Hacker News: jszymborski

New comment by jszymborski in "Running local models is good now"

jszymborski — Tue, 16 Jun 2026 17:55:14 +0000

I run local models and they work fine for me, but specifically for use in coding harnesses, I'm having a hard time. Tools tend to end up in the same loop, trying to `ls` the same folder or `grep` the same file, over and over and eating up the whole context. Super hard to get it to do anything but that. Any tips?

New comment by jszymborski in "Correlated randomness in Slay the Spire 2"

jszymborski — Tue, 16 Jun 2026 13:51:53 +0000

> The phenomenon of "correlated RNG" (or "CRNG")

This is a pretty funny abbreviation since CRNG is sometimes "cryptographic random number generator", which would not be susceptible to this correlation. Albeit I think CSRNG is more common.

New comment by jszymborski in "CrankGPT"

jszymborski — Mon, 15 Jun 2026 18:07:25 +0000

Totally true that what is and is not sustainable is more complicated than it first appears.

I think what I'm honing in on is the idea that hand cranks produce very limited, often interrupted power and are relative low-tech, both of which are directionally the right way for us to be putting our efforts.

New comment by jszymborski in "CrankGPT"

jszymborski — Mon, 15 Jun 2026 15:58:43 +0000

There's a bit of a difference between population control and reaching a sustainable equilibrium. One can also argue the death of all life on earth is a pretty evil logical conclusion of infinite growth on a finite planet.

New comment by jszymborski in "CrankGPT"

jszymborski — Mon, 15 Jun 2026 15:30:38 +0000

As a bit of an aside, I really like the idea of trying to design things with the constraint of it having to be able to run off a hand-crank.

I feel like it is not only an interesting engineering challenge but one that might lead to a more efficient and sustainable framing.

New comment by jszymborski in "Successful Psilocybin Treatment of Alzheimer"

jszymborski — Mon, 15 Jun 2026 15:08:52 +0000

That's totally fair, and I'm often reminded not to paint all "Frontiers" journals with the same brush (although it's difficult for me sometimes).

New comment by jszymborski in "Improvement in advanced Alzheimer’s disease following high-dose psilocybin"

jszymborski — Mon, 15 Jun 2026 14:23:31 +0000

So, I'm not making any accusations or decelerations of "bad science", I've only read the abstract and this isn't my field.

That said, some relevant context here is that:

(1) Case studies are some of the most easily fabricated journal outputs

(2) This is published in Frontiers in Neuroscience, which is listed by some as being a predatory journal [0]. The Frontiers publishers are the fine folk who published an AI generated anatomical figure of a rat that not only was obviously incorrect to anyone you'd stop on the street, it'd give them nightmares [1].

So I'm not saying this paper is bunk, but that I reserve a healthy degree of skepticism pending some clinical trials or replication in animal models.

[0] https://www.predatoryjournals.org/news/list-of-all-frontiers...

[1] https://arstechnica.com/science/2024/02/scientists-aghast-at...

New comment by jszymborski in "Petition to Withdraw Canada's Bill C-22"

jszymborski — Fri, 12 Jun 2026 03:06:12 +0000

Under first-last-the-post, we actually don't :P

New comment by jszymborski in "Pozzo: A Fast Lucky Number Checker"

jszymborski — Thu, 11 Jun 2026 22:06:33 +0000

I'd have voted for it as a judge! I stopped participating in hackathons because they became just about integrating sponsor APIs and half-baked start-up ideas based on quarter-baked demos... All entirely corporate and loveless.

This, however, is a return to the weird shit, and I embrace it!

New comment by jszymborski in "MapComplete – Contibute to OpenStreetMaps"

jszymborski — Thu, 11 Jun 2026 15:09:06 +0000

I think we're hugging it to death.

New comment by jszymborski in "1worldflag: A blue dot on a transparent background"

jszymborski — Mon, 08 Jun 2026 22:55:22 +0000

I guess we'll just have to agree to disagree!

New comment by jszymborski in "1worldflag: A blue dot on a transparent background"

jszymborski — Mon, 08 Jun 2026 14:53:29 +0000

Wouldn't the Esperanto of Flags be the flag of Esperanto?

New comment by jszymborski in "1worldflag: A blue dot on a transparent background"

jszymborski — Mon, 08 Jun 2026 04:51:00 +0000

Esperanto actually already has a flag

https://en.wikipedia.org/wiki/Esperanto_symbols#/media/File%...

New comment by jszymborski in "Social Cache Busting"

jszymborski — Sat, 06 Jun 2026 14:43:07 +0000

> But if you’re talking to a performer, and they have a fake, glassy-eyed smile, and go through all the correct motions, while obviously being totally checked out, you’re not asking the right questions.

Or it's an indication you are asking the correct questions but that the person you're asking it from anticipated it and is evading. Anyone who has heard a politicians and CEOs take questions from the press know this.

New comment by jszymborski in "Ask HN: What was your "oh shit" moment with GenAI?"

jszymborski — Fri, 05 Jun 2026 21:25:03 +0000

There was a viral Medium post that was about LLMs but then there was a reveal at the end was that the whole thing was a ChatGPT post. That was my first "wow" moment.

It was on hackernews... anyone know what I'm talking about?

New comment by jszymborski in "Neocities domain suspended by Namecheap for unrelated court case"

jszymborski — Fri, 05 Jun 2026 20:59:03 +0000

and later in that thread https://bsky.app/profile/neocities.org/post/3mnkqgxt3kn2k.

The post acknowledging the site is up was posted 7 mins ago though :)

New comment by jszymborski in "Neocities domain suspended by Namecheap for unrelated court case"

jszymborski — Fri, 05 Jun 2026 20:07:44 +0000

https://neocities.org seems to up right now...

New comment by jszymborski in "Rootshell: A new E2EE email service hosted in Iceland"

jszymborski — Thu, 04 Jun 2026 00:12:16 +0000

There's a lot of stuff that looks officious enough that will trick folks, especially those distracted or not well-versed in the attack vector.

New comment by jszymborski in "Rootshell: A new E2EE email service hosted in Iceland"

jszymborski — Wed, 03 Jun 2026 22:43:16 +0000

I hate shoving LLMs everywhere, but honestly this is probably a good use case for tiny models like the 0.6B Qwen model to flag account names for human review.

New comment by jszymborski in "I was recently diagnosed with anti-NMDA receptor encephalitis"

jszymborski — Wed, 03 Jun 2026 16:49:04 +0000

I can't begin to imagine the pain and stress caused by those symptoms, but I am so very happy to hear the prognosis is quite good. A linked scientific article makes the case that this is a very new diagnosis, and increased awareness might help a lot of people, so if you are reading this burntsushi, thanks for being open about something deeply personal, I hope I would have had the courage to do so myself.