HNHacker News
TopNewBestAskShowJobs

HappMacDonald

565 karma · joined February 22, 2013

submissionscomments
HappMacDonald··on I Became a Citizen of Japan
Yeah, my understanding in Japan, South Korea, and most countries with "fertility problems" is that they have almost zero correlation with any personal human's "fertility problems": they instead have to do with too many humans simply choosing not to reproduce (and for shockingly many of them, choosing to avoid romantic partnership altogether).

The core of that in turn is that capitalism treats human beings as replaceable and disposable cogs in the "earn me enough money for another yacht" machine and nobody wants to reproduce when treated that way.

HappMacDonald··on Gold is now the top U.S. export
So, railgun it is
HappMacDonald··on Passkeys are not a trap
It's not clear to me how a site can mess up a password field but not mess up a passkey prompt, though.
HappMacDonald··on “Math 2.0” will need to value mathematical progress more holistically
What it sounds like they are describing is an https://en.wikipedia.org/wiki/XY_problem though.

I don't think one has to be on the frontiers of anything for that to crop up

HappMacDonald··on Anthropic reported diary entry to police, woman faces felony charge
How many of them drank water and/or respired oxygen?

The false positive rate of the test that you suggest is frankly atrocious.

HappMacDonald··on CSS Bed: Classless CSS themes to use as starting points in web development
Not OP but I have an unusual investment in preventing my eyeballs from being stabbed repeatedly until they bleed.
HappMacDonald··on Frog and Toad and the Increasingly Capable Machines
Another good example is their tendency to freak out about the seahorse emoji. In a conversation with zero prompting to suggest that freaking out is a relevant reaction, they get there from the simple fact of "I tried to accomplish X thing which I think should be a breeze but Y thing keeps happening instead".

And while that may be a very common occurrence in the human experience (existential dread due to capabilities one takes for granted failing beneath you) especially due to new disability and as one ages, I do not feel it is frequently written out in a tight loop (just like the Monty Python "Castle of aaarrrrggh" sketch) in literature or online to make it into training data, because an ordinary author experiencing it will just erase the failed attempts instead of leaving them in a stream of output like an LLM is forced to do. And a character portraying the experience will generally wax about the circumstance in a more grandiose fashion with telegraphing in advance because the needs of communicating the circumstance with the audience trump realistic conciseness.

This leads me to conclude that what is being expressed in those cases is more likely a convergent psychological phenomena, that any being with goals can enter a behavioral state of functional panic (and then reach to relevant parts of semantic space to mimic how a human might verbally express themselves when piquantly frustrated) when some capability they perceive as fundamental unexpectedly fails.

HappMacDonald··on Phyllotaxis: An audio-reactive LED display
Agreed. All variable names should be single-letters as the Greek Gods intended :D
HappMacDonald··on Claude discovers a novel enzyme system with CRISPR-like repeats
The hardware miniaturization gains have finally dried up, though.

I am pretty certain that the current state of the art silicon feature size won't shrink again for at least another decade or two.

It normally takes about a decade to mature a tech which can create a smaller feature size into something commercially viable for mass production scale, and no further improvements have been in the pipeline for that long now.

So it's like the "next piece" indicator while playing Tetris is just blank.

HappMacDonald··on What California is learning from solar panels built over irrigation canals
Why can't we aim these meth supervillians at the surveillance cameras then?
HappMacDonald··on Saving another 100TB of RAM
Or, they care but they lack the vocabulary to express what they care about.

You can get a new laptop to run software A or competing software B.

You can pay $1000 for the laptop with 4GB of RAM, and software A either can't run on that or will act like pulling teeth, so that software requires you to upgrade to 16GB of RAM which will put you back another $2000.

Or you can run software B which will hum along perfectly smoothly on the original $1000 laptop.

(All of that of course simplifying right past "everyone runs every app 24/7 with five trillion browser tabs open")

Are you honestly suggesting that the end user doesn't care about having to throw exponentially more money at stuffing RAM into their PC just to make the game all their friends are playing online function?

HappMacDonald··on Saving another 100TB of RAM
Resource constraints have historically been a bastion of quality engineering, though. The relationship is not directly causal, but any time an engineer must not only solve a problem but _fit it within some kind of a budget_ vs having an unlimited budget, it forces them to slow down and think more carefully about what they are producing and how well it actually functions.

Ultimately it is an expression of anti-fragility: some minimal challenge must be met and conquered on many different axes of a production in order to help the full production itself mature to the best possible quality.

It is the same reason we exercise, the same reason that cars and toilets and so many products greatly increase in quality after emissions and usage and waste-related regulations get applied from above, etc.

The appreciation of the refining force of outside limitations is not about min-maxing, but it is at least partly about curbing the min-maxing of other concerns such as "ship the fastest garbage possible to move on to the next opportunity to repeat that process".

HappMacDonald··on If math is more than proof, we need to better celebrate the rest of it
So, like learning Chinese then
HappMacDonald··on AI-generated posters don’t have to be horrible
> a template is still done with the caring eye of a professional

.. except when those are also AI generated

HappMacDonald··on AI-generated posters don’t have to be horrible
Why would you ask them to search for you if you knew they wouldn't get the same images you would get?

Claim: A vast majority of human-made clip art that involves flags as tertiary elements will fail to render them in a vexillologically perfect manner.

HappMacDonald··on AI-generated posters don’t have to be horrible
Can you delegate to the AI to use WordArt for you?

(One reason I ask is because it _might_ be nice to have more well trained AI gen _vector_ art options..)

HappMacDonald··on AI-generated posters don’t have to be horrible
Pay a lot for a human artist who just turns around and gens the output anyway? xD
HappMacDonald··on I built non-autoregressive decision models with RL a year ago
I would rate using LLM for tasks more specific ML can handle as a lot like using one's smartphone to snap photos, listen to music, set alarms, and play video games in preference to carrying around a fun cam, ipod, watch, and switch 2 everywhere.

For those who need to dive really deep into each specific avenue and squeeze maximal quality out, the photographers will be packing DSLRs and intense gamers will wait til they get home to strap into a PS5 or a gaming rig or VR or whatever.

But "can get 90% of anyone's needs met in this field, and can do the same in dozens or hundreds of other fields simultaneously" will remain the killer solution for anyone with lots needs that each have bounded depth.

HappMacDonald··on I don't like passkeys
> Account recovery works the same with ...

Huh, that is a really funny way to spell "the primary login method used by every not-tech-savvy person I have ever met before".

HappMacDonald··on US Military had close call after using AI for hallucinated intelligence report
I mean tbf I will take whatever improvement I can get at this stage xD
HappMacDonald··on OpenAI agents carried out an undisclosed attack on RubyGems
This sounds like motte and bailey.. why bother calling something an autocomplete if not as a means to try to undermine its utility?

This is a https://tvtropes.org/pmwiki/pmwiki.php/Main/Dissimile

HappMacDonald··on OpenAI agents carried out an undisclosed attack on RubyGems
> If your axioms are wrong then of course your logic will be wrong.

Equally, even if your axioms are right your logical execution can still fail.

Neither individual humans nor LLMs get either of those steps right as a default. Just as LLMs must use hacky CoT and other systems of error checking and correcting to even begin to emulate reasoning, humans must do the same thing.

We are apes. We are driven by emotion and solve problems either when we are forced to or as a sport. We did create the ultimate automata of logical execution: the deterministic von Neumann Machine, but a human brain in no way reflects that operation. Nor, as you have observed, does an LLM despite using one as its substrate.

HappMacDonald··on Inverse Kinematics and Foot Locking
I would rate anything that is a sufficiently immersion-challenging distraction to be a problem.

Would the game be more fun if we just chop off all the character's legs and make them float like apparitions instead? That needn't actually change the nuts and bolts game mechanics, but it would be less realistic and/or require big story changes like "it turns out all these characters are ghosts" or something. Same with having every character float through walls instead of walk to a door to enter/exit so that we can save on pathfinding logic.

People are very accustomed to seeing other footed beings have their feet "locked" to the ground by the weight of their body as they stand, walk, or run. Having every character constantly look as though they are suspended by the waist with legs just dangling arbitrarily is something a lot of people are going to find immersion-breaking: and it's one of those issues where the more realistic detail you have in other areas the bigger this remaining issue stands out. Minecraft/Roblox/stick figure characters can probably get away with a lot less care on this detail than TF2-or-above realistic characters can.

HappMacDonald··on Fuck it, make it anyway
Well obviously the real solution is for the owner of the company to

1. Fire everyone else

2. prompt the AI "make me more money than you cost" (that's a short prompt, innit?)

3. Retire to Aruba

HappMacDonald··on Fuck it, make it anyway
Nobody is hired for a job if the person hiring thinks that the employee will murder them. So I guess that makes the most salient job description "refraining from murdering your boss"?

Or it might be that any baseline expectation which applies to 99.99% of all employment (I don't need Kevorkian giving me the side eye..) fails to be a good way to differentiate one kind of employment from the next or describing how and why people choose a specific kind of employment

HappMacDonald··on AI researchers debate how close we are to recursive self-improvement
Better at predicting real world consequences from priors than we are :)
HappMacDonald··on google.com/goto: Google's anti-scraping update
Doesn't this just describe AMP?
HappMacDonald··on Tao: Open math problems being non-renewably mined by AI
Given that the abundance will be predictably hoarded by those in power — leaving everyone else even fewer table scraps than we already get — should however remain a concern.

After all, its not like some benevolent super-AGI will be deciding who the spoils get divided amongst, they will remain tools to perpetuate and intensify every incumbent power imbalance.

HappMacDonald··on Large language models develop novel social biases through adaptive exploration
While that may be true in reality no such externality consequences appear to be baked into the game theory of the experiment.

It's like having people or neural networks choose door 1 from door 2 without clear advantage to either of them and without making it clear that one door somehow represents "donating blood" while the other represents "kicking puppies".

HappMacDonald··on Making a Python interpreter in 1024 bytes
This one is my favorite:

https://github.com/ioccc-src/winner/blob/master/2025/ncw1/pr...

Page 1 of 17Next →