Anthropic does a somewhat similar thing. If you visit their ToS (the one for Max/Pro plans) from a European IP address, they replace one section with this:
Non-commercial use only. You agree not to use our Services for any commercial or business purposes and we (and our Providers) have no liability to you for any loss of profit, loss of business, business interruption, or loss of business opportunity.
It's funny that a plan called "Pro" cannot be used professionally.
Ha out of curiosity I loaded that same consumer terms URL on both a USA and a UK VPN exit node - sure enough, the UK terms inject that extra clause you quoted banning commercial usage that is not present for USA users.
There's the usual expected legal boilerplate differences. However, the UK version injects the additional clause at line 134 that has no analog in the US version.
Wow, if you brought a paper contract to court that mutated itself depending which way you look at the paper, I wonder what a judge would think of that?
Personally I would crumple it up and pitch it out the window. I don't know why they can't simply be clear about what clauses apply to which geographies. An IP address should not be assumed as a reliable indicator of the jurisdiction in which an end-user resides. (Eg. In addition of VPNs's and unexpected routing, what happens if you travel?)
I once wrote a contract document in PostScript that changed the wording based on the date. Two parties could cryptographically sign an agreement in the document, which would change when printed on a later date.
One of the reasons we don’t use PostScript so much any more.
It's perfectly normal for contracts in different jurisdictions to use different wording and include different clauses.
Even within the US, employment contracts with the same organisation may contain different wording depending on the state in which the employment is occurring.
> It's perfectly normal for contracts in different jurisdictions to use different wording and include different clauses.
Before signing, yes, but once signed the contract stays constant. Mutating terms of service are weird - I would expect them to be locked to a canonical URL at least, like "https://.../tos?region=eu" or ideally something that locks the version too, like "https://.../tos?version=eu-002".
Let me pose a question from a different angle - these are legal contracts we are talking about, and the version they present to the user apparently changes based only on the client IP address. So if the terms in the EU ToS are better than in the US ToS, what would prevent me from signing up with an EU IP Address the first time? I would expect to be bound to the contract I actually agree to, not just the one they "intended" to show me.
My issue isn't with them having different verbiage for different jurisdictions. It's with the way they sneakily change the verbiage in a manner nobody would expect.
The correct way to do this is to clearly distinguish them e.g. two different contracts, one titled "US Terms of Service" and another "European Terms of Service". Both with static content. Preferably PDF's or some other fixed format which you can download to redline changes when they inevitably tweak it every couple years, and properly print in the event you need to embark on litigation.
Not some "Global Terms of Service that silently change depending on quasi-pseudorandom network stack effects"
How the hell are you supposed to have confidence you're even looking at the right document? Contracts are meant to be clear and unambiguous, this dark pattern works against consensus ad idem.
In the Uk there seem to be separate commercial and consumer terms.
In the UK the consumer terms say its subject to English law and the courts of the UK jurisdiction you live in.
The commercial terms say that in the UK, Switzerland and the EEA there will be binding arbitration by an arbitrator in Ireland appointed by the President of the Law Society of Ireland.
The UK commercial terms explicitly do not apply to individual user plans. The US also has a separate terms sheet for commercial plans.
We are comparing like for like - an individual user using a Claude Pro subscription. A US user can use it for commercial use and be in compliance with the terms, the UK user cannot.
> A US user can use it for commercial use and be in compliance with the terms, the UK user cannot.
But why? My guess is the liability exposure is what they’re trying to control. So you probably can if you’re ok with no liability. It’s still noncompliant to how they wrote it but I would guess it’s the motivation. Unless they really just want to force the UK to pay for all commercial uses, which I suppose is possible.
I think its because the law in the UK limits exclusions of liabilities in consumer contracts far more than in business contracts (in general consumer law has a LOT of protections that do not apply to business contracts). If you look at the clauses excluding liabilities they are very different. I think the same applies to many other countries so they will also have separate consumer contracts.
The law in Australia also has teeth, but visiting the link above just gets me (what seems to be) the US version of the terms without anything around commercial use.
Well, there's your rationale as to why AI cannot replace you.
When sh!t hits the fan, Anthropic will immediately point to this clause. Who knows, maybe a court would see it as valid.
Meanwhile, your customer (and thus, your management) is looking for someone to blame for excrement making contact with the impellers. And that someone's gonna be you.
Well, OpenAI doesn't seem to have clauses like this. Europeans are allowed to use it for commercial purposes under the ToS. (But check it yourself, I'm not a lawyer).
I reimplemented my startup idea from scratch with Codex a few months ago, just for peace of mind.
Honest question, what peace of mind does this give you? If my idea could be implemented from scratch by one of these agentic harnesses I would be concerned about the viability of it more than anything.
But you have limited funds to take in a lawsuit realistically the worst they can do is fire you, it's not like being blameable somehow makes you more valuable.
Employees often make mistakes that cost companies thousands of dollars. And there's no shortage of stories where employees cost companies tens of thousands and millions.
When a construction guy messes up measurements and thousands of dollars of work has the be removed and redone, no one thinks of taking the employee to court. Why would you want to take your Ai to court?
When the construction worker messes up a job that then causes injury or damages the property they absolutely get sued. The state can even get involved if the mistake is deemed criminal negligence.
In your example the owners will often take the construction company or small business owner to court. Most trades people negotiate and redo the work for free or much reduced cost to avoid this.
In office settings if you expose PII you will likely be fired.
I am really losing faith in hacker news intelligence levels or at least reading comprehension.
We were talking about people sueing AI for mistakes.
Employees do not get sued by their employer for mistakes. If your employer wants you to dig a foundation per plan, and you measure it wrong and dig it in the wrong orientation on the lot, you might get fired, but you will not pay the $50k+ to rip out the cement and put a new foundation.
what the hell are you on about? Have you ever been employed? Employees do got reprimanded because of their mistakes. Employers just don't sue via the courts for the same reason you don't sue your spouse first thing when they break a plate. They settle via internal penalties first.
(Not only that, employees who got a reprimand too heavy handed can sue back. Plenty of cases around.)
"AI" company provides a service. They might or might not be adequate, that's not the point, the point is that the ability to sue them must always be on the cards if the agreed upon terms aren't met.
I have no idea what you are talking about. I've been employed my whole adult life and I have never seen an employee get sued ft $50k because his mistake caused the company to lose $50k.
show me any that have claimed that they were for entertainment purposes only. sql server has never had that in its EULA. The GPL does not say that the software is for entertainment purposes only.
The Home/Personal edition of Mathematica is for non-commercial use only although it's a paid subscription. The world around you is not bound by your ignorance.
Will be interesting to see if those still hold any weight (in the US at least) after the latest Meta rulings established defective design as a valid reason to sue big tech for damages
No, it wasn't the government - just a private company running within a state-sponsored accelerator. The government withdrew their funds after finding out about it.
The article says that the water ice could be used as a source of oxygen, needed to power the combustion of hydrocarbons as fuel. But the oxygen is trapped in water molecules, wouldn't it need to be freed from the hydrogen atoms first (which requires energy)?
I have looked it up, and it seems the combustion of methane produces more energy than the combustion of hydrogen. Splitting water into hydrogen and oxygen won't be 100% efficient though, and neither will be turning heat into useful energy, so I don't know if you'll still be energy-positive in the end.
More realistic might be a long-term preparation scenario in which we send a first wave of robots to set up some solar panels and run them to store H2 and O2 for a couple of decades. We won't have gained energy necessarily, but we will have stored it in large enough amounts that early colonisation will have extra energy if it needs it.
In most contexts, an entertainment event being "presented by X" just means that X funded the event in exchange for having their name associated with it. E.g., "Thursday Night Football, Presented by Bud Light" -- Bud Light has nothing to do with the content or transmission of the event, it just paid the most money for advertising.
However: you're totally right that they may do some cool tie-in with the phone itself in this case.
Do you live in Poland? I live in Prague, and here, I'm surprised at how little people care about the foreign investment aspect of the privatization process. Even though almost everyone complains that it is corrupt, people here are mostly pro-west. The economy is strong here, and people don't see a big problem. Sure, we were among the richest countries in Europe pre WWII, but that was a long time ago, and people are now used to being relatively poor. But outside of Prague, especially on the Moravian side of the country I think opinions are very different, and many people actively say that "communism was better" and that the west has robbed the country. It is true, that heavy industry, like the car maker škoda auto (Now owned by VW), and the slevarny (metal making firms that are owned by the west), and the coal mines(owned by the Swiss) are all outside of Prague... Do Pols realise the extent that they were robbed?
I just realised that my post comes across as pro-totalitarian communism. But that wasn't my point at all. My point wasn't that I think that things were "better during communism" my point was simply to try to present the opinions of those who do understand that the pro-western privitization process was unfair.
I just think that a more moderate, slower, and fairer privitization process would have been better for the citizens of the country, and not that we should have stayed totalitarian/communist.
More Poles realise this now, but they are often frowned upon by the mainstream media (surprise, surprise - they are mostly German-owned) and called regressive, nationalistic, etc. But the trends are rather positive, we now have a government that seems to support Polish industry, and peoples' attitudes have changed. For example, there's a very popular mobile app called Pola, which tells you how much a product is "Polish" (it computes a score based on where the owners are located, whether the company does R&D in Poland, and so on) based on barcodes. It's actually pretty trendy now, and it's not uncommon to see people scanning stuff with it in supermarkets. More about it here: https://growthengine.withgoogle.com/intl/en-eu/voices#card-s...
Interesting fact - some people responsible for the sell out to the west have been recently given prominent positions in Ukraine's government structures.
Perhaps they don't want their normal screen name to be associated with pro-nationalistic views, as they are un-popular. I imagine that if I were a Pol working for a German software company, I would feel uncomfortable writing what OP just wrote under my normal screen name :P
I don't have pro-nationalistic views, that's a purely economic issue to me - its result is a long-lasting drain of capital. I did feel somewhat weird after posting that though, and @lossolo caught me.
On the other hand, I remember that there were lot of Czech companies that were privatized into Czech hands and most of them do not exist anymore, because they went bankrupt a long time ago.
At the time (during the 90s), most of those companies that had foreign investors/management grew and survived and those that didn't were often tunneled-out by a corrupt management and were closed down. There were definitely some pretty big cases in 90s. Also there was a big public discussion about this at the time, and many politicians were against the foreign investments in general.
So I simply don't think that it's so clear-cut issue and that it's not like most of Czech people are completely unaware of the problem.
You are, of course right, that the privatization process wouldn't have ended up being magically perfect if foreign investment would have been banned. There is so much awfulness that happened around the privitization process... And I wonder, if škoda hadn't been bought, would it have simply gone bankrupt after having been tunneled.
But if škoda had gone bankrupt, then that wouldn't have been an end to the Kolín, Mladé Boleslav, Leipzig industrial triangle. Cars would still be made here, given the huge amount of equipment and expertise in the area. But with škoda owned by VW, it cements a market dominance which is hard to shake. In a way, capital ownership is perminant.
A worse case is that of the swiss owned coal mines in the north. We behave as though those are still state owned mines. We kick people out of their houses with the power of the law. But the profits flow across the border to the west. Personally, I think that the government should pass a law that would re-nationalize the mines, without paying anything for them. The Chinese would certainly re-nationalize their mines if they were in the same situation. Right now, we behave as if capitalism is absolute, and democracy is relative to capitalism. The people cannot vote to confiscate the mines, but the companies can certainly bribe elected officials into privatizing them. So privatization is a one way process which puts more power into the hands of capitalists and takes power from the voters.
And now the Chinese are buying assets in democratic countries, and we are sitting here with our hands folded, saying "well, capitalism must come first. Democracy must never be allowed to interfere with the needs of capitalism." Which will be the death of democracy, unless we learn to dissobey capitalism and vote for things that break the rules of absolute and perminant ownership.
Well, that's a more general problem with indentation based languages, but Python users seem to be happy. It's a tradeoff, code that's (arguably) more pleasant to read, for a bit more work while writing it.
It's true that gofmt-like tool would be less beneficial, but I think that most of the things you've mentioned could be helped with editor/IDE support, e.g. indent-aware paste, block navigation and so on.
I think the issue with CoffeeScript was that it didn't add anything beyond the arguably prettier syntax.
Have changes the syntax too, but it is also trying to fix things that often cause bugs in Go code, it adds generics, and there are a few other features on the way, too.
> How do you declare methods on non-struct types? Or are you disallowing defining your own non-struct types?
It will be allowed, through type opening (which isn't implemented yet). It's mentioned in the intro, I called it "structure opening" but actually it will work on any non-builtin named type.
I'm not sure how the syntax will look like yet, but say that you have a struct:
struct A:
func MethodA():
pass
Then you will be able to open it and add new methods to it, it could look like this:
open A:
func MethodB():
pass
> How are the generics implemented? The general tone seems to suggest that you do naive template expansion, but the section about when makes it seem that you are using interface{} and type-assertions in the generated code.
No, it's closer to template expansion, "when" is "executed" (for lack of a better word) during compilation. Inactive "when" branches are ignored by the code generator (and type checker, thus they can contain code that's invalid for given instantiation). There's no type-switch in the resulting Go code.
> How is the overloading of make handled (assuming it's actually just implemented with the existing generic mechanisms)?
Right now it's not handled at all. I don't want to add default argument values to the language, so, unless I come up with something better, I'm afraid that there will need to be more than one "make" function, each named differently.
Thanks for your feedback! I'm well aware that the chances of Have becoming popular are tiny, but I'm having lots of fun working on it anyway.
> No, it's closer to template expansion, "when" is "executed" (for lack of a better word) during compilation. Inactive "when" branches are ignored by the code generator (and type checker, thus they can contain code that's invalid for given instantiation). There's no type-switch in the resulting Go code.
As I've now understood better what you mean, let me suggest a possible solution to your default-case conundrum: What about, if I don't want to allow the default-case, I just leave out the default case? And the compiler erroring out, if you pass in a type to a when-statement that doesn't have a valid case defined?
> No, it's closer to template expansion, "when" is "executed" (for lack of a better word) during compilation. Inactive "when" branches are ignored by the code generator (and type checker, thus they can contain code that's invalid for given instantiation). There's no type-switch in the resulting Go code.
I'm not sure I understand. Could you point to where you do this on Github?
That too, but I rather meant predeclared named types. According to Go specification, types that you define with "type Name OldName" are named, but "int", "string" and the rest are named too. Go disallows declaring methods on predeclared types.
Non-commercial use only. You agree not to use our Services for any commercial or business purposes and we (and our Providers) have no liability to you for any loss of profit, loss of business, business interruption, or loss of business opportunity.
It's funny that a plan called "Pro" cannot be used professionally.
https://www.anthropic.com/legal/consumer-terms
reply