I'm going to say it is efficiency and the ability to implement ideas well, even if they are stolen ideas, that account for more of the success for anything else.
I also bet they did come up with some small things here and there themselves, in the process of implementing stolen ideas, because often things become apparent at the moment of implementation.
I have a few questions
- Is this inspired/based on the Blender node editor package?
Some stylistic choices look very much taken from there, even if some other details (eg font) look much more primitive
- If it’s not based on that, why?
Blander has an Excellent Python-based cross-platform ui which is completely open source. I’ve always thought it’s a crying shame that’s not available as a package to build desktop apps. And I think it would be very good to take things from there, like their very mature graph editor
Hello, creator and maintainer of Nodezator here. I love Blender. Used it a lot until 2013. The only reason I haven't used it in the past few years is because I'm not involved with anything that requires me to work with 3D design. In addition to that, I found out recently that Blender doesn't launch anymore on my 4Gb RAM machine.
Blender is the reason why I created Nodezator, that is, because I wanted to use Blender's node editor for general computing as well, not only for 3D related workflows. I even bought Bartek Skorupa's DVD course on compositing with Blender, remember that?
Certainly there must be possible in theory to fork Blender's node-related internals into a standalone app. However, I very much doubt that'd be a trivial task, much less for someone in my situation at the time. When I started making Nodezator I had only been using Python for 2 years and knew nothing of C/C++.
According to Blender's developer website, Python is used for "interface layout, simple tools, key-maps, presets and add-ons", while GLSL is used for shaders and all the rest of the code is C++ (and a bit of C). There's also the size of Blender's repo (1GB + 979MB of LFS storage). With 1GB of source, even with someone skilled enough to disentangle the node editor bits from Blender's source, I believe it would still take a lot of time and probably not an endeavour for a single person. I think that is the reason why no one did it before.
A while ago, I was asked to write a plugin for Blender since I really liked the program. I looked into it but realized that it was a python-only situation and I really just can't stand coding like that, so I passed on it. But if I could get the python ecosystem without having to code like that - that's a really interesting prospect!
Never looked in to its package size, as far as UI goes, but that does sound like something that could be interesting! I do wonder how different it is from imgui, but there could be something there!
I think what you are referring to is the concept of “finetuning”. You use a pretrained network and add a (relatively) small set of new input-output pairs to steer it in a new direction.
It's widely used, you can look it up.
A more challenging idea is whether it is possible to reuse the pretrained weights when training a network with a different architecture (maybe a bigger transformer with more heads, or something).
AFAIK this is not common practice, if you change the architecture you have to retrain from scratch. But given the cost of these trainings, I wouldn't be surprised if OpenAI&co had developed some technique to do this, eg across GPT versions..
Right? I don’t know about this post (and comments) discussing this quote from the western point of view only, ignoring the fact that this is the whole basis of- and arguably explored much richly in- the entire western philosophy
I guess you mean "the entire eastern philosophy" in your last sentence.
I absolutely agree that there is a lot of parallels with Buddhist descriptions of "co-dependent arising", but there are also important differences. Buddhism is focused on suffering, its origin, its cessation, and the path to its cessation. It is a practical path, and everything that is not related to this path is out of its scope.
But I would disagree that eastern philosophy explores this "much richly". The article cites Kant and Husserl in the second paragraph. And the whole idea of seeing "things as they really are" was imported to Buddhist thought by western thinkers[1]. This idea is present in currents of the 3 Abrahamic traditions, in great part through the influence of neo-platonism from the 3rd century onward.
This kind of concept is present in mysticism in all 3 Abrahamic traditions, though in a different language.
As a Buddhist practitioner turned mystically inclined Christian, I would argue that there are lots of facets to look at this, from a wide array of traditions, and all have their place and value. One of the hardest things to do is not to immediately dismiss them by analogy with waht we already think we know, e.g. "oh yeah that's just good ol' teaching of the emptiness of all phenomena, I know that!", without rejecting what one learned until now either. What is beyond word can be described in countless ways, and each of them can help shed a new light and reveal holes in ones current understanding of "reality", if you let it do its work.
[1] you could argue that it is part of some Mahayana traditions, but the way they describe it is so foreign that if you do not engage with it for years, theoretically and practically, you will understand it through a "romantic" lense.
In the case of the Pali canon, the Buddha always re-orients the questions whenever asked about the nature of the universe or of reality, to teach the "noble truths" of suffering, its origin, its cessation and the path to its cessation. I am not aware of any text in the Pali canon where the Buddha would be teaching in order to see "things as they are". It is just not the point of his teaching.
A central concept in Buddhism is actually "emptiness", or the absence of inherent existence of all phenomena. It is sometimes used in Mahayana teachings as "the way things really are", but as I said, this is subtle and confusing, because "the way things really are" is that they do not have independent existence. I would not venture interpreting those strands of teaching too much, as I am not famliar enough with them.
I heard those arguments from various sources, mostly from Thanissaro Bikkhu and Rob Burbea. They are both westerners, formed in the Theravada tradition and scholars of the Pali canon (able to read it in the original language), which I think actually makes them particularly well suited to identify points that might be misinterpreted from a western reader.
> "emptiness", or the absence of inherent existence of all phenomena
Phenomena is empty of a permanent self but not empty of existence. The phenomena exists, but that phenomena is not possessed of a permanent unchanging nature.
The Bahiya Sutta has:
> "Then, Bāhiya, you should train yourself thus: In reference to the seen, there will be only the seen."
It's a short article. Maybe the author didn't think they needed to write about the complete, global history of the idea, as opposed to just introducing it using examples they were most familiar with.
Every time I read takes like this I think people forget why big brands exist?
Small business (or “lifestyle business”) vs big brand is often framed about being high quality vs cheap price, because in practice it often is. But in theory the two things are completely unrelated.
Yes fixing stuff is good for the planet. But big brands could offer customer service just fine if people wanted it.
Small business vs big brand is a problem of predictability.
If you have many independent small businesses, NOT all of them will be good. It will be a mixed bag what you get in your area. OP has felt so fortunate with his local highly-skilled asian-owned small business that he felt compelled to write about it on the internet. Not everyone will be this lucky.
And in a world where information travels very fast (this is really the key point) this system is unsustainable, as there are really only 2 options: either people accept the fact that some neighbourhoods are served worse than others, or the take the car and make the travel up to the nice asian shop they read about on the internet, because that’s apparently worth it.
But, surprise, this second option doesn't scale. Because as soon as the nice asian shop goes viral, they realise they can’t keep up with the demand at all. And so they will probably refuse lots of customers. (Note, I’m not even considering the option they might increase prices)
In this sense, the derogatory “lifestyle business” comment makes sense, since I think it’s meant to highlight how elitist it is. It doesn't scale in the sense that it creates a race for who is able to cop the best option.
When I need a sofa, I want to be able to “just” get a sofa. Simple and predictable. If the sofa is good quality, even better.
> Yes fixing stuff is good for the planet. But big brands could offer customer service just fine if people wanted it.
They could but they won't. Because they realized (and probably all agreed) that it's far more profitable to sell the customer a new product, rather than fixing the old one. And since they are the big brands, and they have practically a monopoly (think about big tech companies) they make the rules. They even have the power to sue the crap to which small business tries to fix their products, like Apple did multiple times.
The repair culture is not something sustainable for a big business, that to stay in the market has to increase year after year their sales, and the only way to do so is... making consumers buy new products, even if they don't need them. How to do so? Decrease the quality of the products, make them impossible to repair. No big business would stay alive if they sold you a couch that lasts a century.
I must have hallucinated that I went to an auto dealer for warranty service last week.
Enterprise software companies have consulting and support services.
At the consumer level though, most people aren't willing to pay for the cost of the manufacturer or retailer repairing clothing and other relatively low cost items. As in this article, there are local businesses that do such things but it can be really hard to justify for lower cost items. I have had shoes resoled and otherwise repaired but haven't done it in years and probably most recently a pair of very expensive custom hiking boots that were made to be repairable. (And the repair was probably $200 or so.)
While i do take your point, i think a big problem with discussions about "big businesses" is that they are completely different most of the time. A sofa =/= a car =/= enterprise software. But then how do discussions happen? A sofa is just a thing to sit on at home, you undoubtedly have other things to sit on. Needing to repair a sofa is not catastrophic to survival. Needing to repair a family car can be catastrophic to an individual or family. Needing to repair enterprise software can be catastrophic to a large business itself. There's hugely different consequence scales here, which i guess correlates with how willing a "large" company is to provide the desired support
At the consumer level it frequently doesn't make sense to repair an item. My son had a part fail on his luggage last winter. It might be covered under warranty (the manufacturer wouldn't commit until inspecting the product) but the repair would require shipping the suitcase to them and paying for return shipping. It was going to cost about $150 minimum to have it repaired on a piece that is already a decade old and could be replaced for $200 on sale. I have seen this repeated many times across products.
I've had minor clothing repairs/alterations done at a local dry cleaner for a fairly nominal sum. (Maybe $10-15) But if you can't just easily do something yourself, yeah, you tend to be looking at a floor of at least $100 and at least a certain amount of hassle.
Things I might have taken in to be repaired 25 years ago like a laser printer just don't make sense to do so today.
You don't even have to go the conspiratorial route to realize that repair doesn't make sense to big businesses. The cost of diagnosing the problem, performing the repair, and validating the repair is fairly high. It is also difficult to ensure consistency in the quality of repairs. Then you have to consider that they think about things on a large scale, while repair is an individualized thing. Just look at how computers are repaired. The actual defective component may cost pennies, yet an entire module is replaced. It's not necessarily because the module is impossible to repair. It's because repair processes are difficult to standardize, the cost of replacing the module may be lower than repairing it, and consistent outcomes are difficult to ensure.
Then there is dealing with the customer. A lot of people like to know how much a repair will cost. You can offer an accurate quote when replacing an entire module. A lot of people cannot understand bills that are $0.05 parts + $100.00 labour, so they feel ripped off. A lot of people cannot understand why a repaired product would exhibit problems when it is returned to them (e.g. there was an independent undiagnosed problem).
I think it depends on the business. Maybe fixing 20 year old electronics might be hard, but it still work for (overengineered? underchanged?) herman miller aeron chairs.
Though even Aeron chairs, I had one fixed under warranty after 12 or so years. Then it pretty much fell apart after another 12 or so years. It might have been fixable again but it didn't seem worthwhile to faff around especially mid-COVID and I'm glad I just got a new one which was rock solid. (At the time I wished I could try a couple other models but I ended up being very happy just getting a direct replacement.)
My backpack wore through the back of my 2 year old Patagonia down jacket. They have a repair program, and fixed the jacket for free, didn't even pay shipping. So some large companies actually do this.
In previous centuries you'd find out sofas are so expensive they're out of your reach, might settle for rustic chairs. Cheap and all, what you have in your house would belong in a palace instead of a peasant home. Farther back, your home would have more riches and tools than entire tribes.
The point of the article is that the big business model is "continued growth", which depends on constantly increasing sales, which means products necessarily get shittier so that they must be replaced more frequently. Small "lifestyle" businesses do not operate under this principle and encourage reuse and renewal. They represent opposing philosophies.
Whether you can "just" get a sofa from big business or not, that is precisely what they hope for, and ideally you should be purchasing a sofa more frequently than you already do to further support this notion.
'The point of the article is that the big business model is "continued growth", which depends on constantly increasing sales, which means products necessarily get shittier so that they must be replaced more frequently.' - Sorry to interject here, but this is extremely wrong and nowhere did I find this take-away from the posted article. There are massive businesses that do sell extremely high-quality products - in fact, Japan went through a transition where their businesses went from producing absolute junk (i.e. just like the stuff we import from China today) to producing extremely high quality products (see Juran, Crosby, and cost of quality measures etc...). The key point of the article is that consumers today choose low-priced products since the market gives it to them. If you allow a person to buy a $800 sofa which looks great on the outside and is made in China albeit with extremely low quality materials vs. a sofa which looks almost exactly the same but is priced at $1500 but is of much higher quality - most consumers will obviously choose the $800 dollar sofa vs the $1500 since that's how the free-market functions. Is this rational though?
Well - the consumer will need to buy 4 of the $800 dollar sofas just from having to replace them throughout a 20 year period vs. having the ability to buy one (the $1500) one but that's not obvious to the consumer and it's not clear how to even make this type of judgment. Which sofa really costs the most to you given the information I just provided? The high-quality $1500 one or the $800 dollar one? To a rational person having all of the above information - the more costly one is cheaper - but to an average consumer not having this information the clearly cheaply made product is the better choice. People also are prone to more short-term thinking in many societies which also doesn't help things but the takeaway in general which you posted there is very wrong: mass production and scale usually result in higher-quality products not lower quality ones.
The problem isn't just short-term thinking. You alluded to another point yourself:
> Looks great on the outside
Consumers aren't often equipped to evaluate quality, partly due to skill issues, partly because the corners are cut in places which are hard to spot before purchase. Price doesn't work as a discriminating factor (except to filter out a portion of the worst inventory) because of the number of brands explicitly trying to pass off junk as high-quality luxuries.
If you really can't tell which one is better, and you're as likely to get scammed buying something expensive, why not put less money on the line for something that has a chance of being good enough?
> Consumers aren't often equipped to evaluate quality, partly due to skill
issues
Maybe, maybe not. But regardless, in parallel to this, we have corporations whose modus operandi is at worst to lie to consumers about quality, and at best to mislead. And not just about specific products, but about quality as a general concept.
E.g., with the right training you can tell the difference between the sort of particle board flooring that will balloon up and be destroyed when a drop of water lands on it, the sort of particle board flooring that resists minor water infiltration, and various grades of "real" floors, but most people don't have that training.
Separately (and I _think_ my messaging was clear about this -- talking about the corners being cut being ones that are hard to discover and describing the companies doing that shit as scammers), yes I totally agree; corporations are absolutely not passive participants in consumers being unable to make educated decisions. Even major brands will actively defraud consumers (e.g., Garmin revoking a bunch of lifetime licenses on Navionics software and trying to whitewash public opinion by claiming it was for the customers' own good, or Atlassian blatantly ignoring the CCPA because it's a fairly toothless law), and there exists a plethora of maybe-legal-but-obviously-wrong behavior from most successful companies, including but not limited to "lies, or at best misleadings."
This is a key factor: businesses have learned to “optimize” by cutting corners wherever people can’t perceive the deficit. The product just has to last long enough.
In the off-chance this is a good place to ask, how does a consumer (or, ideally, the existing governing body) fight back? I'll briefly walk through a hypothetical scenario to have something concrete to talk about, then ask about the normal alternatives?
Say you have an IoT device. It's marketed as a device capable of doing a task (e.g., scanning car OBD codes). That task can be done offline. The device initially does that task offline. The app had a backdoor, and the owning company used that backdoor to force logins on previously happy users. Later, they restrict functionality-which-could-be-completely-offline-and-used-to-work to people who pay for a monthly subscription, or maybe they go out of business or otherwise just decide to shut down the servers (see the recent Spotify debacle).
With that backdrop:
- The ToS usually ban class actions and require arbitration.
- The fraud in question is on the order of $20-$200 -- not worth being pursued for most people.
- The ToS are somehow magically invoked when you buy the product, regardless of whether you even saw a warning message suggesting that there might exist a legal agreement which you should read.
The usual outcomes are (1) you get a default judgement and are unable to exercise it because the company goes bankrupt or does some sort of shenanigan which requires a lawyer costing more than the damage in question (a common solution is spinning off a subsidiary owning all the bad debt and responsibilities, keeping the assets elsewhere, kind of like what Johnson and Johnson tried after the talc/cancer debacle), (2) despite the company's best efforts you get a class-action judgement, and the company settles for much less harm than they inflicted, happily pocketing the difference, (3) other more complicated and/or less desirable situations.
What does an individual do to limit their liability in a world where that sort of fraud seems to be condoned, and what options do we have as a society to reduce the overall problem?
It seems hopeless to me. Clearly the company restructuring to separate assets from liabilities should be criminalized and corporate directors / executives should be subjected to criminal prosecutions but that very rarely happens.
The entire concept of corporations as a shield from liability is actually really dubious in my opinion.
> If you allow a person to buy a $800 sofa which looks great on the outside and is made in China albeit with extremely low quality materials vs. a sofa which looks almost exactly the same but is priced at $1500 but is of much higher quality - most consumers will obviously choose the $800 dollar sofa vs the $1500 since that's how the free-market functions. Is this rational though?
This is rarely the choice though. In my experience, the choices tend to be the $800 low quality sofa, the $3000 low quality but with a name brand sofa, and the $6000 low quality but with an even fancier name brand sofa.
Presumably there are some manufacturers that still produce furniture that's actually made of massive wood rather than cardboard and veneer, but it's becoming increasingly rare.
I don't disagree with you - but this also explains the determining factor in which product wins and why today's markets or sofas are lower quality than they were 30+ years ago. If price doesn't matter - what's going to be the driving force in buying behavior? Consumer behavior in other words is no longer driven by quality or long-term cost: today, people will simply choose the lowest cost items and deal with the pain of having to replace it every X years. This drives the market to place a premium on what then? LOWEST COST. Lowest cost = the manufacturers that cut corners and reduce quality, so the market driving force (consumers) lead to a game where the lowest cost producers win and thus saturate the marketplace with junk.
I bought a sofa from a local builder a few years ago for around $2500. The frame is well built but the cushions lost their original shape within 6 months. All told, I'd rather have the sturdy sofa that looks a bit sloppy over a sofa that will break if more than three friends sit on it but I'd really rather have a sturdy one that still looks great after five years.
There are loads of manufacturers that still do this. Go to any furniture row, you’ll see the Ikea parking lot is full, the rc willey parking lot less so, and the premier quality furniture brands parking lots nearly empty.
The problem is that sofas haven’t come down significantly in price. These shittier products aren’t actually much cheaper than, for example, a custom made sofa. I know this because I just bought a custom made sofa.
Yes but this is once again where the free-markets and economics come in: if the mass-produced ones match the custom-made ones in price, consumers will start switching to custom-made sofas. The mass-production suppliers will either have to 1) lower the price of their sofas or 2) increase the quality of production to match the quality of the custom-made ones. Both 1 and 2 are great for consumers and this is why competition is so great :). Notice that all of this is driven by the choices the CONSUMER (me and you) make.
This is a very theoretical argument but I don't think it happens that way in reality, because of all the ways that real human beings are not economically perfect agents. Particularly the information asymmetries - it's much harder to gauge the reputation of a small business vs. a big one.
That’s not quite true, because many people don’t have good custom sofa makers near them. Moreover, because they are small businesses, they don’t have the same capacity for marketing as big brands; most of their business is word of mouth.
Also, custom sofas take time to build - not much, but 2-4 weeks or so.
So there are a lot of reasons big brands are convenient. But you pay for that convenience in quality.
Not sure about sofas but when I bought a bed and some bookcases and nightstands last year all from national retailers, the lead time was 4-12 weeks depending on the product. Getting a custom sofa in 2-4 weeks would beat the competition in many cases. Again, the market constraint lies in knowing about the small vendor in the first place and having a way to purchase it conveniently.
> Well - the consumer will need to buy 4 of the $800 dollar sofas just from having to replace them throughout a 20 year period vs. having the ability to buy one (the $1500) one
Even an $800 sofa will probably last much longer than five years. An added benefit to disposable items is that when you move and/or your tastes change, it's easier to abandon the old sofa and buy a new one than to take the old one with you. (With a little luck and perseverance you may even be able to recoup a few bucks by selling the old one.)
Just because you want to "'just' get a sofa" it does not follow that you "'just' get some food" as well. Or at least not always: sometimes you may 'just' want to, and sometimes you'll want something more that 'just' calories.
And you may not care about sofas as compared to other things: you may 'just' want a sofa, but if you're really into cooking then you may want more than (say) 'just' some random knife, perhaps going for hand-forge Japanese steel.
Further, the cost of making a mistake with food (a few (dozen) dollars) versus a mistake with a sofa (hundreds/thousands) are on two different levels.
Eh. I have terrible taste, so I've eaten McDonalds around the US as well as in Paris and Bangalore.
Within the US, yeah, it's very consistent. I've not seen much variation, other than the one or two specially decorated locations and the menu is very consistent.
Internationally, the branding is very consistent, but the product isn't that consistent. The fries in Paris were very different (and not very good; my feeling is they probably used the same procedure but very different potatoes), but the burgers were pretty similar. In India, they don't serve burgers, but at least when I was there, they did have the delicious old school chicken nuggets that they replaced with 'all white-meat' bleh nuggets in the US. I didn't try any of their chicken sandwiches, because why when I could have the nuggets of my youth?
I understand there's significant regional differences in all the territories they operate in.
They purposely tailor items to the country they're in, and even add entirely new items. For example, the fries in Country X might be cooked with different ingredients or in a different manner, in an attempt to satisfy consumers in that demographic. You'd better believe they could serve US-style fries wherever they wanted, if they wanted. The reason they don't is not because they can't but because they don't want to.
It is remarkably consistent, tourists go to McD's very often when short on time because they know almost exactly what they’re getting, even though they might be a 10h flight from their home.
> If you have many independent small businesses, NOT all of them will be good. It will be a mixed bag what you get in your area. OP has felt so fortunate with his local highly-skilled asian-owned small business that he felt compelled to write about it on the internet. Not everyone will be this lucky.
I don't really agree with this. Unlike big corporations that stick around despite bad service, small businesses have a reputation to maintain. The only bad small businesses that can remain long enough are ones in high density, high turn-over areas (ie: high tourism, fast changing populations like NYC). For sleepier places, these businesses rely on word of mouth and repeat customers.
The reason why people buy a sofa from a generic corp rather than this kind of shop is price.
> ... either people accept the fact that some neighbourhoods are served worse than others, or the take the car and make the travel up to the nice asian shop they read about on the internet, because that’s apparently worth it.
What's better: having to do a few minutes of research to find a good sofa repair shop in your city or having to buy a new sofa every 5 years?
Further: what's better for you personally, and what's better for the planet? Are they compatible?
> But, surprise, this second option doesn't scale.
Why is it important for every type of business to scale? Is "scale" a virtue we must judge every business by?
> We can’t use common sense because it doesn’t fit on a form.
> We can’t use a simple approach to solve the problem in front of us unless the same approach would also work on a problem 100x larger that we may never have.
> If the smart thing to do doesn’t scale, maybe we shouldn’t scale.
When the sofa refurbisher can only handle 100 sofas a year, the 101st customer doesn't have any where to go. Perhaps the market will then lead a second refurbisher to set up shop but that only moves the constraint somewhere else in the supply chain. By its very nature, these small shops can never serve "everyone" the way the big box retailers and flat pack builders can. It's not really a solution to the problem at hand.
Aren't those "constrants" and having people set up shop to solve them literally the most important and critical basic block of our western economies and are critical for social wellbeing? Why do you keep trying to paint this as a negative in response to essentially command economy the monopolies create?
> But, surprise, this second option doesn't scale. Because as soon as the nice asian shop goes viral, they realise they can’t keep up with the demand at all. And so they will probably refuse lots of customers. (Note, I’m not even considering the option they might increase prices)
I've always wondered -- who cares if you hit a scaling limit? Why can't you just coast on huge demand, refuse some customers, and have a sustainable business that's founded on trust in the quality of your product?
So, how is that story going to help me? Apparently, I can't buy a sofa of this quality any more, and if I want it fixed, I apparently have to go to Canada.
You can actually buy sofas of quality. The easiest way is to go to the local design center (most cities have one and if you’re not in a city you can drive to the closest and visit) and you’ll often find many retailers selling high quality furniture. It’s the stuff that’s kind of expensive but not so heavily styled as to incur a crazy premium just for looking expensive. You will be able to see it as it’ll look like Tim’s sofa but costs 2x or more what said sofa would cost on Wayfair. They’re often but not always made in the North Carolina region stateside, other locales seem to be Ohio and Pennsylvania.
We bought such a sofa per advice from a friend that’s an interior designer, and it’s amazing. At 10 years it looks like it was brand new and has withstood the first 10 years of baby life including playdates and kids drawing on it, etc (we got it with a special treatment to make it not absorb such things and it actually worked). Kids jumping off the back frame, throwing all the cushions around, etc. Literally unblemished and the internal frame is rock solid.
But also the single most expensive piece of furniture I’ll ever buy. I’ll never need to buy a replacement for it though. I expect to be using it for the rest of my life and passing it onto my descendants.
Read the Dwell article referenced in the article to learn that the whole ecosystem of the North Carolina furniture industry is dying out rapidly due to the onslaught of cheap, light, shippable, assemble-at-destination flatpack furniture.
We have a 20-year-old quality sofa from a major NC company that we got reupholstered by them last year, just before they went out of business.
Yeah. I expect only the best craftsman paired with the best businessmen will survive. But there’s a decent market among the affluent for quality furniture and most commercial furniture for high traffic environments demand pretty high end and durable stuff.
A key problem is that when people make purchasing decisions, price ranks extremely high on the priority list, even when it will be costlier and worse for the consumer in the long run. Capitalism has found a thousand ways to exploit that inherent trait we all share, and we have to work damn hard to counteract it—-and most people won’t even know they should be making that effort.
A lifestyle business isn’t elitist, nor necessarily for elitist customers. It is in most people’s interest to invest in quality, but not everyone can afford it and even among those that do, the final price tag has an undue weight in the equation. (Not to mention that big brands are removing quality as an option even in the higher price ranges)
> Every time I read takes like this I think people forget why big brands exist?
I think the article made a great point why big brands exist -- to deliver on the promise of unbridled growth, often leading to enshittification.
> But big brands could offer customer service just fine if people wanted it.
The experience suggests that they usually offload that to a third-party vendor to cut costs and we all know that does not track as good as small, family owned, locally sourced, your trustworthy shop.
> But, surprise, this second option doesn't scale.
True. Probably does not have to. A sufficiently wide distribution of such businesses is just as good.
>when i need a sofa, i want to be able to “just” get a sofa
Which is understandable, and also the whole problem. Everything that used to go along with getting that sofa, like the human interaction, is thrown out in the name of efficiency, and eventually we all end up locked in our houses with nowhere to go but our jobs.
It's not just efficiency, the very last thing I want to do when buying any good or service is talk to another human being. Hell if self checkout is any indication people will trade efficiency for not having to talk to a person.
Eventually we'll all end up conserving our social battery for friends and loved ones rather than work and shopping.
Not from grocery store interactions that's for damn sure.
My friend family tree is messy as hell but I can loosely bucket them in people I met in high school or college, people I met at parties, people I met at raves, people I met at shared interest gatherings — skate park, community garden, clubs, people I met at rallies/protests, people I met online, and then all of their friends and partners, and their friends of friends, etc..
I get that it's unbelievably hard to bootstrap from literally zero but like, the cashier at the gas station really doesn't feel like it's the play unless you live in a tiiiiny ass town.
I wonder if we are (again) just carrying over some preconceptions about how humans do things, and instead Agi will come from a completely different direction..
Wow.. Really great work, finally someone is doing it!
Since I've thought about this for a long time (I've actually even made a very simplified version last year [1]), I want to contribute a few thoughts:
- cool that you have a Vscode extension, but I was a little disappointed that it opens a full browser view instead of using the existing, good Notebook interface of Vscode. (I get you want to show the whole Frontend- But I'd love to be able to run the Reactive Kernel within the full Vscode ecosystem.. Included Github Copilot is cool, but that's not all)
- As other comments said, if you want to go for reproducibility, the part about Package Management is very important. And it's also mostly solved, with Poetry etc...
- If you want to go for easy deployment of the NB code to Production, another very cool feature would be to extract (as a script) all the code needed to produce a given cell of output! This should be very easy since you already have the DAG.. It actually even existed at some point in VSCode Python extension, then they removed it
nbgather used static slicing to get all the code necessary to reconstruct some cell. I actually worked with Andrew Head (original nbgather author) and Shreya Shankar to implement something similar in ipyflow (but with dynamic slicing and a not-as-nice interface): https://github.com/ipyflow/ipyflow?tab=readme-ov-file#state-...
I have no doubt something like this will make its way into marimo's roadmap at some point :)
Unfortunately, the obvious economic incentive to effectively paywalling code through the Internet, is so strong that on this very website (that was supposed to have hacker mentality) most of the comments are in favour of it, because “you have to make a living”…
If you have any feasible ideas that could plausibly work where people don't "have to make a living" so that we can find ourselves and not be lost, we're all ears. Until then we're stuck in a place where people need money to pay for frivolous things like "food" or "rent" or "transportation". Still, in the face of that, there's a site called GitHub where people do freely share code, despite the economic misincentives, so I think the kids will be alright.
I wouldn't be so quick to conspiracy. I'm the author of a work and a famous blog post that trains a particular common architecture much faster (don't want to dox myself too much) and with far fewer parameters, but it has been rejected several times and is now arxiv only. Our most common complaint was "who would use this? Why not just take a large model and tune it?" That question alone held us back a year (had over a hundred citations by then and remains my most cited work) until it switched to "use more datasets" and "not novel" (by that time true, others had built off of us, cited us, and published in top venues).
I don't think this was some conspiracy by big labs to push back against us (we're nobodies) but rather that people get caught up in hype and reviewers are lazy and incentivized to reject. You're trained to be critical of works and especially consider that post hoc most solutions appear far simpler than they actually are. But context matters because if you don't approach every paper with nuance it's easy to say "oh, it's just x." But if those ideas were so simple and obvious they would also be prolific. I see a lot of small labs suffer the same fate simply due to lack of compute. If you don't make your new technique work on many datasets it becomes the easiest thing to reject a paper by. ACs aren't checking that reviews are reasonable. I've even argued with fellow reviewers about papers in workshops -- papers I would have accepted in the main conference -- that are brushed off and the reviewers admit in their reviews that they do not work on these topics. I don't understand what's going on but at times it feels like a collective madness. A 10 page paper with 4 very different datasets that solves a problem, is clearly written, has no major flaws, and is useful to the community should not need defending when submitted to a workshop just because reviewers aren't qualified to review the work (this paper got in btw). We are moving into a "pay to play" ecosystem and that will only create bad science due to group think. (another aspect of "pay to play" is in the tuning. Spending $1M to tune your model to be the best doesn't mean it is better than a model that could not afford the search. Often more than half of resources are spent on tuning now)
Is there a place where you guys discuss... things? I'm layman interested in this topic akin to pop-physics/maths, but have no chance to just read papers and "get it". On the other hand, immediately available resources focus more on how-to part of it rather than on what's up overall. Also, do you have something like 3b1b/pbs/nph for it? Content that you can watch and say "well, yep, good job".
I don't have any great recommendations and unfortunately my advice may be not what you want to hear. What I tell my students is "You don't need to know math to build good models, but you need to know math to know why your models are wrong." But this is even a contentious statement within the community. (Personally I'm more interested in exploring what we can build and understand rather than focusing on throwing more compute and data at problems. There's a lot of work to be done that does not require significant compute, but it isn't flashy and you'll get little fame. Every famous model you know has some unsung hero(s) who built the foundation before compute was thrown at the problem). I was previously a physicist and we similarly frequently express that you do not know the material unless you can do the math. Physicists are trained in generating analogies as they help communication but this sometimes leads to people convincing themselves that they understand things far more than they actually do. They say the devil is in the details, and boy are there a lot of details. (Of the science communicators, I'm happy those are the ones you mention though!) But do not take this as gatekeeping! These groups are often happy to help with the math and recommend readings. ML is kinda a while west and you can honestly pick a subdomain of math and probably find it useful, but I would start by making sure you have a foundation in multivariate calculus and linear algebra.
As to paper reading, my suggestion is to just start. This is a fear I faced when I began grad school and it feels overwhelming and like everyone is leagues ahead of you and you have no idea where to begin. I promise that is not the case. Start anywhere, it is okay, as where you end up will not matter too much on where you begin. Mentors help, but they aren't necessary if you have dedication. As you read you will become accustomed to the language and start to understand the "lore." I highly suggest following topics you find interesting backwards through time, as this has been one of the most beneficial practices in my learning. I still find revisiting some old works reveals many hidden gems that were forgotten. Plus, they'll be easier to read! Yes, you will have to reread many of those works later, as you mature your knowledge, but that is not a bad thing. You will come with newer eyes. Your goal should be to first understand the motivation/lore, so do not worry if you do not understand all the details. You will learn a lot through immersion. It is perfectly okay if you barely understand a work when first starting because a mistake many people make (including a lot of researchers!) is that a paper is not and cannot be self contained. You cannot truthfully read a work without understanding its history and that only comes with time and experience. Never forget this aspect; it is all too easy to deceive yourself that things are simpler than they are (the curse of hindsight).
I'd also suggest to just get building. To learn physics you must do physics problems. To learn ML you must build ML systems. There are no shortcuts but progress is faster than it looks. There's hundreds of tutorials out there and most are absolute garbage but I also don't have something I can point to that's comprehensive. Just keep in mind that you're always learning and so are the people writing tutorials. I'm going to kinda just dump some links, they aren't in any particular order sorry haha. Its far from comprehensive, but this should help you getting started, nothing in here is too advanced. If it looks complicated, spend more time, you'll get it. It's normal if it doesn't click right away and there's nothing wrong with that.
Unless they were very confident of acceptance, a top research prof would rewrite and resubmit before publishing on arxiv so that others could "build on it" (scoop you at a top conference).
Welcome to ML. And idk, I'd feel pretty confident that a paper that gets so many citations gets accepted. The review system is like a slot machine if you aren't a big tech lab
They certainly have an incentive to keep these kinds of improvements in-house and not publish them, since they are commercial entities and this represents a competitive advantage.
Nvidia can't make GPUs fast enough. I doubt 10xing training and/or inference efficiency would result in a decrease in demand. I would be surprised if it didn't instead increase demand. Mind you, Nvidia is pushing hard on TensorRT which optimizes models at inference time and results in major increases in throughput (not 10x though lol).
But if things get too efficient for individual users, you won't need an Nvidia GPU anymore. People will use cheaper hardware instead. I'm looking forward to running good models at decent speed on a low-end CPU or whatever crappy GPU is in my phone.
I had the same thought this morning and was debating selling my nvda stock when I saw this - feels like they are well-positioned right now, as with crypto a few years ago, but if there were an efficiency breakthrough that allowed commodity CPUs to do the inference instead, this advantage could vanish quickly.
It hits so many right sposts. Thanks for sharing it