HNHacker News
TopNewBestAskShowJobs

short_sells_poo

2,257 karma · joined July 19, 2013

submissionscomments
short_sells_poo··on Show HN: Krita RGBA Tech – Bringing Realistic Metal to Life in Open-Source Art
Photoshop and other painting software had "intelligent" brushes for a while now. These try to simulate stuff like paint mixing (even watercolor), opacity and texture.
short_sells_poo··on The Silurian Hypothesis
Perhaps it's down to scale? If a single relief of a single building block survives, is that still considered a ruin of a football stadium, or just an imprint of a building block? The Colosseum would certainly not be able to survive for 100 million years. If some intelligent beings find a single building block while doing archaeology, they'll likely not be able to recognize it's purpose, if they even recognize the building block for what it is.

The structural integrity of (comparatively) smaller fossils is much more likely to survive for 100s of millions of years. A large building will be eroded by roots, wind, acid rain, etc. A fist sized rock containing the imprint of a trilobite, buried in soil and safe from most elements, will easily survive for 100s of millions of years.

Or perhaps the author meant that the vast majority of our buildings are not built to last. An asphalt highway - if unmaintained - will be unrecognizable in a hundred years. Roman roads survived for a few 1000 years because of their very crude nature: they are just a pile of rocks. But even such structures will fall prey to soil erosion and the relentless assault of nature. Unless said roads see foot traffic, they'd be also quickly overrun by trees and grasses.

short_sells_poo··on Wooden satellite heads to space in Mars exploration test
Ha, interesting proposition. What do you think would be the alternative to DNA? I'm curious as I'm a complete layman. Intuitively, I'd think that a self replicating life form will need something to carry it's program, and this is what DNA is. Alien life might use different molecules to carry that program, but conceptually it would be the same, no?

I'd even go as far as to say that DNA arose out of the fact that carbon based chemistry is good for life (we see N=1 case at least), and once we accept carbon based chemistry, amino-acids are the next optimal step, followed by something like the DNA to program the construction of proteins for amino acids.

Again, I'm just a layman so it'd be nice to know the views of experts in this area.

short_sells_poo··on Former Bayesian captain on sinking, vessel ballast and stability
I think I phrased that badly. What I meant to convey is that a cheap and cheerful solution will need replacement at interval X, and a much more expensive solution will need replacement at 1.2X. At least in my experience, the relationship between the money spent and reliability tended to be logarithmic.

In other words, spending more on fancier equipment didn't really minimize the main pain point of failures: the ship being unavailable for sailing due to maintenance. It was worse in fact, because more expensive equipment wasn't more reliable usually. Shiny lcd touch panels looked great but were a ballache to use when there was a swell. I had chunky 800hp mercury engines fitted to crack that magical 100mph barrier, but because they were fly by wire, we had constant issues with electrical contacts that were difficult to diagnose out at sea (unlike the older mechanically actuated throttles).

Even worse, the more "over-engineered" a piece of equipment becomes, the more difficult it tends to be to reason about it's boundary conditions. More moving parts, heavier weight, more exotic materials - these all add further unknowns to an already very chaotic system that is seafaring. To come back to my electronic vs simple mechanical throttle system, the former is difficult to repair with a set of pliers, the other is trivial. Even though the former may have redundancies and fancy recovery modes, if it does fail, you are SOL.

So you are right, ultimately it's a cost question in the sense that if the more expensive components aren't more reliable, why spend money on them. I was keen to make the point that I'd have gladly spent more money on said equipment to make it more reliable, but there seemed to be no correlation between the two aspects :D

short_sells_poo··on An embarrassingly simple approach to recover unlearned knowledge for LLMs
You raise very good points, and I agree that scale is not necessarily bad, in fact it can be a source of much good. Scale simply increases the frequency and thus likelihood of things, whether good or bad.

I'm sure that big players will be able to assert their rights with their armies of lawyers. Just like the music and movie industries have after the rise of file sharing.

My worry is perhaps more subtle: I'd argue that generative AI draws much more from the masses of small content creators and they will not be able to assert their rights. In some sense, if people pirate the next blockbuster movie, the producers might only make 1 billion instead of 1.1 (and piracy has never been proven to actualy impact sales), but if all content starts being consumed via massive and centralized anonymizers, the masses of people who made the internet what it is will eventually disappear. So the scale on which these tools can hoover up information and reproduce it is unprecedented, as is the fact that nobody important seems to be thinking about how to make sure that we can keep actual humans motivated to keep generating the content that actually feeds the AI.

It's one of those cursed things: the long term interest of the AI companies is that humans will keep feeding the beast with more information, but the short term interest is to capture their audience and do everything possible to keep them inside the walled garden of a single AI provider. It is not in the companies' interest for people to step outside and go straight to the painter/writer/moviemaker, because at that point the AI is no longer needed.

short_sells_poo··on Former Bayesian captain on sinking, vessel ballast and stability
It's not (just) cost saving. Having owned a yacht, everything breaks all the time. The environment is extremely harsh on everything. Seawater is corrosive. Even high quality stainless steel (expensive!) will corrode. Sealife is invasive. Every nook and cranny is under constant assault from all sorts of weird little critters.

The most robust solution is to keep things as simple as possible. Any bells and whistles you add will become liabilities very quickly, and if crucial systems depend on these bells and whistles, you can quickly get cascading failures.

I learned in about 2 months that you want the most crude, easy to repair and reliable solution to important parts and you want them cheap - because you'll be replacing them regularly. Expensive solutions will provide very little relief from this (if any at all). If the TVs in the guest rooms break regularly because of constant vibrations, that'll annoy the guests, but the ship will float and go forward. If your fancy fly by wire engine controls corrode and you are out at sea, it's mayday time. A lot of smaller boats still use mechanical wire actuation because it can be mended literally with a string.

short_sells_poo··on An embarrassingly simple approach to recover unlearned knowledge for LLMs
I agree, but that doesn't make it good - or perhaps even acceptable. To quote myself answering another commenter:

> Never before has been the reuse (I'm trying to avoid using the word theft) of content produced by others have been conducted on such an industrial scale. The entire business model of LLMs and generative models has been to take information created by masses of humans and reproduce it. They seem to have zero qualms taking all the work of professional and amateur artists and feeding it into a statistical model that trivializes replication and reproduction. You could argue that humans do this as well, but I feel scale matters here. The same way that a kitchen knife can be used to murder someone, but with a machinegun you can mow down masses of people. Please excuse the morbid example, but I'm trying to drive a point: if we make a certain thing extremely easy, people will do it, and likely do it on a mass scale. You could argue that this is progress, but is all progress inherently beneficial?

short_sells_poo··on An embarrassingly simple approach to recover unlearned knowledge for LLMs
You raise some great points and I agree it that we are on tricky ideological grounds. I'll try to provide sensible counter-arguments to your encyclopaedia and teacher examples, and hopefully not fall into strawmans (please do object if I do):

1. First there's the motivation or intent. Teachers want to earn a living, but their purpose in some sense and (hopefully) their main intent is that of education. I argue that teachers should be paid handsomely, but I also argue that their motivation is rarely to maximize profits. This is contrary to the bog standard Silicon Valley AI company, who are clearly showing that they have zero scruples about breaking past promises for those sweet dollar signs.

2. My second point actually builds a bit on the first: both encyclopaedias and teachers tend to quote the source and they want their audience to expand their research horizon and reach for other sources. They don't just regurgitate information, they'll tend to show the reader where they got the information from and where to go for more and neither the teachers nor the books mind if the audience reaches for other teachers and books. LLMs and generative models are/will be/have been capable of this I'm sure, but it is not in their creators' interest to enhance or market this capability. The more the users are walled in, the better. They want a captive audience who only stays in the world of one AI model provider.

3. Scale. Never before has been the reuse (I'm trying to avoid using the word theft) of content produced by others conducted on such an industrial scale. The entire business model of LLMs and generative models has been to take information created by masses of humans and reproduce it. They seem to have zero qualms taking all the work of professional and amateur artists and feeding it into a statistical model that trivializes replication and reproduction. You could argue that humans do this as well, but I feel scale matters here. The same way that a kitchen knife can be used to murder someone, but with a machinegun you can mow down masses of people. Please excuse the morbid example, but I'm trying to drive a point: if we make a certain thing extremely easy, people will do it, and likely do it on a mass scale. You could argue that this is progress, but is all progress inherently beneficial?

There's value in these models, so we should use them. But I feel we are rapidly hurtling towards a walled garden corporate dystopia in so many areas of our society. Industries which tended to have negative impact on our lives (waste, tobacco, alcohol, drugs) have become heavily regulated and we have paid for these regulations in blood. Will we have to pay the same blood price for the harmful industries of the new age?

short_sells_poo··on ChatGPT Search
Just copy pasting the response to another commenter asking in similar vein:

Thank you for the kind words. I've been thinking of doing something long form, I'm just unsure if there's enough of an audience in this age of catchy tweets and tiktok videos.

Full disclosure, I had no formal education in writing... I suppose all credit goes to the actual great writers I take inspiration from.

short_sells_poo··on ChatGPT Search
Ha, thank you for the kind words. I've been thinking of doing something long form, I'm just unsure if there's enough of an audience in this age of catchy tweets and tiktok videos.

Full disclosure, I had no formal education in writing... I suppose all credit goes to the actual great writers I take inspiration from.

short_sells_poo··on An embarrassingly simple approach to recover unlearned knowledge for LLMs
Thanks for the link, I appreciate it. I suppose the issue is that this just further enshittifies the internet into a small handful of walled gardens. Big players get their payday, because they could feasibly sue OpenAI and generate them enough headache. But the vast amount of content on the internet was not built by a small handful of media companies, but rather by masses of small creators. It is their work that OpenAI is profiting from and I have yet to see a credible suggestion on how they will compensate them.
short_sells_poo··on An embarrassingly simple approach to recover unlearned knowledge for LLMs
> We know that our model would be useless without your source. So we will take the useful part of your source and obfuscate the rest so that we can charge our users for utility provided by you without having to pay you anything.

Isn't this basically the entirety of the latest AI craze? They basically took a public good - the information available on the Internet - and hid behind some thin veneer of "we are not stealing, we just trained an AI on the information" and then they sell it. Note, I'm intentionally not writing "free information available on the Internet", because information is not free. Someone has to pay (in time or money) to generate it and host it. They might have provided it gratis to the public, but nobody asked them if an AI can come along, harvest it all and regurgitate it without a hint of reference to the original source.

Much of that information is not even free in the monetary sense, it is supported by ads. The AI will not only not click through the \ds, it won't even generate repeat traffic as once the information is harvested, there's no need to access the source anymore.

If you really think about it, it's a brilliant business model. It's a perfect theft, where the affected group is too diffuse and uncoordinated, it's extremely difficult to prove anything anyway, and the "thieves" are flush with investment capital so they sleep well at night.

LLMs have undoubtedly great utility as a research tool and I'm not at all against them. I think they (or a model similar in objectives) are the next step in accessing the knowledge humanity has amassed. However, there's a distinct danger that they will simply suck they sources dry and leave the internet itself even more of a wasteland than it has already become. I have no illusions that AI companies will simply regress to the lowest cost solution of simply not giving anything back to whoever created the information in the first place. The fact that they are cutting off the branch that they are sitting on is irrelevant for them, because the current crop of owners will be long gone with their billions by the time the branch snaps.

short_sells_poo··on ChatGPT Search
Because it is never enough. We see this time and time again. Once they are making billions, the people in charge will demand that they start making dozens of billions, and then hundreds. The growth must never cease, because the moment you stop growing, you can't sell the dream that supports ridiculous PE ratios anymore.

Google was a very profitable business 10 years ago and the search was still decent. In the last decade they absolutely butchered their core product (and the internet along with it) in an effort to squeeze more ad dollars out, because it's not the level of profitability that they need to maintain, but the growth of that profitability.

Microsoft was a ridiculously profitable company, but that is not enough, they must show growth. So they add increasingly user hostile features to their core product because the current crop of management needs to see geometric growth during their 5 year tenure. And then in 5 years, the next crop of goobers will need to show geometric growth as well to justify their bonuses.

Think about this for a moment: the entire ecosystem is built on the (entirely preposterous) premise that there must be constant geometric growth. Nobody needs to make a decision or even accept that this is long term sustainable, every participant just wants the system to keep doing this during their particular 5-10 year tenure.

It's an interesting showcase of essentially an evolutionary algorithm/swarm optimizer falling into a local optimum while a much better global optimum is out of reach because the real world is something like a Rastrigin function with copious amounts of noise with an unknowable but fat tailed distribution.

<rant/> by a hedge fund professional.

short_sells_poo··on Hell Freezes over as AMD and Intel Come Together for x86 – ServeTheHome
And here I thought windows on arm was increasingly sliding into irrelevance as nobody is buying it.
short_sells_poo··on Show HN: Graphite, a Blender-inspired 2D procedural design Rust app
I'm so glad this is a genuine product that stands by it's own merit rather than trying to sell itself on the basis of being written in Rust. I feel too many products that use some fancy new tech in development make that aspect part of the sales pitch. It is certainly interesting for us here, but the users of the tool will only care about the technical side to the degree that the tool is stable, performant and continues to be developed.

I'm not an illustrator so I can't add much beyond that, but I'm rooting for the devs to turn it into a success!

short_sells_poo··on Christopher Columbus may have been Spanish and Jewish, documentary says
Too many big discoveries were made by people who were "too idiotic or hard headed" to realize that something is impossible. Clever and level headed people will look at an endeavour, calculate that the risk/reward is terrible and not even try. It takes a special kind of person to do this kind of stuff, and many die trying such that we don't even hear about them. The few who get lucky and succeed push humanity's reach to (literally) new frontiers and get remembered as great explorers.

So while I agree that a "lucky idiot" is often a good description for these efforts, I think it is very important that we do not diminish their importance. It took someone like Columbus to make this trip, and yes he had to get supremely lucky, it doesn't diminish from the fact that he took risks that paid off handsomely.

short_sells_poo··on Julia 1.11 Highlights
Julia seems to have fallen out of the mindshare completely in my industry (finance). It's a huge shame because I really like the language and what they are trying to do.
short_sells_poo··on How a Gas Compressor Station Works
You pointed out something I've never considered and I feel the same. Gas networks seem to be very reliable and crucially safe (given the contents are pressured and flammable).

I suppose gas distribution networks tend to be over-engineered for safety because if something goes wrong, it goes really wrong: https://en.wikipedia.org/wiki/Merrimack_Valley_gas_explosion...

short_sells_poo··on Homemade AI Drone Software Finds People When Search and Rescue Teams Can't
I mean, if something as traditional as simple clustering is AI, then so is linear regression and Excel Sheets have been doing AI/ML for the past 2 decades.

At some point we just have to stop with the breathless hype. I'm sure labelling it as AI gets more clicks and exposure so I know exactly why they do it. Still, it's annoying.

short_sells_poo··on Waverley, the last seagoing paddle steamer
I spent the entire 2 hour journey from Rapperswil to Zurich just mesmerized by the engine. It moves with a certain ballet like fluidity and beauty. As an amateur metalworker I also appreciate the craftsmanship that must've gone into making the engine frame, the pistons and all the moving parts.
short_sells_poo··on AI chipmaker Cerebras files for IPO
You are so on point! A huge number of amateur investors get obliterated on this. Your call may be right, but that's no help if you don't survive to see it realized.

You may have a hugely profitable idea that could realize crazy gains over a 5 year horizon, but if you get margin called and liquidated in year 3, you'll end up with nothing.

The magic of investment is compound returns, not crazy leverage. Take some of the crazy Nvidia profits and reinvest it elsewhere where you expect geometric growth. Keep things decently diversified.

short_sells_poo··on Orion, our first true augmented reality glasses
I think the hardware and idea is great. But I simply can't ignore the fact that it is made by Facebook, which is basically the posterchild for modern addiction optimizer adtech business.

The last company I'd want to put right in front of my eyeballs for any extended periods of time is Facebook.

short_sells_poo··on Hacker plants false memories in ChatGPT to steal user data in perpetuity
Ah, because some text saying "cigarettes cause cancer" is all that's needed to educate people about the dangers of smoking and it's not a problem at all if you enjoy it responsibly, right?

I'm talking about the industry and a surrounding crowd of breathless sycophants who hail them as the second coming of Christ. I'm talking about malign comments like "Our AI is so good we can't release the weights because they are too dangerous in the wrong hands".

Let's not pretend that there's a strong and concerted effort to educate the public about the dangers and shortcomings of LLMs. There's too much money to be made.

short_sells_poo··on Hacker plants false memories in ChatGPT to steal user data in perpetuity
This is like saying a gun that appears safe but that can easily backfire unless used by experts is completely fine. It's not an issue with the gun, the user should be competent.

Yes, it's technically true, but practically it's extremely disingenuous. LLMs are being marketed as the next generation research and search tool, and they are superbly powerful in the hands of an expert. An expert who doesn't blindly trust the output.

However, the public is not being educated about this at all, and it might not be possible to educate the public this way because people are fundamentally lazy and want to be spoonfed. But GPT is not a tool that can be used to spoonfeed results, because it ends up spoonfeeding you a whole bunch of shit. The shit is coated with enough good looking and smelling stuff that most of the public won't be able to detect it.

short_sells_poo··on Hacker plants false memories in ChatGPT to steal user data in perpetuity
> How is that any different though from regular false or fabricated information gleaned from Google, social media or any other source?

It lowers the barrier to essentially nothing. Before, you'd have to do work to generate 2 pages of (superficially) plausible sounding nonsense. If it was complete gibberish, people would pick up very quickly.

Now you can just ask some chatbot a question and within a second you have an answer that looks correct. One has to actually delve into it and fact check the details to determine that it's horseshit.

This enables idiots like the redditor quoted by the parent to generate horseshit that looks fine to a layman. For all we know, the redditor wasn't being malicious, just an idiot who blindly trusts whatever the LLM vomits up.

It's not the users that are to blame here, it's the large wave of AI companies riding the sweet capital who are malicious in not caring one bit about the damage their rhetoric is causing. They hype LLMs as some sort of panacea - as expert systems that can shortcut or replace proper research.

This is the fundamental danger of LLMs. They have crossed past the uncanny valley. It requires a person of decent expertise to discover the mistakes generated and yet the models are being sold to the public as a robust tool. And the public tries the tools and in absence of being able to detect the bullshit, they use it and regurgitate the output as facts.

And then this gets compounded by these "facts" being fed back in as training material to the next generation of LLMs.

short_sells_poo··on London saw a surprising benefit to ultra-low emissions zone: More active kids
I skipped over this issue, but it's absolutely right. Train tickets are ridiculously expensive - to the degree that it's cheaper to drive into town and then pay the parking. And for this price you get absolutely abysmal service.

I don't know where public transport has gone wrong in the UK, but it is truly catastrophic even compared to the post-soviet bloc countries in Europe. In fact, those tend to have much better public transport these days.

short_sells_poo··on London saw a surprising benefit to ultra-low emissions zone: More active kids
Ah yes, thanks for mansplaining that to me. I was totally unaware that the concept of filtering exists - particularly as a motorbike rider of 20 years...

With that out of the way, like all things, filtering works when all participants are careful. Motorbikes are actually quite a bit easier to deal with (as a car driver), because a) they can be heard and b) they are quite a bit bulkier.

Even as a motorbike rider, it's not cars that I'm most afraid of, but cyclists. Again, this is not about arguing against being careful - on the contrary. I'd like the cyclists to exercise just as much care as they are asking others to exercise. It's basic self preservation. What good is it to a dead cyclist that it was the truck driver's fault?

short_sells_poo··on London saw a surprising benefit to ultra-low emissions zone: More active kids
This also requires said vulnerable participants to stop having a deathwish. I'm scared to hell from cyclists in London, because they are inconsiderate and extremely unpredictable. Try rolling up to a major 4 lane intersection, and you are going to have cyclists materializing out of thin air on both sides.

Furthermore, visibility on UK roads is very poor. You often have very tall hedges lining streets, which means you can't see more than a few meters until the very last moment.

You'd need to basically rip up the entire city and rebuild it from scratch, and then replace all the inhabitants with rational actors. It's simply not going to happen otherwise.

short_sells_poo··on London saw a surprising benefit to ultra-low emissions zone: More active kids
Anecdotal sample from my village just outside Zone 6 (but inside M25): every single one of my neighbours has 2 cars, some have 3 (or 4). Basically every adult or young-adult member of the household who has to go somewhere on their own has a car. You could argue that they don't need that many, but take it this way: the reliability of the UK train network has become absolutely disgraceful. They are either on strike, or if they are not striking, then they are doing engineering works, and if neither of those are happening, then there is a major disruption. This means that in the last 6 months, I think there have been 2-3 weeks perhaps where the connection into London was reliable.

Unless you have a car, you are completely hosed.

short_sells_poo··on FTC: Vast Surveillance of Users by Social Media and Video Streaming Companies
Apologies I was unclear: I'm not criticizing this report, I'm criticizing the lack of action over the past decade or so.
← PreviousPage 9 of 29Next →