/prisma-2020-flow-diagram