Hacker News Reader

dnauticson Singapore govt dating app uses Gale-Shapley stable marriage algorithm

Which side do they screw over (the choosers get their worst possible match and the bidders get their best possible match)

dnauticson Biology might not be quantum, but its math is quantumlike

As a working scientist: chasing beauty is not a good way to do science.

> It leads to improved decision making, experimental design, etc

It does not. It often does the exact opposite.

dnauticson C's Flexible Integer Sizes Were Not a Design Mistake

That is way too much to remember

dnauticson Biology might not be quantum, but its math is quantumlike

That's not correct. Lock and key absolutely does take into account em and hydrophobic interactions. Also, molecular dynamics is nearly useless at modeling these things (which is why heuristic models such as alphafold, Rosetta, dominate)

dnauticson Biology might not be quantum, but its math is quantumlike

What are you talking about. The phenomena are adequately explained. I can measure a Kd and make mathematically modeled predictions that will come true. I can even phenomenologically test contributions to the Kd (ablate hydrogen bond, delete or add a charge, etc)

Nature is under no obligation to make explanations trivial to a human brain conditioned on quotidian macroscopic observation. Doesn't mean you have to appeal to quantum woo. We know more or less how much "quantum mechanics" (for some definition of QM, obviously an electron shell is QM, but for all intents and purposes you can just treat it as a classical ball that does a few weird things like bonding) contributes, to, say reaction rates. It's nonzero. It's nearly zero, though.

dnauticson Biology might not be quantum, but its math is quantumlike

> If you took a classical bag and filled it with classical locks and classical keys and just shook it around for a while, none of those keys would end up in the locks

I think you are vastly underestimating the number of collisions required to get an enzyme binding event. We did a back of the envelope calculation in grad school and it was something like >> 10^6 ~ 10^9.

And you can of course do macroscopic things like this:

https://www.youtube.com/watch?v=3X6qEE2fHvE

dnauticson Does Georgism work? Five years later

About half might. But then another half will have dealt with actually moving a confused, possibly suffering from dementia, elderly person, managing all of the affairs (and forgetting to do some, changing health care insurance), creating an unnecessary multiyear fuck you in your life while trying to manage a job and a family, and the policy will get torched.

Allowing the elderly to stay where they are isn't just about protecting vulnerable boomers, it's about not being a problem for responsible people in the next generation. You could also be one of those no-contact assholes and never have to deal with your parents when they get screwed by displacement, land value tax away.

dnauticson Gravity seems holographic. What does that mean for reality?

I am not a physicist but i presume that is the hope of some who pursue holographic modes

dnauticson ASML says it sold 'absolutely nothing' in Europe in 2026

it's not like bridges don't collapse in china. There were two major ones last year.

dnauticson We're gonna need a lot more mathematicians

> doesn't convincingly justify why, in my opinion.

I'm convinced this agency argument is correct [for the next N months]. But yeah, it's vibes. And you could probably create a reasonable proxy measure for this.

So I wouldn't call his argument unconvincing, I would call it unformalized. In order to walk this world you're gonna have to contend with some informal arguments that are powerful, correct, and should be convincing.

dnauticson Gravity seems holographic. What does that mean for reality?

Yeah because you can prove that any such bijection cannot be a topological homeomorphism, for example. So necessarily some "nice to have" properties must drop out.

dnauticson Mercury 2.5 LLM hits 770 tokens per second

You can upload different weights and even do LoRAs. The chip architecture is interesting, the first (n) layers are the sane, so you can change architecture by adding (m) layers. Plausible that this is sufficiently flexible enough for several generations of real world applications. For example, we still use 45nm general purpose silicon for automotive, e.g.

dnauticson Claude discovers a novel enzyme system with CRISPR-like repeats

I guess the Poincare conjecture or the theory of relativity will continue to have "little value" until they get published in a peer reviewed journal

dnauticson The Millennium Problems for Biology

There are bacteria which naturally use wires. For example to strip electrons off of metal nodules (the videos are amazing, you can see the bacteria attach to the nodule, the metal shrinks, and you see the Schlieren effect of the high concentration of dissolved ions).

AIUI the biggest obstacle is likely the fact that bacteria want to operate in a high current, low voltage regime and everything we do with bulk power wants to operate in a low current high voltage regime.

dnauticson GPT-6 Astra has gained the ability to drive a car

There's also token RTT on top of network latency.. but what if you had a model running at 10k tps (like taalas' llama3b-8

dnauticson MCP was always a bad idea?

I'm not. You can build an MCP and throw a ton of junk in there and bloat the shit out of its token cost. So, really, "it depends".

dnauticson AI Has No Wisdom and Neither Will You

It's a mix. There are some projects where I let the LLM mostly run free and I focus on architecture. Even through some scary parts. Recently I picked a parallelization scheme for this and let the LLM cook. I have no idea how it implemented it, except that it (supposedly) "does it the way I asked:

https://github.com/ityonemo/bpa

For my projects that I use day to day in prod there is considerably more hand-holding and code review.

dnauticson AI Has No Wisdom and Neither Will You

Its not a rubber duck anymore when it talks back to you.

dnauticson AI Has No Wisdom and Neither Will You

Hard disagree. Usually I am wiser than the AIs but there have been times where the AI has pushed back and made me see the light on some poor design I was about to pursue

dnauticson MCP was always a bad idea?

yeah MCP contains a provision for some blocktext for instructions. If your MCP has walls of text in those instructions, it will burn more tokens that a terse MCP.

dnauticson MCP was always a bad idea?

yes and using that tool does what with tokens?

dnauticson Show HN: Mini-AGI – Dynamic continual learning model trained on 8GB VRAM

I love the approach of this but "It has to not forget. A model that learns continually and overwrites itself is worse than one that does not learn at all."

Is highly misguided.

While the platonic ideal of Lt Commander Data is appealing, The parable of funes the memorious (Jorge Luis Borges) comes to mind.

dnauticson What Zig felt like, coming from Rust

Maybe the authors just don't want to do it?

Anyways: its possible, I am building it as a very side project.

github.com/ityonemo/clr

dnauticson MCP was always a bad idea?

The other problem with MCPs is that the harnesses don't auto-reconnect if you've enabled access. Connection status should be persistent

dnauticson MCP was always a bad idea?

How do you think the LLM becomes aware of which mcps are available? Vibes?

dnauticson The Millennium Problems for Biology

Electrofixation of carbon. Build a bug with one of the alternate (non rubisco) pathways and pump electrons from the mains to turn CO2 into sugar

dnauticson The Millennium Problems for Biology

"how was this built from these lego blocks"

That is not engineering. Nobody would call astronomy or cosmology engineering and it is very much the study of "how the universe is built from some basic building blocks.

Don't assign a reductive meaning to engineering that is so encompassing that everything everywhere is engineering, else you render "engineering" a meaningless concept

dnauticson What Zig felt like, coming from Rust

In practice as long as you stick to using zig std, an analyzer can get quite far

dnauticson The Millennium Problems for Biology

Its not unsolved. It's very solved. It was not an engineering question.

dnauticson The Millennium Problems for Biology

Chemiosmotic effect