HNHacker News
TopNewBestAskShowJobs

fnordpiglet

13,709 karma · joined March 30, 2021

Oink.
submissionscomments
fnordpiglet··on I am often wrong
Interesting what I found objectionable was his expectation that everyone around him follow his internal process, regardless of its steps or urgency.

Someone who has an ounce of leadership realizes that a team is a puzzle of different pieces of different shapes. You fit them together in varying ways, by observing how their processes and methods and strengths interlock to be greater than the sum of their parts. You do this fractally as you become more senior, and expect people to follow this meta framework rather than some fixed process or way of perceiving. You see quickly the organization takes on different strengths depending on the puzzle underneath them, composed of nonlinear amalgams of unique individuals. You do create protocols for communicating and decision making that are relatively uniform, but you get your grip off of people’s mojo and let them be people, not one of whom is Boris (unless Boris works for you). But beyond building a Postels law structure for communication and decision making, you expect the meta framework to be the way people evaluate their teams and you evaluate your team in the same way.

You eventually need people to shift around and reorg people to the shape of the system (Conways law) based on their strengths and processes that work for them and their teams. If you need the system to be different you organize the people in that way in and expect the system to change to fit (again Conways law).

None of what I read reflects a thinking like this, and I’d prefer to not work with someone who feels their role is to mint out clones of themselves. Further someone who posts about how they are frequently wrong is the first sign of a narcissist that believes they’re actually never actually wrong, but admits they might be misinformed at times. For me, it’s a red flag to avoid.

fnordpiglet··on Singapore’s National Library Board offers micropayments to build reading habits
I’m confused by the mechanics. How do they know you’re reading unless you have a device that records you reading while you read, so is likely the device you’re reading on - encouraging people to stare at devices?

I actually read on my phone and have since ~2005 and a Palm device. I don’t use e ink or paper much simply because my phone fits in every pocket and is always there. I read constantly at every moment I get a chance to. No other form factor works for that. Even in a suit it fits well in the jacket pocket.

fnordpiglet··on Sherline Tools Is Going Out of Business
I think you’re touching on a lot of things here which form a self reinforcing web

A) pensions don’t exist so people need cashflow to retire. You either save enough capital that you can build an annuity for yourself or you have cash flow investments and the lowest effort one is rental properties.

B) taxes for capital gains on a home that’s appreciated enough incentive rolling your home into a rental then rolling it into more rentals as it appreciates further afield into lower cost areas. The basis isn’t taxable as long as you roll it to another investment, and as prices go up, you are incentivized to move it into more properties to diversify at lower rents but more units. As that effect continues as people retire the cost of rentals further from the city core keeps rising. An easy fix would be to raise the capital gains deduction for a primary residence to incentive retirees to outright sell and take the annuity option, which is actually easier overall but the capital gains tax makes it unattractive today.

C) Obviously the financing your described, but compounded by private equity and hedge fund dominance in the unit and development rental market skewing towards higher margin lower occupancy developments. The marginal cost of a mass produced “luxury” rental allows for a lot of vacancy, compounded by the IO and other benefits to reduce vacancy load costs. There are also tax benefits to vacancy that improve as your portfolio gets larger. I think this is more than the retiree effect honestly, I think funds dominate large developments now and increasingly unit level portfolios.

D) The rezoning effect is IMO making it worse because it’s making the optionality of low density developments greater allowing for land value appreciation that wouldn’t happen otherwise because of (C). It’s taking a property with lower utility and making it more valuable for a larger development, but there’s no incentive to make the development more affordable per unit. Instead it increases the price per square foot of property value, increasing costs for everyone, while marginally increasing units but they’re constructed and priced to keep price per square foot of livable areas equal or increasing.

I think (A) can’t improve without a real social security system that replaces pensions. (B) can’t improve without tax reform realizing the nominal cutoffs don’t account for inflation and aren’t “rich” taxes any more but retiree taxes. (C) is probably intractable without socialist regulations which are perceived as unfair, and are hard to craft in a way that is not in fact unfair or have unintended consequences (D) is hard because it’s actually a demonstration of the unintended consequences of regulation and ethical capitalism - but the effect is paradoxical.

These are all structurally hard problems and they aren’t made better by the rate of global population growth. There are way more people today than there were before, and economic activity is much denser than before. The urban areas are crushed, and the rural areas empty except for a hoary crust of MAGA with disproportionate state and federal power. It is frustrating all over - and I don’t see an end state that is very positive.

fnordpiglet··on One-Electron Universe
You may, but if you read the link, you’ll find it consistent with the second law of thermodynamics despite being clever about it. The demonstration of consistency requires a modern understanding of information theory, but it’s been long realized it’s likely not a real “out.”
fnordpiglet··on One-Electron Universe
Complications vs simplifications was a restatement of Occam’s razor as you assuredly know and knew. However complications added can be doors to other hidden processes that are simply poorly understood or observe. The increase in complications isn’t necessarily a pure gradient, it can indicate where more interesting questions lie.

In this specific case experimental evidence shows it’s almost certainly not true. It was less the complications that killed it than objective observation.

fnordpiglet··on Sherline Tools Is Going Out of Business
Something in this is contradictory and just hoping for a clarification - over investment in real estate should drop costs. I assume you mean under investment leading to prices requiring an overly burdensome amount of personal investment in housing?
fnordpiglet··on Measure internet censorship
Is this an “instantly be arrested by the secret police” app for those where such things are censored?

The glowing reviews would reflect quite the survivor bias.

fnordpiglet··on One-Electron Universe
I’d note that occams razor is a probabilistic statement and not a law.
fnordpiglet··on One-Electron Universe
The answer is that they don’t literally go backwards in time showing reverse causality. They are representable in QFT by reversing the time component of their counterpart, which doesn’t impact how they interact with broader causality, just their internal configuration. Further - this does not play out in the lab as reversed causality even between anti particles. You can collide two positrons then detect them later at their destinations just fine in normal causality. It’s a shame because it would be really sweet to find some reversal of entropy and causality somewhere or somehow!
fnordpiglet··on One-Electron Universe
IANP, but Ive always thought you should be able to observe reverse causality in a positron if true. That’s the primary effect of going backwards in time - a reverse in entropy.

The other side of the theory is the positron hidden in the proton. I assume a particle collider should be able to falsify this.

But it turns out antimatter isn’t supposed to actually move backwards in time but follows the same causality as everything else. It’s just that they represent their matter pairs configuration mathematically by reversing the time orientation. This isn’t super obvious but it’s very much different than going backwards in time literally. Sad!

fnordpiglet··on New evidence for hidden chambers beyond Tutankhamun's tomb
It’s hard to believe in 2026 someone would bemoan that nothing has progressed in cancer research. Cancer has no single cure, but survival durations are absurdly crazy for most common cancers compared to 30 years ago. Current treatments are like science fiction and are extraordinarily effective. The horizon is remarkable as well. It’s genuinely surprising to hear someone say “never any progress” re: cancer.

I’ve nothing to say about Egyptian tombs though.

fnordpiglet··on Disney+ changes subscriber agreement to allow ads on every plan
I don’t get it - what does paying for the premium tier buy you then? I’ve always been willing to pay for content, and happy to pay to not see ads. But this is bizarre - pay less, see ads; pay more see ads?

Disney+ is probably the least valuable streaming subscription I have but I keep it for the backbook of star wars and marvel content. They keep trying to package it with random crap, and I assume they’re going to wave hands at some garbage package as the benefit. But I for one absolutely refuse to be marketed to, especially during my rare moments of free time.

fnordpiglet··on Claude Code now reads AGENTS.md if there is no Claude.md
Because, knowing what to induce in its behavior is the key - by understanding things like Postels law, I can direct the agentic loop towards patterns of development that avoid the cascade of issues that seem endemic in modern agentic harnesses.

In fact, this is a key insight, so it’s surprising you didn’t get the punchline. Trying to build things in a new way on a machine composed of the corpus of all the old ways is stupid. By eliciting the corpus of tried and true methods over the history of computing and process engineering as the grounding for how to develop and behave in processes, you offload a huge amount of the work in getting software that’s “right” for the domain you’re working in. But not being aware of them and being heedless that the techniques hard won were hard won by people at least or more smart than you facing similar problems is leading to a cycle in software that’s needlessly dumb, making stupid mistakes that are unnecessary.

There was a great Star Trek NG episode where Picard is stranded on a planet with a creature that can only speak in metaphorical language. Every sentence it uses refers to an event in the past and you have to understand it’s history to understand it today. Because LLMs are entirely trained on historical corpus, there is no time in history when all the techniques and processes and system design thinking was more relevant. Like that episode, Darmok, you can elicit a wealth of experience by simply referring to a technique from the past - if you know what it is.

Another way to put it , those that aren’t aware of their history are doomed to repeat it. The sad part of the situation is the coding agent you’re working with is fully trained on it, but unless you intentionally activate the semantic space and bring the concepts into its J-space you won’t maximize the value of that corpus. So, we are using these tools which are mostly crap duct-taped together, which suffer from flaws well understood for decades, mostly because of the driving human being ignorant of the available corpus.

fnordpiglet··on The Contract on America
Clinton, who won the majority of votes or Harris who in a few months forced Trump to a plurality win?

Both were admittedly weak and forced, but Clinton would have won had any one of number of factors been different, and Harris walked into a tire fire and wasn’t the right candidate no matter what.

Democrats certainly need to field a better slate, but the fact people voted for self immolation can’t be attributed to a lesser of two evils argument - because Trump was never a compelling leader, and he has never been popular - hence his inability to ever secure a majority and operating the two most unpopular presidencies in modern history. He was voted into power by people who were sold on “at least he’s different,” which is more of an idiocratic solution to the current politic than anything else.

The answer is as the article says, it’s both national suicide and murder by the administration. There is no point in second guessing it, a technical victory and a plurality chose a national policy of global stupidity, taking a wrecking ball to the complex international system.

The hope I hold is this weakens the presidency irrevocably as Congress takes control again and places the executive in its proper functionary role. It’s abdication over history as Washingtonian presidents set a higher and higher bar of relative trustworthiness was fraught and stupid, and is IMO the actual manifestation of exceptionalism. And we f’ed around and found out, as MAGA is so fond of.

fnordpiglet··on Claude Code now reads AGENTS.md if there is no Claude.md
My view is at the end of 35 years in the industry the current wave of “software” and “systems” engineering coming out of the AI harness community is pretty garbage. But web technology and distributed systems followed a very similar arc, as did protocols and memory management; and microcode before them. But it feels like this is particularly bad because the mistakes being made are plainly obvious to the grey heads who have been here a while. As opposed to before when mistakes were in new domains being explored, these are mistakes made before and have good solutions to.

I feel people too readily blame the LLMs themselves for this, but I’ve found LLMs know the history of computing thought evolution better than anyone I’ve ever encountered. Once you push them in the right direction, ground them in the philosophy of thought of hard won engineering ideas, they are astoundingly precise and accurate in their read and application (keeping every session grounded is the trick!). So it’s not the machines making these same mistakes with ready conceptual frameworks around them, it’s the 22 year old gatekeepers dashing head first into wall after wall, when we painstakingly built the door two feet to the left about the time they were gestating.

fnordpiglet··on Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint
They’re fairly useful as constrain domain classifiers due to the low memory requirements means you can stuff a lot of them into a less expensive GPU farm and get really decent throughout with pretty good results over all. At least that’s my experience. I wouldn’t bother using a tiny model for coding - but the world is full of abductive reasoning tasks that don’t involve coding.
fnordpiglet··on Artificial intelligence now beats some of the best human forecasters
Forecasting = modeling + intelligence
fnordpiglet··on Part-human part-mouse brain developed in science breakthrough
I for one welcome our part human part mouse brain overlords.
fnordpiglet··on Houthis 'take control' of key island in global shipping route
I have a suspicion that the caring is starting to mount, otherwise there’s no need to offer a $1.3T election bribe (which I’ll file a the other dividend / bribes that were never sent and the bridge Trump sold me)
fnordpiglet··on Houthis 'take control' of key island in global shipping route
Triple digit when summed I assume?

If they were going to try this it needed to be done months earlier when the protests were raging. Instead the brain trust fumbled the initiative and let the Iranian leadership kill and scourge their local allies into compliance. Then fumbling the decapitation strike coupled with the decades of preparation for a decapitation strike (via redundancy of command) the genius brigade seemed to be unaware of because they fired all the experts, and the total repression of the protesting populace, led to an unusually strong system in Iran. They also almost certainly let racism play a big role in underestimating the Iranians, who are a very technically sophisticated society and have been preparing for this scenario since day one of the revolution.

Everything that’s happened was predictable and predicted, which is why other administrations who were more than happy to topple foreign governments when it suited them to not do so in Iran. MAGA feels like the junior dev that blames the compiler for their own syntax errors.

fnordpiglet··on Claude is no longer available for minors
Glad kids can’t learn how to use AI effectively at school! Would only want them on laundered accounts to cheat, rather than official school accounts to learn the tools for practical purposes. That would be a real shame.

My daughter’s school is working hard to make AI use a complementary aspect of learning, rather than adversarial. Like the calculator it can be used to accelerate learning once you’re past the basics and using it for practical automation and task purposes is not dissimilar to any other technical skill and is highly practical.

The knee jerk “no minors for stochastic parrot calculators” is absurd. And I’m no anarchist parenting style parent - we do parental controls and mirroring and limit / ban social media. We are more Orwellian than free range. But this is one step too far, too unthinking, too simplistic.

fnordpiglet··on Growing proof that autonomous cars save lives
Except the thread is about trains
fnordpiglet··on Growing proof that autonomous cars save lives
When you say most places, I assume you leave out most of the world in this? The number of countries with broad complete train networks is small compared to the land mass of the earth inhabited by humans. Even considering urban areas, most people don’t live in London or New York or Tokyo. They live in a city like Jakarta, Dhaka, Mexico City or Lagos. Even Bangkok which has the MRT and Skytrain only services the relatively affluent areas with service, everyone else is car bound.

Most urban US cities have some attempt at an urban train network, and most significantly more useful than Seattle. The issue is typically more of funding and willpower - major infrastructure in an established city is expensive and disruptive to the point most don’t want to deal with it. Additionally the feeder transit infrastructure tends to be weak enough that the daily practicality is low because you have to live right on the line and your destination must also be. Otherwise you’re driving in. Finally the urban planning for surface roads has been well enough structured that the trade off for time saved in public transit doesn’t make sense. You spend time walking and doing transfers in excess of double a driving time.

Is that a flaw? If you electrify the cars to minimize pollution, automate their driving to reduce risk, ensure they collaborate to improve efficiency, I’m not sure I buy the argument any more.

fnordpiglet··on No Man's Sky Cosmos
Yeah this is my experience as well. At first it felt empty, but then I sort of realized it’s not a quest game it’s a building game and it’s all about building bases and ships and what not, and they’re added progressively more ways to build and customize the world. And there are explorations - there are some truly bizarre biomes but they are rare. The variants that feel repetitive are falling along some average world characteristic - probably not too dissimilar to the universe at large. It takes time to find the weird ones.

If you’re looking for a scripted arc, wrong game by construction.

I think the multiplayer aspect is a bit weaker than it could be, maybe improved more recently, but it is a something that Minecraft has done well

fnordpiglet··on Muse – Meta’s personal AI agent
I’ve desperately tried to use the Claude work, ChatGPT work, even OpenClaw, and I’ve yet to find a meaningful use of such an assistant. Maybe my life is too boring, but what I see here is a bunch of stuff that makes life more complicated. IF the things I use are well supported here, and IF I have manual repetitive tasks to delegate with them, it feels a bit like taking a nap while using FSD - I don’t trust an LLM to make life decisions for me.

Summarizing FB? Christ I hope not. Paying my bills? Auto pay was like 25 years ago. I’m still staring at the whole ecosystem and shrugging a bit. That it’s offered by Facebook - the “we will never has ads ads are for losers” company promising security and privacy today with a T&C that changes by the metrics wind - is much less compelling than the other offerings that at most I believe they will use my data for training.

But I’m solicitous of use cases - has anyone actually used one of these harnesses to their benefit on some on going basis ?

(Btw I did use Claude work once - I had it shop for my daughters textbooks on a complex UX website that would have required me to tediously do a lot of repetitive tasks, and it sort of did it, over a long period of time - and felt like a perl script could have done it better faster)

fnordpiglet··on GPT-6 Astra
5-1 is considerable cheaper I’ve found and feels a lot more like 4-6.

But opus 5 described as a terrorist is being generous.

fnordpiglet··on New type of dice guarantees no tie when deciding who goes first
The provenance doesn’t matter because they settled on a difficult mathematical problem to solve because it was difficult to solve and would require clever solutions that answered more fundamental mathematical questions than the banal problem of choosing who goes first in a game. That was just the framing that motivated the larger math question, in which the number of dice does matter, because it’s a more interesting question.
fnordpiglet··on GPT-6 Astra
We have a mandate to only use 4-8. Sonnet 5 is pretty good and fable 5-1 has been pretty good so far fwiw.
fnordpiglet··on GPT-6 Astra
Unless things changed this used to work but was blocked. You also used to be able to force in a model catalog to get around the token cap. Each, as of at least April, were closed off and ineffectual. I’ve not tried lately so you may be right.
fnordpiglet··on New type of dice guarantees no tie when deciding who goes first
Because the question isn’t can you make a rice with 5! Sides but can you make it so 5 players can roll simultaneously individual dice with a minimal number of sides per dice and never tie and always be fair.

Sure you can fly a helicopter to the top of Everest, but it’s not the same as climbing Everest even if the outcome is the same.

← PreviousPage 4 of 34Next →