The recent post criticising Free Software advocates for advocating user-modifiable software and then being annoyed at LLMs annoys me and the reason is best illustrated by this analogy:
-
@raymaccarthy @david_chisnall And I'd argue we do need arguments against LLMs that do not hinge on them being useless garbage, because improvement is happening and a lot of people are already claiming they increase their productivity. I disagree with them - but they firmly believe that, and the "LLMs are useless garbage" argument IS NOT going to get through to them.
@raymaccarthy @david_chisnall incide tally, i tried to order an Uber for the first time ever yesterday.
"tried to" being the operative word.
-
@ermo @david_chisnall how come what? Me finding it weird that FOSS people see excluding non-human entities, that were the reason why FOSS movement started, to be incompatible with their goals?
-
The recent post criticising Free Software advocates for advocating user-modifiable software and then being annoyed at LLMs annoys me and the reason is best illustrated by this analogy:
Public-transport advocates spend years advocating for a connected public-transport infrastructure, where it’s easy to take a small combination of busses, metros, trams, and trains to get from anywhere to anywhere. The network would be efficient and operated as a non-profit-making public good, making individual movement cheap (or, ideally, free). They work with municipalities to build out some of this infrastructure, persuade national governments to invest in the longer routes, and so on.
Someone comes along with a massive subsidy for a handful of private taxi companies to hire a bunch of drivers and give free (paid for by investors) ride to everyone. The drivers are immigrants who don’t speak the language very well, which is great for the taxi companies because they are easy to exploit (they are, in fact, underpaid and put in dangerous situations routinely). The owners of the taxis are pocketing a load of investor money for every ride though.
When you get in one of these taxis, there’s a 90% chance they’ll take you where you want, a 9% chance they’ll take you somewhere nearby, and a 1% chance they’ll just drop you off in a dangerous part of town. A bunch of people are mugged and a few more murdered as a result of this, but the companies aren’t liable. The investors behind this tell everyone ‘don’t bother learning to drive, there’s no point, our taxis will take you anywhere, for much less money!’. At the same time, ridership on existing public transport drops off, leading to calls to cut its funding and there are mass redundancies for bus drivers and so on. The taxis are all diesel and heavily polluting, leading to worse air quality everywhere they go. To make sure that they can pick people up easily, the ones not actively giving rides are constantly circulating, placing huge strain on road infrastructure and further increasing pollution.
And then someone says to those public-transport advocates: ‘this is what you wanted, why are you unhappy just because it’s not delivered in the way you imagined?’
@david_chisnall The fact that this actually happened still boggles my mind! -
@ermo @david_chisnall how come what? Me finding it weird that FOSS people see excluding non-human entities, that were the reason why FOSS movement started, to be incompatible with their goals?
The problem is that it's very difficult to exclude corporations in a way that doesn't exclude people, or cause other harms to people.
Let's start with the second one. Imagine you have a perfect license that exactly excludes corporations, but doesn't exclude individuals. You release a replacement for MS Office with it. But now you need to interoperate with corporations and they are precluded from using your code, so they are now incentivised to keep sending you MS Office documents. Your program now has to perfectly interoperate with MS Office, or normal people need to buy MS Office. Fortunately, LibreOffice and friends didn't pick a license that excludes corporations and so governments are now in a position to mandate ODF for interoperability and say 'just use LibreOffice' if a corporation complains that MS Office doesn't import it properly.
But excluding corporations is itself hard. If I'm a sole trader, I'm presumably not excluded even if I use the program for work, but if I collaborate with another person in a partnership are we now both excluded? If I work for a community interest corporation such as lowRISC, am I excluded? What about an animal sanctuary? If you exclude, say, companies worth over a billion dollars (hard, because you'll need to keep adjusting for inflation), what stops employees of such a company from using the program on their own time? Or simply subcontracting work that needs it to a smaller company. You cannot craft a license that specifically excludes the people you want to, so you end up with one of two things:
- Some people are accidentally excluded, which harms them.
- Some corporations are accidentally included, which gives them a competitive advantage and skews markets to favour the very groups you were trying to harm.
-
The recent post criticising Free Software advocates for advocating user-modifiable software and then being annoyed at LLMs annoys me and the reason is best illustrated by this analogy:
Public-transport advocates spend years advocating for a connected public-transport infrastructure, where it’s easy to take a small combination of busses, metros, trams, and trains to get from anywhere to anywhere. The network would be efficient and operated as a non-profit-making public good, making individual movement cheap (or, ideally, free). They work with municipalities to build out some of this infrastructure, persuade national governments to invest in the longer routes, and so on.
Someone comes along with a massive subsidy for a handful of private taxi companies to hire a bunch of drivers and give free (paid for by investors) ride to everyone. The drivers are immigrants who don’t speak the language very well, which is great for the taxi companies because they are easy to exploit (they are, in fact, underpaid and put in dangerous situations routinely). The owners of the taxis are pocketing a load of investor money for every ride though.
When you get in one of these taxis, there’s a 90% chance they’ll take you where you want, a 9% chance they’ll take you somewhere nearby, and a 1% chance they’ll just drop you off in a dangerous part of town. A bunch of people are mugged and a few more murdered as a result of this, but the companies aren’t liable. The investors behind this tell everyone ‘don’t bother learning to drive, there’s no point, our taxis will take you anywhere, for much less money!’. At the same time, ridership on existing public transport drops off, leading to calls to cut its funding and there are mass redundancies for bus drivers and so on. The taxis are all diesel and heavily polluting, leading to worse air quality everywhere they go. To make sure that they can pick people up easily, the ones not actively giving rides are constantly circulating, placing huge strain on road infrastructure and further increasing pollution.
And then someone says to those public-transport advocates: ‘this is what you wanted, why are you unhappy just because it’s not delivered in the way you imagined?’
That's ridiculous, there's no way any municipality would fall for such a transparent plot to loot public coffers for private gain.
It would be almost as stupid as running all healthcare through for-profit private insurance companies and saying it's to keep costs low.
-
The recent post criticising Free Software advocates for advocating user-modifiable software and then being annoyed at LLMs annoys me and the reason is best illustrated by this analogy:
Public-transport advocates spend years advocating for a connected public-transport infrastructure, where it’s easy to take a small combination of busses, metros, trams, and trains to get from anywhere to anywhere. The network would be efficient and operated as a non-profit-making public good, making individual movement cheap (or, ideally, free). They work with municipalities to build out some of this infrastructure, persuade national governments to invest in the longer routes, and so on.
Someone comes along with a massive subsidy for a handful of private taxi companies to hire a bunch of drivers and give free (paid for by investors) ride to everyone. The drivers are immigrants who don’t speak the language very well, which is great for the taxi companies because they are easy to exploit (they are, in fact, underpaid and put in dangerous situations routinely). The owners of the taxis are pocketing a load of investor money for every ride though.
When you get in one of these taxis, there’s a 90% chance they’ll take you where you want, a 9% chance they’ll take you somewhere nearby, and a 1% chance they’ll just drop you off in a dangerous part of town. A bunch of people are mugged and a few more murdered as a result of this, but the companies aren’t liable. The investors behind this tell everyone ‘don’t bother learning to drive, there’s no point, our taxis will take you anywhere, for much less money!’. At the same time, ridership on existing public transport drops off, leading to calls to cut its funding and there are mass redundancies for bus drivers and so on. The taxis are all diesel and heavily polluting, leading to worse air quality everywhere they go. To make sure that they can pick people up easily, the ones not actively giving rides are constantly circulating, placing huge strain on road infrastructure and further increasing pollution.
And then someone says to those public-transport advocates: ‘this is what you wanted, why are you unhappy just because it’s not delivered in the way you imagined?’
@david_chisnall As someone who has been an enthusiast and contributor to open source since before it had that name:
We have never delivered on the idea of user-modifiable software. We have sometimes freed software builders from unnecessary toll booths and restrictions. In other words, we built something that works for us and stopped.
I think your analogy of literal crazy taxis is good when talking about LLMs for noncoders. But there was no actual product from FL/OSS world that it’s displacing
-
@david_chisnall As someone who has been an enthusiast and contributor to open source since before it had that name:
We have never delivered on the idea of user-modifiable software. We have sometimes freed software builders from unnecessary toll booths and restrictions. In other words, we built something that works for us and stopped.
I think your analogy of literal crazy taxis is good when talking about LLMs for noncoders. But there was no actual product from FL/OSS world that it’s displacing
@david_chisnall Maybe a better analogy: imagine we only had trains and last-mile delivery on horse carts
For 30 years, nerdy rail engineers (on their days off) have tinkered with door-to-door rail networks. They build impressive examples in their back yard – tech that becomes light rail, and is absorbed by existing rail companies
Buses and cars are invented and solve the last-mile problems. They crash due to flaws and operator error all the time.
The backyard rail engineers are angry.
-
The recent post criticising Free Software advocates for advocating user-modifiable software and then being annoyed at LLMs annoys me and the reason is best illustrated by this analogy:
Public-transport advocates spend years advocating for a connected public-transport infrastructure, where it’s easy to take a small combination of busses, metros, trams, and trains to get from anywhere to anywhere. The network would be efficient and operated as a non-profit-making public good, making individual movement cheap (or, ideally, free). They work with municipalities to build out some of this infrastructure, persuade national governments to invest in the longer routes, and so on.
Someone comes along with a massive subsidy for a handful of private taxi companies to hire a bunch of drivers and give free (paid for by investors) ride to everyone. The drivers are immigrants who don’t speak the language very well, which is great for the taxi companies because they are easy to exploit (they are, in fact, underpaid and put in dangerous situations routinely). The owners of the taxis are pocketing a load of investor money for every ride though.
When you get in one of these taxis, there’s a 90% chance they’ll take you where you want, a 9% chance they’ll take you somewhere nearby, and a 1% chance they’ll just drop you off in a dangerous part of town. A bunch of people are mugged and a few more murdered as a result of this, but the companies aren’t liable. The investors behind this tell everyone ‘don’t bother learning to drive, there’s no point, our taxis will take you anywhere, for much less money!’. At the same time, ridership on existing public transport drops off, leading to calls to cut its funding and there are mass redundancies for bus drivers and so on. The taxis are all diesel and heavily polluting, leading to worse air quality everywhere they go. To make sure that they can pick people up easily, the ones not actively giving rides are constantly circulating, placing huge strain on road infrastructure and further increasing pollution.
And then someone says to those public-transport advocates: ‘this is what you wanted, why are you unhappy just because it’s not delivered in the way you imagined?’
@david_chisnall "You wanted freedom! This is the price of being free"
Editing to clarify I don't actually believe the sentiment. More the whole thing smacks if dipshits who way that.
-
@shanesemler @david_chisnall And I'll even help the ones who aren't dicks about it when that whole process blows up in their faces.
The "normies" who go and try to make code themselves with AI are the ones who are REALLY getting screwed here. It does it just well enough to make them think that it did what they asked. It'll then make up really stupid excuses why it didn't. Like calling an if/else branch a "rule based system that simulates AI".
The "normies" are going to make a fucking mess.
@crazyeddie @david_chisnall Making a mess is how you learn.
-
@shanesemler @david_chisnall It's malicious compliance. The outcome they see is the one stripped from the context of why buses run by the government are preferable to a private fleet of taxis and contrary to their offence at people not thinking its the same thing, the model that conveniently allows them money and power that would otherwise go to the people also introduces problems that wouldn't exist in the model of what people actually wanted.
While this next question sounds like a gotcha, I do genuinely want to know your answer because it helps me figure out what I'm responding to.
If an author, artist, or musician told you "if you don't like it, write/make it yourself" would you feel the same way if the person responding to that statement told an LLM "write a spinoff of this famous novel for me" or "make a painting kind of like this one, but with the changes the artist said no to"?
-
@crazyeddie @david_chisnall Making a mess is how you learn.
Making a mess in a context where the mess is understandable and you can incrementally improve on it is how you learn.
-
Making a mess in a context where the mess is understandable and you can incrementally improve on it is how you learn.
@david_chisnall @crazyeddie The tools exist, and people will use them. You all can argue about it and seethe into the ether, but people are going to do it, whether pro devs like it or not.
-
@raymaccarthy @david_chisnall And I'd argue we do need arguments against LLMs that do not hinge on them being useless garbage, because improvement is happening and a lot of people are already claiming they increase their productivity. I disagree with them - but they firmly believe that, and the "LLMs are useless garbage" argument IS NOT going to get through to them.
RE: https://mastodon.social/@ratsnakegames/116431740480623276
@ratsnakegames @raymaccarthy @david_chisnall
Eventually the people will understand that the way big tech is approaching AI is wrong, just wait for the bubble to burst.
Yo don't need to use weak arguments they can easily dismantle. There are lots of solid arguments against the way big tech is trying to ride an insane race to make gigantic LLMs. Like for example the expense in water and energy, or the fact they are making gigantic investments and they don't even have a clear business model.
The best of all arguments is that "AGI" or "ASI" (the excuse they use to get investment) is a lie, a children's story as credible as Narnia or the elves of the north pole. Believing LLMs are going to become "superintelligence" just by pouring more compute and data on them is like believing children are going to fly just because they learn to walk and then quickly they learn to run.
-
RE: https://mastodon.social/@ratsnakegames/116431740480623276
@ratsnakegames @raymaccarthy @david_chisnall
Eventually the people will understand that the way big tech is approaching AI is wrong, just wait for the bubble to burst.
Yo don't need to use weak arguments they can easily dismantle. There are lots of solid arguments against the way big tech is trying to ride an insane race to make gigantic LLMs. Like for example the expense in water and energy, or the fact they are making gigantic investments and they don't even have a clear business model.
The best of all arguments is that "AGI" or "ASI" (the excuse they use to get investment) is a lie, a children's story as credible as Narnia or the elves of the north pole. Believing LLMs are going to become "superintelligence" just by pouring more compute and data on them is like believing children are going to fly just because they learn to walk and then quickly they learn to run.
@paelnever @raymaccarthy @david_chisnall "Yo don't need to use weak arguments they can easily dismantle."
I don't think that any of the arguments David makes are weak or easily dismantled - in fact, I think pointing out that all the providers are muscling in on loss leadership so they can start extorting their customers just like Uber does is a very solid argument that actually gets managers to listen.
The argument that "AGI is a lie" will not have pull over people who use LLM codegen.
-
@paelnever @raymaccarthy @david_chisnall "Yo don't need to use weak arguments they can easily dismantle."
I don't think that any of the arguments David makes are weak or easily dismantled - in fact, I think pointing out that all the providers are muscling in on loss leadership so they can start extorting their customers just like Uber does is a very solid argument that actually gets managers to listen.
The argument that "AGI is a lie" will not have pull over people who use LLM codegen.
@paelnever @raymaccarthy @david_chisnall My boss and his bosses do not give a single fuck about terminology. They give a fuck about money and entrepreneurial risk. Arguments aimed at THESE things actually get through to them.
If an argument does not get through to people, it is not the best argument, unless your goal is to be a smug asshole rather than actually convincing people.
-
@paelnever @raymaccarthy @david_chisnall My boss and his bosses do not give a single fuck about terminology. They give a fuck about money and entrepreneurial risk. Arguments aimed at THESE things actually get through to them.
If an argument does not get through to people, it is not the best argument, unless your goal is to be a smug asshole rather than actually convincing people.
@paelnever @raymaccarthy @david_chisnall my employer is heavily pivoting to AI, and they ACTIVELY PREEMPT the criticism about water and energy usage by making token effort to reduce these things, and also shrugging and saying "we don't like it but we have to do it or we'll be left behind".
You have to attack the idea that codegen is beneficial TO THE COMPANY, or you are wasting your time.
-
@paelnever @raymaccarthy @david_chisnall My boss and his bosses do not give a single fuck about terminology. They give a fuck about money and entrepreneurial risk. Arguments aimed at THESE things actually get through to them.
If an argument does not get through to people, it is not the best argument, unless your goal is to be a smug asshole rather than actually convincing people.
@ratsnakegames @raymaccarthy @david_chisnall
If people don't understand arguments that's not a problem of arguments. During decades the system is being dismantling education and turning people into idiots unable to think.You don't "attack the idea that codegen is beneficial to the company" by making a fake analogy about FOSS and LLMs compared to public transport and uber. You DO attack that idea by showing statistics that demonstrate software developers are not being more productive with LLMs. Those papers exist.
-
@ermo @david_chisnall how come what? Me finding it weird that FOSS people see excluding non-human entities, that were the reason why FOSS movement started, to be incompatible with their goals?
@hjvt @ermo @david_chisnall RMS wrote an interesting piece about this topic:
That said, the popular view is that if you want to restrict usage, the only reliable way will be writing proprietary software where you can control who gets it. No one will stop you from doing that.
I'd suggest that the more useful thing to do is to convince other people of the rightness of your predilections and get them to vote. Reshape the world such that people are punished for doing unethical things, rather than rewarded, and you won't have to special-case your software licenses.
-
@paelnever @raymaccarthy @david_chisnall "Yo don't need to use weak arguments they can easily dismantle."
I don't think that any of the arguments David makes are weak or easily dismantled - in fact, I think pointing out that all the providers are muscling in on loss leadership so they can start extorting their customers just like Uber does is a very solid argument that actually gets managers to listen.
The argument that "AGI is a lie" will not have pull over people who use LLM codegen.
@ratsnakegames @raymaccarthy @david_chisnall
I'm sorry but a fake analogy is not a valid argument and will never be.
Most of the people who actually use LLM codegen they DO understand "AGI" is a lie. Maybe is a surprise to you but there are people who understand that all technologies can be used in the right or the wrong way.
-
@ratsnakegames @raymaccarthy @david_chisnall
I'm sorry but a fake analogy is not a valid argument and will never be.
Most of the people who actually use LLM codegen they DO understand "AGI" is a lie. Maybe is a surprise to you but there are people who understand that all technologies can be used in the right or the wrong way.
@paelnever @raymaccarthy @david_chisnall it is not a fake analogy.
I'm gonna block you now because you're tedious.
