I thought the point of the large language model version of AI was that they can understand human communication.
MCP seems like we have given up on making the models good or smart. We are bending over backwards to make the internet easier to interact with for AI than for humans.
If general intelligence is on the horizon, this all seems a colossal waste of time. (Not your resume. I mean the general direction of AI development.)
I'm pretty sure EU apps can link elsewhere for payment, but they have to show a warning message first saying that the payment is outside of Apple's jurisdiction and you can't get a refund from apple, etc.
Am I crazy or was this not already possible in the US? When I tried to sign up for Disney Plus to watch the new season of Andor, Disney told me I had to complete the signup outside of the app. Then I got a big warning from Apple saying I was no longer transacting with Apple and wouldn't be able to cancel the subscription or get refunds via the App Store if I continued. Am I misunderstanding what this trial was about?
It really isn't. If they were purely focused on getting training data, they would give more captchas to everyone, not just the users with no google cookies, connecting from VPN, and with weird browser configurations. The fact of the matter is that all those attributes are more "suspicious" than average, and therefore they want to up the cost for getting past the captcha.
In a very real way, adolescents rise to the responsibilities given to them. Usually teenagers that need modulating is a reflection on their upbringing rather than any innate flaw.
They are old enough to drive certain vehicles and old enough to buy alcohol. If we trust them with that surely we can let them do things during the day without constant adult supervision.
That's exactly the point. It's a middle ground ABV where there aren't a ton of products and below which are mostly fermented beverages and above which is distilled liquor.
In Switzerland, it is 16 for beer and wine, and 18 for spirits (or drinks that contain spirits, like "alcopops", even if they have a low ABV). I think Austria and Germany are the same.
People who make statements like that are the kind of people you dread will pick up your pull request. You just know you're going to go from maybe spending an hour cleaning up some suggestions to a 3-day philosophical battle to get them to a point where they deign to accept your PR.
Not at all. If the code is decent and shows effort I have no problem. If it's sloppy it shite code.
I really don't have time to care about what my peers think of me. It's work. I don't want to communicate with them outside work. Work is just another mind space that stays at work. I am strict when it comes to code, I expect the same.
I want working maintainable code to enable me to do my job. If people dread submitting a PR because they can't write code with effort, good. I like my ships built strong not weak.
If they fix their problem, good. Trust given, more than happy to salute however time and time they've proven to me they don't.
These developers have proven to me they won't. These are developers who are those who do not fix the issuing code and will just move on to the next problem hacking it to make it work.
If you've never worked with such, then lucky. If this sting for you, time to put more effort in to your work.
No, they really don't. They submit their code. I submit mine. If they have a problem with mine, I'll fix what they have issues with if they are reasonable, why wouldn't I?
Where I work in enterprise your peers change daily. With my role and importance to the company implementing hacky code puts me at risk and so I will of course push back. The people I knew last months may not even work in the company.
The view of I must be a horrible person comes from the Comment OP being angry at me for having a reasonable standards to an Enterprise standard of code. "It works im done. Next please"
If their code isn't up to scratch I will tell them and reject it. The issue I have is lazy developers who implement hacks and don't actually go and fix the code.
I am being made the bad person from someone's angry hospitality. All I was saying is that lazy developers are lazy developers and that I axe their work because it's sloppy and doesn't deserve to be on show.
I'm not to engage further as you're only ever going to repeat yourself in more obnoxious ways, but I will say that if you treat people in real life the way you comment here, you will only ever be tolerated at best. Never respected and never liked.
Even the people you consider peers will abhor that you think and speak like this about people.
I'm not even sure what you're talking about. Is this a JavaScript complaint and they were meant to use a for..of? Are you an FP purist and think they were supposed to use map/filter?
This sounds like you have some very specific trauma around a very specific "foreach if loop", because I would personally never throw around such a specific-but-not-specific example of tech debt. Tech debt is extremely contextual.
That's not true at all. You think games would feel as immersive if everything was Calibri? Magazine-style articles would feel as tactile if they all used the same system fonts? Etc.
You may not care about fonts, but to say they don't matter is a misunderstanding. For example, I could glibly say we only need one programming language (the user doesn't care what syntax you used before it was compiled down to 1s and 0s!), but any engineer would make the case why that's not true at all.
No, I think we're just looking at it from different perspectives.
Yes, most people are fine choosing from the fonts available on their computer when writing a document.
But that's not what me nor OP are talking about. We're talking about shipping software (like a mobile app), or publishing a blog post. In that case, the best you can specify is either a very common font (Helvetica, etc), or a high-level classification (serif, sans-serif, etc).
There are many free fonts out there, yes, but there's a reason they're free. The quality for a majority of them is significantly lower, and many designs come with constraints (either utilitarian or stylistic). You don't have to agree, but I'm not being absurd or lacking sincerity.
You're also just going around and commenting the same thing on each of my posts. But don't limit your understand to just my writing here; there's thousands of books about the importance of typography if you're curious to learn more.
> There are many free fonts out there, yes, but there's a reason they're free.
Go on and tell me what that reason is then. Are you also going to tell me free open-source software, like Linux is low-quality because its free?
> The quality for a majority of them is significantly lower
Again, a completely baseless, unprovable assertion.
> You don't have to agree, but I'm not being absurd or lacking sincerity.
What do you call your example of using Calibri for everything in response to someone suggesting the use of free fonts?
You are lacking sincerity and making absurd claims. Almost everything else you say is literally baseless rhetoric that you are unable to back up with data or any objective argument.
> there's thousands of books about the importance of typography if you're curious to learn more.
It's amazing that you apparently know of thousands of such books, but are unable to make one coherent, objective argument to back up your claims.... did you read them?
You’ve been combative throughout this thread, and it's clear that you don’t see typography or design as disciplines that warrant serious thought. I don't think you're actually willing to engage with an explanation of why it matters but I'll try anyway.
System fonts are the absolute bottom of the barrel. Some are well designed but using any of them is a visual shorthand that you didn't care enough to put thought into your design. You're associating your product with the ocean of amateur work on the internet, giving the impression you copy pasted a template.
There are some high quality free fonts typically backed by massive organizations with actual typographic expertise. Most free fonts however, are amateur work that are technically and functionally lacking. Professional fonts are well designed at all weights, they're carefully spaced, they include much larger character sets to support more languages, contain features like lining and non-lining figures, variable font weights, small caps... are those all slight differences?
There’s a reason so many articles exist with titles like “Google Fonts That Don’t Suck”. Most of them do. If you are a professional whose job requires working with type, then choosing a font is foundational to your product. Arguing that all design is BS is just lazy; it's not a coherent argument.
I highly recommend practicaltypography.com, a free web book that discusses all of this and more, including why system fonts are bad and why a professional typeface is worth paying for.
This claim that system fonts are the "bottom of the barrel" is just so clearly false that I don't understand how you can be an advocate of typography and say it. Both Microsoft and Apple put huge amounts of effort into typography, contract or employ well-regarded designers, and their outputs are themselves well-regarded.
If you wanted to say "most of what's on Google Fonts is bottom of the barrel", you'd have a colorable argument. But that isn't what you said.
San Francisco is a great font. Arial is a perfectly functional semi-clone of Helvetica, Times New Roman is a decent interpretation of Plantin. Roboto is an interesting mash-up of Helvetica, DIN, and a few others.
System font from a web standpoint means you get one of these depending on the user's choice of phone, desktop, and/or browser.
It is somewhat like buying art because the frame covers a blemish on the wall. That the print inside the frame might be of a famous impressionist painting does not mean that the frame or the print necessarily go with the room.
The car analogy involves a car rental place - that they may give you any one of several newish, functional and even stylish vehicles does not change that you may often wind up being paired with a vehicle mismatched for your function.
Around the time Matthew Carter was creating Georgia, one of the most widely-used system fonts in the world, for Microsoft, he was also widely considered one of the best typographers in the world. Georgia is not hotel room wall art.
There are many laughably horrible attempts at fonts out there on free font sites (I remember my days learning to write software in the early 00s), sure. But there are also high quality professionally designed and typeset fonts available for free, including those of the system variety. The argument is comparing the latter to expensive designer fonts, not the former to high quality fonts.
Disagreeing with you doesn't mean I'm combative. (Not that I care)
> typography or design as disciplines that warrant serious thought.
We are talking about fonts here, more specifically fonts used in software, more specifically the quality of free fonts used in software. Not 'design' as a whole which is much more than that.
> System fonts are the absolute bottom of the barrel.
If you say so.
> You're associating your product with the ocean of amateur work on the internet, giving the impression you copy pasted a template.
Reusing a font means you're copy-pasting your article/app/etc from a template? Erm ok.
> There are some high quality free fonts typically backed by massive organizations with actual typographic expertise.
'Some'? Like 1000? 10000? How many fonts does one application need? 'typically'? How 'typically'? And I'm not being pedantic - your statements are pretty meaningless without actual numbers.
> Professional fonts are well designed at all weights, they're carefully spaced, they include much larger character sets to support more languages, contain features like lining and non-lining figures, variable font weights, small caps... are those all slight differences?
What is a 'Professional font'? lmao
Plenty of free fonts have all of the features you've listed, and plenty of non-free fonts don't.
> There’s a reason so many articles exist with titles like “Google Fonts That Don’t Suck”. Most of them do.
Again 'so many' and 'most'... you should provide specific (at least approximate) numbers, otherwise this says nothing about how many good free fonts are actually out there.
> Arguing that all design is BS is just lazy
Well I didn't say that, pretending that I did is pretty lazy tho.
> I highly recommend practicaltypography.com, a free web book that discusses all of this and more, including why system fonts are bad and why a professional typeface is worth paying for.
Oh geez! A FREE book which tells you why you should pay for 'professional' fonts while at the same time selling them to you with affiliate links! Thank you sir!
You should care if you're being combatative, but, even more importantly, quoting previous comments the way you're doing doesn't work well on HN and is also a flamewar trope. Everybody can read the comments you're responding to. Just refer back to them in prose. A single quote, maybe 2 in a long comment, fine, but what you're doing now creates the impression that you're sort of rebutting what the previous commenter said as you read them, sentence by sentence, which is a tell that you're not actually thinking about what they said.
Also: they're pretty clearly wrong, so you shouldn't need any of this to refute them.
I am rebutting what the previous commenter said, sentence by sentence (almost), I don't know why that tells you that I'm not actually thinking about what they said though. Did I misunderstand or misrepresent something they said?
I'm not going to argue with you, but I just want to point out that the person I was responding to specifically used the phrase "system supported fonts". That's why I mentioned Calibri.
> '..freely available and/or system supported fonts.'
Not just 'system supported fonts' (whatever that means), and not just Calibri. That's why your 'use Calibri for everything' example is absurd and does not at all address the point they made.
The last sentence is the variety that is super tempting to make but counterproductive because it shuts down discussion or poisons it thereafter its made to impress bystanders not actually communicate with the person.
Agree that it might not be the best, but seems like a fairly appropriate response for someone trying to back up their rhetoric with 'thousands of books out there'. How is it 'made to impress bystanders'?
Who cares? They're part of a line of argumentation that dunks on the typography work of Matthew Carter. This is very much the same thing as a thread on industrial design dunking on Dieter Rams. You don't get angry at that kind of argument; you laugh at it.
Oh I was never angry, I was enjoying the argument (maybe that makes me combative, oh well), and I was completely open to being proven wrong and thereby becoming more informed on the topic... alas...
Something tells me that some designers care about fonts a heck of a lot more than most consumers do. As a consumer, I care about legibility above all else. There are plenty of metrics that affect that, but many of the freely available (albeit, not necessarily free) fonts are perfectly fine on that front. More bluntly, some of those freely available fonts are going to be better than the vast majority of fonts that you can pay for because: (a) companies like Apple, Google, and Microsoft have invested in their development or licensing to ensure their customers have access to high quality fonts with coverage for most languages; and (b) they have wide availability, since font substitution is going to have a much larger impact upon the perceived quality of a document than its use of quality fonts.
Maybe this pedantic snobbery will matter again when we switch back to creative mode, but it all seems highly elitist right now while many are trying to just survive.
I admire your passion, but... as someone who is not deeply interested in fonts, I view them in largely functional terms. Can I read it? Does it look ok?
Programming language choice has an aesthetic side, but it is also very much a functional concern. Can I write secure code? Will it be performant? Will it be maintainable?
Different languages represent different functional tradeoffs. Are fonts really the same kind of thing? IOW, how would you make a choice between using Arial vs. Helvetica?
Arial v Helvetica is an interesting example, because Arial was designed basically as a cost-efficient alternative to Helvetica. So, the reason you'd choose between the two is exactly the thing the original comment was complaining about – licensing! They were designed to be metrically compatible... meaning, the character widths and spaces are exactly the same. This means that switching to Arial won't affect the layout of your document. This was more important when things were more analog, but it's still important with digital documents: for example, it could mess up the number of pages, which would affect meta content or create line breaks that seem meaningful but aren't. Additionally, having things like a widow (a word by itself on a new line) can disrupt the visual flow and draw focus to or away from content in ways you don't desire.
But just because those two typefaces are quite similar (and the reason to pick between them is largely financial/convenience) doesn't mean you'd never want to have more fine-grained control over the text you're working with.
You mentioned security. When I'm editing this comment, 0 and O are very different (the zero has a slash through it), however when I hit save they look quite similar. (But because we're all using system fonts on HN, it might be different for you). While it's often just a stylistic choice, in many situations the two characters would be indistinguishable and that would be an issue, which is why someone might choose a typeface where characters are significantly different. Think a password you have to transcribe.
If you know your font will be used in a quite small size, you may want one that is optimized for being read at tiny sizes. If you're displaying something technical, a monowidth font is better suited.
And all of this focused on utility for the most part; I'm leaving out all the reasons you'd want it for stylistic reasons. If you're trying to make people feel at ease, you may want typeface where the end of the strokes are rounded, for example. Sometimes you want people to feel a certain way, in the same way you modulate your tone when talking.
Very few system fonts are any good. Would you use Arial instead of Helvetica Neue? I certainly wouldn't. Put two posters side-by-side and you'd notice the Helvetica one as looking more professional, even without any design background.
Additionally, very few system fonts include all the weights. Fonts aren't just come in a single weight. The font you use for a giant page-filling title is generally skinnier than the font used for a caption.
Good design creates a reaction, such as causing you to buy something or interacting more with something or whatever, even for people that say they don't care about design.
An obviously false statement which you can't possibly back up.
> Would you use Arial instead of Helvetica Neue? I certainly wouldn't. Put two posters side-by-side and you'd notice the Helvetica one as looking more professional, even without any design background.
First of all that's just completely your own subjective opinion. Second, there are many other free sans-serif fonts out there to choose from (examples[1]).
> Good design creates a reaction, such as causing you to buy something or interacting more with something or whatever
'Design' can encompass many things, but can you show me some data that backs up your claim that slight differences in fonts will make a difference in product quality/performance/revenue/etc? Because I have seen a loooot of data that says it's almost always completely irrelevant.
This is just clearly wrong. Even Georgia and Verdana are very serious works of typography. The Cleartype fonts hold their own against modern text faces. San Francisco and New York are also obviously strong fonts. These are gigantic companies that take typography seriously, they can easily afford to invest in competent system fonts, and they both obviously have.
They updated it for the sequel, and one example doesn't nullify thousands of years of design.
But to go down that path from a logical standpoint... Papyrus isn't on my computer (OSX) for whatever reason, and it doesn't come on Linux. Papyrus isn't a free, public font... it's licensed by its owner (ITC), so the only reason you can use it on your computer is because someone is paying a license for you to see it.
I guess it comes down to how you view the concept of "the medium is the message". Should the tone be set by the creator of the software / writer of the blog post / etc, or should the end user choose one typeface for everything (or have fine-grained control over everything they read and view?)
I don't think this makes much sense as an argument, because you can have it either way with the status quo. The question isn't whether creators can use typesetting expressively; they clearly can, with a degree of freedom and optionality that would have blown me away when I started font nerding back in the 1990s. The question is whether I should sympathize with designers who are irritated by the licensing terms for Gotham or Brandon Grotesque (or whomever is doing per-impression licensing these days). I do not, and I think I'm on solid ground.
Doesn't look very expensive to me. An LLM capable of this level of summarization can run in ~12GB of GPU-connected RAM, and only needs that while it's running a prompt.
The cheapest small LLMs (GPT-4.1 Nano, Google Gemini 1.5 Flash 8B) cost less than 1/100th of a cent per prompt because they are cheap to run.
Yes! And also, Apple loves selling expensive hardware and has zero shyness asking people to pay a few thousand bucks to buy into part of their ecosystem.
They could easily offer an on-prem family 'AI' product that you plop in your house and plug into your router, and does all AI processing for the whole family, and uses a secure VPN to connect to any of your devices outside the LAN.
If such a product delivered JUST what this guy's cool hack provides, and made Siri not a stupid piece of sh*t for my family, I'd buy it for $1999 even if I knew it cost Apple $700 to make.
You of course sold all your stocks beginning of November because you knew a deep crash was coming and wanted to be ready to buy into it. And now you have a lot of cash.
MCP seems like we have given up on making the models good or smart. We are bending over backwards to make the internet easier to interact with for AI than for humans.
If general intelligence is on the horizon, this all seems a colossal waste of time. (Not your resume. I mean the general direction of AI development.)
reply