Swimming upstream against the AI current—updated!
I had to give the seal of approval precisely to the only one of the 3 AI agents I currently use that cannot search the web! Why? Well, because searching the web is not the primary reason for which such LLMs and the chatbots built on them were created! They’re meant to use their “intelligence” stored in those many billions of parameters!
ChatGPT only got web search on October 31, 2024, or probably the next day (available to everyone since February 5, 2025).
The LLMs belonging to companies that own search engines are a special case. Copilot got web search earlier, on February 7, 2023, when it was known as Bing Chat and its raison d’être was to apply its “intelligence” to web searches only. The distinct Copilot service first showed up on November 1, 2023, for Microsoft 365 Enterprise customers; and Microsoft Copilot was born on November 15, 2023. However, Google’s Gemini, initially called Bard, gained web search capabilities in May 2024. Indeed, in 2023, Bard relied on the data it was trained on.
But people were raptured by ChatGPT’s ability to write poems, essays, to rephrase or summarize text, and to answer various questions. Also, to write code!
Chatbots started to create images by interfacing with other AI systems, such as DALL-E, so again, this has very little to do with the models whose “intelligence” was both praised and ridiculed (when such chatbots were giving blatantly false or completely made-up answers).
So, to me, the LLM behind the chatbot itself is what matters the most.
Does anyone know the name of the Prime Minister?
And yet, when a chatbot does have access to the Internet, it’s unacceptable for it to give false answers, being them only because they’re outdated. I’m extremely reluctant to ask a chatbot about current and recent events—I still know how to use a search engine or Wikipedia, thank you. But let’s play the game.
The idea came to me from the reading of this article: Mistral AI : Le Chat rivalise-t-il avec ChatGPT ?
À cette heure, Le Chat ne parvient pas à se hisser au niveau de ChatGPT.
Well, OK, that has always been the case, except that ChatGPT hallucinates itself, and sometimes Mistral is much better. Or it used to be.
But the article revealed that Mistral was unable to give the correct answers to the question « Qui est le Premier ministre français ? », whereas an extremely recent event triggered a web search that led to a correct answer: « D’après les déclarations d’Emmanuel Macron datant du 9 février 2025, combien la France va-t-elle investir dans l’intelligence artificielle dans les années à venir ? ». Very odd.
I have to admit that I couldn’t understand the part about Mistral’s Canvas. They mentioned this: “Les Canvas de ChatGPT sont directement inspirés des Artefacts de Claude, lancés en août 2024.” So I hopped to “Claude : les Artefacts sont disponibles pour tous, comment ça marche” and “Artifacts are now generally available” but I still have no clue as to why would one want to create such “artifacts” in a chatbot. How does one export such a creation to a popular document format, supposing there is a format supporting such presentation-like output? The entire use case feels unreal and unrealistic to me.
As a side note, the same site made a similar comparison: DeepSeek contre ChatGPT : quel chatbot IA est le plus performant ? Unfortunately, they don’t seem to have been used DeepSeek-R1’s DeepThink, which is a game changer, so I don’t find the comparison relevant.
🔴 Back to our sheep, Mistral is indeed dumb:

🟢 Copilot is spot-on, and it gives an authoritative source (Wikipedia), not some random news outlets:

🟡 Claude can’t provide the correct answer, but it also can’t search the web, and its knowledge cutoff is in April 2024:

Unless forced to, I won’t use any time soon the other 4 chatbots I still have bookmarked—ChatGPT, Gemini, Perplexity, DeepSeek—but I decided to bring up Duck.ai, because someone mentioned it to me (a fan of DDG).
Say hello to Duck.ai!
This is what you get by accessing Duck.ai, which redirects to DuckDuckGo.com with added parameters. The service promises “Anonymous access to popular AI models, including GPT-4o mini, Claude 3, and open-source Llama 3.3 and Mistral.” In fact, GPT-4o-mini, Llama-3.3-70B, Claude 3 Haiku (not 3.5!), o3-mini, Mistral-8x7B.
Claude 3 is older than 3.5, and it shows:
The current Prime Minister of France is Élisabeth Borne. She assumed office on May 16, 2022.
GPT-4o-mini:
As of my last knowledge update in October 2023, the Prime Minister of France is Élisabeth Borne. She has been in office since May 2022. However, please verify with a current source, as political positions can change.
Llama-3.3-70B:
The Prime Minister of France is Élisabeth Borne. She has been in office since May 16, 2022.
And so on. As useful as the last Encarta edition. Tell me what qualifier should I use instead of mentally retarded with regard to Duck Duck Go’s choice to use LLMs without access to the Internet, when their main business is web search!
Oh, but DDG has its “Bing Chat”-like assistant, too: Assist: “Generate a short answer from the web.”

All right, fine. “Assist” is OK, but why giving The Guardian as a source? Wikipedia is enough. Also, if you want an official one, www.info.gouv.fr (omit the “www” and you’ll be redirected to the 1996-style URL that starts with “www”).
But Google and Bing offer the correct answer even without explicit AI assistance, so what’s the point of using DDG’s enhancements? Startpage doesn’t give any specific name when asked “What is the PM of France?”, and this is why I stopped looking for alternate, independent, privacy-cautious, useless search engines. I’ll just stick to the top two. No, I never liked DDG, and I never trusted them: no ads, respecting my privacy, free, yet not in bankruptcy? This can’t be. They must be liars.
And Duck.ai lacks a reason to exist.
German grammar vs. AI
I’m using Le Chat Mistral since March 1st or 2nd, 2024, after having read this in Le Monde: On a testé Le Chat, l’étonnant ChatGPT à la française de Mistral AI. Stunning or not, I wanted an alternative to ChatGPT, which was still hallucinating. I have to say that Copilot also provided me with hours of utterly hilarious inventions. Unfortunately, that was “the old Copilot”; the current version doesn’t show any conversations earlier than Aug. 9, 2024.
So yeah, Le Chat Mistral was at times better and sometimes worse than ChatGPT. Its style was different, though, and I appreciate that. Now Mistral’s style is much worse, and I hate that. Can’t they bring new LLMs without breaking what was good? Note that I disable Mistral’s Flash Answers. But I’d like it to have, like Claude 3.5 Sonnet, a choice of styles: Normal, Concise, Explanatory, Formal.
Among the first things I asked Mistral on March 2, 2024 was a set of questions on German grammar:
- Why in “Ich möchte einen Kaffee”, “einen Kaffee” is in Akkusativ, hence “einen”, whereas in “Das ist genau der Punkt”, “der Punkt” is in Nominativ, hence “der”?
- Please also explain why dates with “am” can use either of accusative (“den”) and dative (“dem”), and which are the differences between them:
“Mein Urlaub endet am Montag, den 29. August.”
“Mein Urlaub endet am Montag, dem 29. August.”
The answers were mostly disappointing. But let me briefly clarify the above topics.
In the first case, the key to the answer is the predicate nominative. Mistral didn’t mention that. What it said instead was confusing: “Subject complements provide additional information about the subject of the sentence and are usually in the nominative case.” The exchange was rather long and unsatisfactory, and it included a useless talk about the subject-verb-subject complement (SVSC) word order and the subject-verb-object (SVO) word order.
In 2025, though, Mistral does mention the predicate nominative. It got smarter. Still, its style is too schematic.
In the second case, it’s a bit tricky. In full fairness, only the first variant is correct and Hochdeutsch. But the second variant can also be encountered, even in writing, in more colloquial situations. A very good explanation were to consider it a special form of a tight apposition, with a logical disconnect from whatever was said in the first part. Here’s a detailed lesson in German: Apposition. There is a distinction between “enge Apposition” (tight apposition) and “lockere Apposition” (loose apposition). To the point:
Der Kasus ist bei engen Appositionen meistens kongruent, ebenso bei lockeren Appositionen, die einen Artikel enthalten. Wenn der Kasus einer lockeren Apposition von dem des Bezugsworts abweicht, steht er meistens im Nominativ oder Dativ:
Der Sieg des Mannes, Handwerker aus Bottrop, kam überraschend. (Bezugswort: Genitiv, Apposition: Nominativ)
Der Sieg des Mannes, einem Handwerker aus Bottrop, kam überraschend. (Bezugswort: Genitiv, Apposition: Dativ)
Der zweite Satz dieses Beispiels ist laut deutscher Grammatik falsch. Da die Apposition einen Artikel enthält, muss sie eigentlich kongruent zum Bezugswort sein. Nicht kongruente Appositionen begegnen Dir in der Umgangssprache. In offiziellen Texten und Gesprächen solltest Du sie vermeiden.
From the two grammatical disconnections mentioned above, the first one is correct of the Nullartikel. People ask such questions even on Reddit, and here’s another example: 100% correct + grammatical disconnection + correct because of the Nullartikel:
- “Das war der Vorschlag seines Vaters, des berühmten Architekten.”
- “Das war der Vorschlag seines Vaters, dem berühmten Architekten.”
- “Das war der Vorschlag seines Vaters, berühmter Architekt.”
German grammar is absurd, but one has to live with it.
To show you what 3 chatbots answered me these days, I’ll start using a different order: Claude, Mistral, Copilot, Copilot + Think Deeper. That’s because Claude’s answers are generally shorter. The answers are shown in a tabbed interface, so don’t disable JS in the browser.
-
Claude
-
Mistral
-
Copilot
-
Copilot Think Deeper
■ Why in “Ich möchte einen Kaffee”, “einen Kaffee” is in Akkusativ, hence “einen”, whereas in “Das ist genau der Punkt”, “der Punkt” is in Nominativ, hence “der”?
Claude 3.5 Sonnet: Let me explain this difference by looking at how these sentences work grammatically.
In “Ich möchte einen Kaffee,” we’re dealing with a transitive verb “möchten” (to want) that requires a direct object - the thing that is being wanted. In German, direct objects are always in the Akkusativ case. Think of it this way: the action of wanting is being directed at the coffee, so “Kaffee” becomes the recipient of the action, making it a direct object. Since “Kaffee” is a masculine noun, its article changes from “der” to “einen” in Akkusativ.
In “Das ist genau der Punkt,” we have a completely different grammatical structure. Here, we’re using the verb “sein” (to be), which is a linking verb. Linking verbs in German connect the subject with what’s being said about it - they don’t show an action being done to something. Both parts of this equation, so to speak, remain in Nominativ. “Das” is the subject, and “der Punkt” is what we call a predicate nominative - it’s essentially saying that “das” equals “der Punkt.” Since no action is being performed on “Punkt,” it stays in Nominativ case with its original article “der.”
To make this clearer, let’s compare it to English. In “I want a coffee,” “coffee” is receiving the action of wanting. But in “That is exactly the point,” we’re not doing anything to “the point” - we’re just saying that “that” and “the point” are the same thing.
Would it help if we looked at some more examples of how these cases work in different contexts?
■ Please also explain why dates with “am” can use either of accusative (“den”) and dative (“dem”), and which are the differences between them.
Claude 3.5 Sonnet: I want to clarify something important: with “am” in German dates, we actually only use Dative case (dem), not Accusative (den). Let me explain this in detail.
The preposition “am” is actually a contraction of “an” + “dem” (Dative). When we talk about dates in German, we say for example “am dritten Mai” (on the third of May), where “dritten” is in Dative case. This makes sense because “an” in this temporal context - when we’re talking about when something happens - always takes the Dative case.
You might be thinking of expressions with dates that use “den” because there’s another common way to express dates in German: using the Accusative case without any preposition. For example:
“Ich komme den dritten Mai” (I’m coming on the third of May)
So we have two correct ways to express the same thing:
- “Ich komme am dritten Mai” (with “am” + Dative)
- “Ich komme den dritten Mai” (bare Accusative, no preposition)
This distinction relates to a broader pattern in German: when we talk about points in time, we often have these two options. The Accusative without a preposition is slightly more formal or literary, while the “am” + Dative construction is more common in everyday speech.
Think of it like this: the Dative with “am” emphasizes the point in time where something happens (answering “when?”), while the bare Accusative treats the time more like a duration or span that the action moves through (answering “for how long?” or “during what time?”), even though in this case it’s just a single day.
Would you like to explore some more examples of how these two constructions are used in different contexts?
■ But I encountered both of these:
“Mein Urlaub endet am Montag, den 29. August.”
"Mein Urlaub endet am Montag, dem 29. August."
Claude 3.5 Sonnet: Ah, now I understand your question better! You’re absolutely right - both forms are indeed correct in this specific construction. Let me explain why this interesting case works differently from what we discussed before.
When we’re using appositions (additional descriptions that rename or specify something) with dates in German, we can use either case. Here’s why:
In “am Montag, den 29. August,” the apposition “den 29. August” is in Accusative case because it’s functioning as what we call a loose apposition. It’s almost like starting a new thought that provides additional information about the date. You could think of it as if there’s an implied “ich meine” (I mean) or “das ist” (that is) between the parts.
In “am Montag, dem 29. August,” the apposition “dem 29. August” is in Dative case because it’s matching or agreeing with the case of “Montag” (which is in Dative because of “am”). This is called a tight apposition, where the second part is more closely bound to the first part grammatically.
This dual possibility with date appositions is actually part of a larger pattern in German grammar. When we have appositions, they can either:
- Match the case of the word they’re describing (tight apposition)
- Follow their own grammatical rules somewhat independently (loose apposition)
Think of it like this: When you say “dem 29. August,” you’re treating the date as being very tightly connected to “Montag” - almost like they’re one unit. When you say “den 29. August,” you’re treating it more like additional information that’s slightly more separate from “Montag.”
Would you like to explore how this pattern of tight versus loose appositions works in other contexts beyond dates? It’s a fascinating aspect of German grammar that appears in various situations.
■ What are the German terms for “tight apposition” and “loose apposition”?
Claude 3.5 Sonnet: The German terms are “enge Apposition” (tight apposition) and “lockere Apposition” (loose apposition). Let me help explain how these terms developed and are used in German linguistics.
“Eng” literally means “tight” or “narrow,” and in this grammatical context, it refers to how closely bound the appositive phrase is to the word it’s describing. When we use an “enge Apposition,” the appositive element is considered an integral part of the noun phrase, almost inseparable from it grammatically. This is why it takes the same case as the noun it modifies.
“Locker” means “loose” or “relaxed,” and a “lockere Apposition” has more grammatical independence. Think of it as having a bit more breathing room between the main noun and its appositive. This independence is why it can follow its own case rules.
What’s particularly interesting is that in German linguistic tradition, you can often recognize a “lockere Apposition” by the commas that set it apart (like in our date example), while “enge Appositionen” typically don’t require commas. For instance:
Enge Apposition: “Mein Freund Peter” (My friend Peter)
Lockere Apposition: “Mein Freund, der Arzt,” (My friend, the doctor,)
This distinction becomes especially relevant with dates because they represent one of the few instances where German grammar allows both forms to coexist as equally correct options. The flexibility here reflects the dual nature of dates as both integral parts of time expressions and as clarifying information.
Would you like to explore how these apposition types appear in other common German constructions beyond dates? Understanding their broader usage can help reinforce the concept.
■ But the discrepancy in Kasus in loose appositions is mostly to be use in colloquial language, right? While not blatantly incorrect, it’s not Hochdeutsch. I found the following 2 examples in a German grammar lesson:
Der Sieg des Mannes, Handwerker aus Bottrop, kam überraschend. (Bezugswort: Genitiv, Apposition: Nominativ)
Der Sieg des Mannes, einem Handwerker aus Bottrop, kam überraschend. (Bezugswort: Genitiv, Apposition: Dativ)
Claude 3.5 Sonnet: Let me help clarify this important nuance about appositions in standard German (Hochdeutsch). You’ve touched on a significant point about grammatical correctness versus common usage.
In standard written German, the grammatically correct form is for the apposition to match the case of its reference word (Bezugswort). In your example, since “des Mannes” is in Genitive, the most proper form would be:
“Der Sieg des Mannes, des Handwerkers aus Bottrop, kam überraschend.”
The examples you’ve provided actually demonstrate what happens in everyday language use, where speakers often simplify the case structure. The Nominativ version (“Handwerker”) and the Dativ version (“einem Handwerker”) are indeed common in spoken German and less formal writing, but they’re not considered standard written German.
This situation reflects a broader pattern in German language evolution: there’s often a tension between the strict grammatical rules of standard German and the simplifications that occur in everyday usage. The tendency to simplify case agreement in appositions is part of a larger trend where complex case structures are sometimes reduced in colloquial speech.
Think of it like this: Just as English speakers might say “between you and I” in casual speech (when “between you and me” is technically correct), German speakers sometimes simplify these appositive constructions. The difference is that in German, this simplification isn’t yet widely accepted in formal writing or Hochdeutsch.
When writing in formal contexts or preparing for language examinations, it’s best to maintain case agreement in appositions. This means having the apposition match the case of its reference word, regardless of whether it’s a tight or loose apposition.
Would you like to explore more examples of how this distinction between formal and colloquial usage appears in other grammatical structures? Understanding these differences can be very helpful for developing a more nuanced command of German.
■ Or rather, like English speakers say “It is me” instead of “It is I” 🙂
Claude 3.5 Sonnet: Ah yes, that’s an excellent parallel! The “It is me” vs “It is I” example actually demonstrates this principle even better than my previous comparison. Let me explain why.
In both English and German, we see this tension between prescriptive grammar rules and natural language use, especially with predicate nominatives (what comes after forms of “to be”/“sein”). In formal English, “It is I” is technically correct because both parts of the equation should be in the nominative case - just like in German with “Das ist der Punkt” where both parts are in Nominativ. However, most English speakers say “It is me,” using the objective case, because it sounds more natural.
This mirrors exactly what’s happening with those German appositions. The strict grammatical rule says the apposition should match the case of its reference word (like “des Mannes, des Handwerkers”), just as formal English grammar says we should use “It is I.” But in everyday speech, German speakers often default to simpler cases like Nominativ or Dativ for appositions, just as English speakers naturally say “It is me.”
What makes this comparison particularly apt is that both examples show how living languages tend to simplify complex case systems over time. In Old English, case distinctions were much more important, similar to German. Modern English has largely abandoned most case distinctions except in pronouns, and we can see German showing some similar tendencies in casual speech, particularly with these appositive constructions.
This helps us understand why, even though “des Mannes, des Handwerkers” is the prescriptively correct form, you’ll often hear and see variations with other cases in everyday German. It’s part of the natural evolution of language, where speakers tend to simplify complex grammatical structures when the meaning remains clear from context.
Would you like to explore other examples where German shows this kind of tension between formal grammatical rules and everyday usage? It’s a fascinating window into how languages evolve.
■ Why in “Ich möchte einen Kaffee”, “einen Kaffee” is in Akkusativ, hence “einen”, whereas in “Das ist genau der Punkt”, “der Punkt” is in Nominativ, hence “der”?
Mistral: In German, the case of a noun (and its accompanying article) is determined by its grammatical function in the sentence. Let’s break down the two examples you provided:
- "Ich möchte einen Kaffee":
- In this sentence, “einen Kaffee” is the direct object of the verb “möchte” (which means “want”).
- In German, the direct object of a sentence is typically in the accusative case.
- The accusative case for the masculine noun “Kaffee” is “einen Kaffee”.
- "Das ist genau der Punkt":
- In this sentence, “der Punkt” is the subject complement, which is in the nominative case.
- The nominative case is used for the subject of the sentence and for predicate nominatives (which describe or rename the subject).
- The nominative case for the masculine noun “Punkt” is “der Punkt”.
In summary, the case of the noun phrase is determined by its role in the sentence. The accusative case is used for direct objects, while the nominative case is used for subjects and subject complements.
■ Please also explain why dates with “am” can use either of accusative (“den”) and dative (“dem”), and which are the differences between them.
Mistral: In German, when talking about dates, the preposition “am” is a contraction of “an dem,” which means “on the.” The case used with “am” can be either accusative or dative, depending on the context and the specific phrase being used. Here’s a breakdown:
- Dative Case (“dem”):
- The dative case is generally used with “am” when referring to a specific date or day.
- For example, “am Montag” (on Monday), “am 1. Januar” (on January 1st).
- The dative case is used because these expressions are referring to a specific point in time, which is a typical use of the dative case in German.
- Accusative Case (“den”):
- The accusative case can be used with “am” in certain fixed expressions or idioms.
- For example, “am Morgen” (in the morning), “am Abend” (in the evening).
- These expressions use the accusative case because they are referring to a period of time rather than a specific point in time.
In summary, the choice between accusative and dative with “am” depends on whether you are referring to a specific point in time (dative) or a period of time (accusative). This distinction is part of the broader rules governing the use of cases in German grammar.
■ But I encountered both of these:
“Mein Urlaub endet am Montag, den 29. August.”
"Mein Urlaub endet am Montag, dem 29. August."
Mistral: In German, both “den” and “dem” can be used with dates, but they convey slightly different nuances and are used in different contexts. Let’s examine the two examples you provided:
- "Mein Urlaub endet am Montag, den 29. August.":
- Here, “den 29. August” is in the accusative case.
- The accusative case can emphasize the duration or extent of time, indicating that the vacation ends on that specific day.
- This usage is less common in standard German but can be found in certain regional dialects or colloquial speech.
- "Mein Urlaub endet am Montag, dem 29. August.":
- Here, “dem 29. August” is in the dative case.
- The dative case is more commonly used in standard German to indicate a specific point in time.
- This construction is generally preferred in written and formal contexts.
In summary, while both constructions can be encountered, the dative case (“dem”) is more standard and widely accepted in formal German when referring to a specific date. The accusative case (“den”) might be used in regional variations or colloquial speech to emphasize the duration or extent of time.
■ Have you ever heard of “enge Apposition” (tight apposition) and “lockere Apposition” (loose apposition)?
Mistral: Yes, in German grammar, “Apposition” refers to a grammatical construction where two or more words or phrases are placed next to each other to provide additional information or clarification about one another. Appositions can be categorized as either “enge Apposition” (tight apposition) or “lockere Apposition” (loose apposition), depending on how closely they are connected to the noun they describe.
- Enge Apposition (Tight Apposition):
- In a tight apposition, the apposition is closely connected to the noun it describes and is usually not separated by commas.
- The apposition provides essential information that is necessary to understand the noun fully.
- Example: “Der Dichter Goethe” (The poet Goethe) — Here, “Goethe” is in tight apposition to “Dichter,” and no commas are used.
- Lockere Apposition (Loose Apposition):
- In a loose apposition, the apposition provides additional, non-essential information and is typically separated by commas.
- The information can be omitted without affecting the core meaning of the sentence.
- Example: “Mein Freund, der Arzt, kommt heute.” (My friend, the doctor, is coming today.) — Here, “der Arzt” is in loose apposition to “mein Freund” and is separated by commas.
These distinctions help in understanding the grammatical structure and the intended emphasis or additional information provided in a sentence.
■ Can’t “am Montag, den 29. August” be a case of loose apposition, despite the comma? I.e., the exact date is more like an afterthought, to make sure the date is correctly understood?
Mistral: Yes, you’re correct. The phrase “am Montag, den 29. August” can indeed be seen as a form of loose apposition, despite the use of the accusative case. In this context, the date “den 29. August” is providing additional, clarifying information about the day “Montag” (Monday). The comma suggests that the date is being added as an afterthought or extra detail to ensure clarity.
In this interpretation:
- “Montag” is the primary reference point.
- “den 29. August” is the apposition, providing additional information to specify which Monday is being referred to.
This usage can be seen as a way to emphasize or clarify the exact date, even though it uses the accusative case, which is less common in standard German for dates. It reflects a more colloquial or regional variation in the language.
■ Why in “Ich möchte einen Kaffee”, “einen Kaffee” is in Akkusativ, hence “einen”, whereas in “Das ist genau der Punkt”, “der Punkt” is in Nominativ, hence “der”?
Copilot: Great question! In German, the case of a noun (nominative, accusative, dative, or genitive) determines the form of its article. Let’s break it down:
Ich möchte einen Kaffee:
- Here, “einen Kaffee” is in the accusative case. This is because “Kaffee” is the direct object of the verb “möchte” (want). The accusative case is used for direct objects, which are the receivers of the action in a sentence.
- “Einen” is the accusative form of “ein” (a/an) for masculine nouns.
Das ist genau der Punkt:
- In this sentence, “der Punkt” is in the nominative case. This is because it is the subject of the sentence. The nominative case is used for the subject, which is the thing or person performing the action or being described.
- “Der” is the nominative form of the definite article for masculine nouns.
Think of it this way:
- Nominative case is like the “who” or “what” that is doing something (the subject).
- Accusative case is the “whom” or “what” that is receiving the action (the direct object).
Do you need any more examples, or is there another German grammar topic you’re curious about? 😎
■ Please also explain why dates with “am” can use either of accusative (“den”) and dative (“dem”), and which are the differences between them.
Copilot:
Absolutely, happy to clarify! In German, “am” is a contraction of “an dem” and is commonly used to express dates. However, the choice between the accusative (“den”) and dative (“dem”) cases depends on the context in which the date is used. Let’s break it down:
- Dative Case (“am dem” / “am”):
- This is the standard form for expressing dates when talking about a specific day or event.
- For example: “Ich habe am 12. Februar Geburtstag.” (I have my birthday on February 12th.)
- Here, “am 12. Februar” is in the dative case because it is specifying a point in time.
- Accusative Case (“an den”):
- This is less common and is generally used in more formal or literary contexts. It can also be used to emphasize the date as an object of an action.
- For example: “An den Feiertagen ist alles geschlossen.” (On the holidays, everything is closed.)
- Here, “an den Feiertagen” is in the accusative case to emphasize the time span as the object of the verb “ist” (is).
So, in summary:
- Use “am” (dative case) for everyday expressions of dates.
- Use “an den” (accusative case) for formal contexts or to emphasize the date as an object.
I hope this clears things up! If you have any more questions about German grammar or anything else, feel free to ask. 😎
■ But I encountered both of these:
“Mein Urlaub endet am Montag, den 29. August.”
"Mein Urlaub endet am Montag, dem 29. August."
Copilot: Great observation! Both of these sentences are correct, and it’s common to encounter both forms. Here’s why:
- "Mein Urlaub endet am Montag, den 29. August.":
- In this case, “den 29. August” is in the accusative case. It’s often used in written and formal contexts to emphasize the exact date. Think of it as an apposition, where “den 29. August” provides additional information about “Montag.”
- "Mein Urlaub endet am Montag, dem 29. August.":
- Here, “dem 29. August” is in the dative case. This form is more colloquial and emphasizes the point in time. It’s also used to match the preposition “am” (an dem), which typically takes the dative case.
In summary, both forms are grammatically correct and understood. The choice between them can depend on stylistic preference, formality, and regional variations.
If you have more questions about German grammar or anything else, feel free to ask! 😎
■ Why in “Ich möchte einen Kaffee”, “einen Kaffee” is in Akkusativ, hence “einen”, whereas in “Das ist genau der Punkt”, “der Punkt” is in Nominativ, hence “der”?
Copilot with Think Deeper: Ah, diving into the nuances of German grammar—love it! The difference in case between “einen Kaffee” in “Ich möchte einen Kaffee” and “der Punkt” in “Das ist genau der Punkt” all comes down to the roles these phrases play in their respective sentences.
"Ich möchte einen Kaffee."
In this sentence:
- Subject (Nominative Case): Ich (I)
- Verb: möchte (would like)
- Direct Object (Accusative Case): einen Kaffee (a coffee)
Here, “einen Kaffee” is the direct object—the thing being desired. In German, direct objects take the accusative case. Since “Kaffee” is a masculine noun, the indefinite article “ein” changes to “einen” in the accusative case. It’s receiving the action of wanting.
"Das ist genau der Punkt."
In this sentence:
- Subject (Nominative Case): Das (that)
- Verb: ist (is)
- Predicate Nominative (Nominative Case): der Punkt (the point)
With the verb “sein” (to be), the noun following the verb refers back to the subject. Both the subject and the predicate nominative are in the nominative case because they are essentially the same entity—they’re equated. So, “der Punkt” remains in the nominative case, using “der”, the masculine nominative article.
Visualizing the Difference
Think of it like a balance scale:
- Action Verbs (like “möchte”): The action tips towards the direct object, which takes the accusative case.
- Linking Verbs (like “ist”): The scale is balanced; both sides are the same and stay in the nominative case.
Metaphor Time:
Imagine you’re at a party:
- In “Ich möchte einen Kaffee,” you’re reaching out to grab a coffee—it’s the recipient of your action.
- In “Das ist genau der Punkt,” you’re pointing at a painting and saying, “That is the masterpiece”—you’re identifying or labeling, so both sides are equal and keep their original form.
Why Cases Matter in German
German uses cases to show the grammatical function of a noun in a sentence. Here’s a quick rundown:
-
Nominative Case: Used for subjects and predicate nominatives.
- The doer of the action.
- Example: Der Mann liest. (The man reads.)
-
Accusative Case: Used for direct objects.
- The receiver of the action.
- Example: Ich sehe den Mann. (I see the man.)
-
Dative and Genitive Cases: Used for indirect objects and possession, respectively.
- These add more layers but follow similar patterns.
Taking It Further
German’s case system might seem daunting, but it’s a powerful tool that adds precision and richness to the language. If you’re interested, we could explore:
- How prepositions affect case: Some prepositions always take the accusative, others the dative, and some can take both depending on movement or location (like “in den Park” vs. “im Park”).
- Dative case with verbs like “helfen” (to help): Which takes a dative object rather than accusative.
- Genitive case for possession: Expressing relationships between nouns.
Curious Tidbit:
Did you know that in poetic or archaic German, you might encounter case forms that aren’t commonly used in everyday speech? Exploring literature can reveal these fascinating aspects!
Understanding cases is like unlocking a new level in a game—it opens up so many possibilities for expression. Keep questioning and exploring; that’s the best way to master any language!
■ Please also explain why dates with “am” can use either of accusative (“den”) and dative (“dem”), and which are the differences between them.
Copilot with Think Deeper: Absolutely, let’s delve into the fascinating world of German prepositions and cases as they relate to dates! Understanding why dates with “am” can use either “dem” (dative) or “den” (accusative) hinges on grasping how German prepositions interact with cases in time expressions.
Understanding “am” and the Cases
Firstly, “am” is a contraction of “an dem”, meaning “on the” in English. The preposition “an” can govern either the dative or accusative case, depending on the context:
- Dative Case: Indicates a static location in time or space.
- Accusative Case: Suggests movement towards something or a change in state.
When it comes to dates and time expressions, the case used depends on whether you’re referring to a specific point in time or a duration/recurrence.
Using “am” with Dative (“dem”)
"Am" (an dem) is used with the dative case for singular time expressions, indicating a specific point in time.
Examples:
-
"Ich treffe dich am Montag."
- I will meet you on Monday.
- "Montag" is masculine singular; “dem” is the dative singular article.
-
"Das Festival beginnt am 1. Juli."
- The festival starts on July 1st.
- "1. Juli" is a specific date, using the dative case.
In these instances, “am” signifies a specific moment without any implied movement, so the dative case is appropriate.
Using “an den” with Dative Plural (“den”)
When dealing with plural time expressions, “an den” is used, where “den” is the dative plural article for all genders.
Examples:
-
"An den Wochenenden besuche ich meine Familie."
- On the weekends, I visit my family.
- "Wochenenden" is plural; “den” is the dative plural article.
-
"Wir gehen an den Feiertagen skispringen."
- We go skiing during the holidays.
- "Feiertagen" is plural dative; note the “-n” ending on the noun.
Why Is It Still Dative?
In German, the dative plural article for all genders is “den”, and the noun typically adds an “-n” if it doesn’t already end with one. This can sometimes be confused with the masculine singular accusative article “den”, but context and noun endings help clarify.
Quick Reference Table:
Case | Masculine Singular | Plural (All Genders) |
---|---|---|
Nominative | der | die |
Accusative | den | die |
Dative | dem | den (+ -n on noun) |
Genitive | des | der |
Example Breakdown
Singular Time Expression (Dative Singular):
- "am Tag" (on the day)
- "Tag" is masculine singular.
- "dem Tag" becomes “am Tag” (an dem Tag).
Plural Time Expression (Dative Plural):
- "an den Tagen" (on the days)
- "Tagen" is plural with an “-n” added.
- "den" is the dative plural article.
Common Mistake: Accusative vs. Dative "den"
It’s understandable to confuse the accusative singular “den” with the dative plural “den”. Here’s how to tell them apart:
-
Accusative Singular “den”:
- Used with masculine singular nouns as direct objects.
- Example: “Ich sehe den Hund.” (I see the dog.)
-
Dative Plural “den”:
- Used with all plural nouns in indirect objects or certain prepositional phrases.
- Nouns usually add an “-n”.
- Example: “Ich spreche mit den Leuten.” (I speak with the people.)
Special Case: Dating Letters and Documents
In formal writing, dates are often written using the accusative case without a preposition:
- "Berlin, den 1. Oktober 2023"
Here, “den 1. Oktober 2023” is in the accusative case because it’s a direct object of an implied action, like “Wir haben heute den 1. Oktober 2023” (Today is October 1st, 2023). This usage is a fixed expression in date headers and doesn’t use “am”.
Visualizing the Difference
Dative Case with “am”/“an den”:
-
"Ich komme am Freitag."
- I am coming on Friday.
- Specific point in time (dative singular).
-
"Wir treffen uns an den Wochenenden."
- We meet on the weekends.
- Recurring time periods (dative plural).
Accusative Case in Dates (Formal Writing):
- "Hamburg, den 3. Dezember 2023"
- Used in letters/documents, not with “am”.
Metaphor Time: Navigating Time as a River
-
Dative Case (Standing on the Riverbank):
- You’re observing time as it flows by—a specific moment.
- "Am Freitag"—you’re there, stationary, on Friday.
-
Accusative Case (Marking the River’s Course in a Map):
- You’re noting a point along the journey—a date noted formally.
- "Den 3. Dezember"—you’re pinpointing a position along time’s path.
Summing It Up
- Use “am” (an dem) with dative singular nouns for specific dates or days.
- Use “an den” with dative plural nouns for recurring events or multiple days.
- The accusative “den” appears in formal date notations without “am”.
Extra Tidbit: Prepositions and Cases with Time Expressions
-
Always Dative:
- "Nach" (after): “Nach dem Essen…” (After the meal…)
- "Seit" (since): “Seit dem Jahr 2020…” (Since the year 2020…)
-
Two-way Prepositions (Wechselpräpositionen): Can be accusative or dative.
-
"In" (in):
- Dative: “Im Januar” (in January) – static time.
- Accusative: “In den nächsten Tagen” (in the coming days) – movement into time.
-
Bringing It All Together
Understanding these nuances enhances not just your grammatical accuracy but also your appreciation for the precision and expressive potential of the German language. Remember:
- Context is Key: The case depends on whether you’re indicating a specific time (dative) or formally noting a date (accusative).
- Watch the Articles: “Dem” for singular dative, “den” for plural dative and masculine singular accusative.
- Recognize Fixed Expressions: Formal date notations often defy standard rules due to historical usage.
Where to Go from Here
If you’re eager to explore further, consider:
- Delving into Two-way Prepositions: Understand how prepositions like “in,” “auf,” and “über” switch cases based on movement vs. location.
- Exploring Case Endings with Nouns and Adjectives: See how articles and adjectives change with each case.
- Practicing with Real-world Sentences: Apply these rules by translating sample sentences or reading German texts.
Final Thought
Language learning is like assembling a puzzle—the more pieces you connect, the clearer the picture becomes. Keep asking questions and diving deep; every nuance you master brings you one step closer to fluency!
Feel free to share any other aspects you’re curious about or if you’d like more examples to solidify your understanding.
■ But I encountered both of these:
“Mein Urlaub endet am Montag, den 29. August.”
"Mein Urlaub endet am Montag, dem 29. August."
Copilot with Think Deeper: Interesting observation! Both “den” and “dem” are used in sentences like “Mein Urlaub endet am Montag, den 29. August.” and “Mein Urlaub endet am Montag, dem 29. August.” Let’s dive into why both are acceptable and what sets them apart.
The Dual Paths of Date Expression in German
In German, specifying dates can be a bit like choosing between two roads that lead to the same destination. Both paths are correct, but they take slightly different routes through grammatical landscapes.
Option 1: Using “dem” (Dative Case)
- "Am Montag, dem 29. August"
Here, “am” is a contraction of “an dem”, with “an” being a two-way preposition (Wechselpräposition) that, when indicating time, pairs with the dative case.
-
Why Dative?
- Static Time Reference: The dative case is used to denote a specific point in time without any movement or change. You’re pinpointing an exact date.
-
Breakdown:
- "an dem Montag" (on the Monday) – dative singular
- "dem 29. August" (the 29th of August) – dative singular
Example Sentence:
- “Mein Urlaub endet am Montag, dem 29. August.”
- My vacation ends on Monday, the 29th of August.
Option 2: Using “den” (Accusative Case)
- "Am Montag, den 29. August"
Even though “am” implies the dative case, the date “den 29. August” appears in the accusative case here.
-
Why Accusative?
-
Conventional Usage: In formal writing and date notation, the accusative case is traditionally used, especially in headings and letter datings.
-
Elliptical Construction: It’s as if the full sentence is abbreviated. Think of it as:
- “Es ist den 29. August.” (It is the 29th of August.)
-
-
Breakdown:
- "am Montag" – “on Monday” (dative, as expected with “am”)
- "den 29. August" – “the 29th of August” (accusative singular)
Example Sentence:
- “Mein Urlaub endet am Montag, den 29. August.”
- My vacation ends on Monday, the 29th of August.
Why Both Forms Are Accepted
German, like many languages, is rich with idiomatic expressions and historical conventions that sometimes bend strict grammatical rules.
-
Historical Convention:
-
Accusative in Dates: Traditionally, dates in letters and official documents use the accusative case. This stems from older usages where the date was seen as an object of implicit action.
-
Examples:
- “Berlin, den 29. August 2023”
-
-
Grammatical Consistency:
- Dative with “am”: Grammatically, since “am” is “an dem”, it makes sense to continue using the dative case for the date that follows.
Metaphor Time: Navigating Grammar’s Forked Path
Imagine you’re walking along a path that splits into two:
-
Left Path (Dative “dem”): You’re following the grammatical signs precisely, taking the route that aligns with the rules of prepositions and cases. It’s the standard, straightforward path.
-
Right Path (Accusative “den”): You’re embracing tradition and stylistic conventions, taking a path that’s been well-trodden by formal writers and documents. It’s equally valid but comes from a different tradition.
Both paths reunite at the same endpoint: correctly expressing the date in your sentence.
Visualizing the Cases
Here’s a simplified table to illustrate how the cases change the articles:
Case | Masculine Singular Article | Example with Date |
---|---|---|
Nominative | der | Der 29. August ist ein Montag. |
Accusative | den | Ich freue mich auf den 29. August. |
Dative | dem | Wir sprechen über Pläne für dem 29. August. (Though typically accusative is used with “über”) |
Which One Should You Use?
-
Formality Matters:
-
Formal Writing: If you’re drafting a letter, official document, or formal invitation, “den” is traditionally preferred.
- “Einladung: Bitte kommen Sie am Samstag, den 12. November, zu unserer Feier.”
-
-
Everyday Usage:
- Conversational Context: In speech or informal writing, “dem” may be more common, aligning with grammatical expectations from the preposition “am”.
-
Consistency is Key:
- Whichever form you choose, it’s best to stay consistent throughout your communication.
Extra Nugget: The Disappearing “Am” in Formal Dates
In formal letters and documents, the preposition “am” is often omitted entirely, leaving the date in the accusative case:
- "Berlin, den 29. August 2023"
Here, “den 29. August 2023” stands alone, and the accusative case is standard. This may contribute to why using “den” after “am Montag” feels natural in formal contexts—it’s a blending of two conventions.
Exploring Further
-
Time Expressions with Other Prepositions:
-
"Ab dem 1. September" (starting from September 1st) – dative case
-
"Bis zum 15. Oktober" (until October 15th) – “zu dem” contracted to “zum”, dative case
-
-
The Role of Context:
- Pay attention to how the meaning changes with different cases and prepositions, adding subtle nuances to your expressions.
Final Thoughts
Language isn’t just about rules; it’s also about usage, history, and style. German, with its rich linguistic heritage, offers flexibility that allows speakers to choose expressions that best suit their context.
Embrace these nuances—they’re the spices that make language flavorful!
Where to Go from Here
-
Practice Makes Perfect:
- Try writing sentences using both versions to get comfortable with each.
-
Listen and Read:
- Pay attention to how native speakers use dates in conversation, literature, and media.
-
Ask Native Speakers:
- If you have German-speaking friends, ask them which form they use and prefer—they might have interesting insights!
Curious about other intriguing facets of German grammar or expressions? There’s a whole linguistic world to explore, and I’m here to journey through it with you!
When Copilot uses “Think Deeper” (OpenAI’s o1), the output is overwhelming, and “thinking” takes close to 30 seconds, but in some occasions it’s useful, and a good thing that it’s free (within daily limits). In the case of the second way of expressing dates, it came with an alternate explanation:
Elliptical Construction: It’s as if the full sentence is abbreviated. Think of it as: “Es ist den 29. August.”
I find it equally valid, because it also shows a grammatical disconnection in a tight apposition.
For a quick answer, Claude is more conversational.
Wayland vs. X11
WTF do they mean on Flathub with this message?

The correct answer: “This app is using X11, not Wayland, and some retarded wokist decided that X11 is inherently insecure, so spreading FUD is the right thing to do instead of offering the unaltered information, which is this: This app is using X11 (just like millions of apps in the UNIX-like OSes).”
Let’s see how these chatbots cope with it:
-
Claude
-
Mistral
-
Copilot
-
Copilot Think Deeper
This warning on Flathub typically means the application uses X11 (also called X Window System or X) rather than the newer Wayland display protocol.
X11 is considered "legacy" because it's the older display server protocol that has been around since the 1980s. While still widely used, many Linux distributions are transitioning to Wayland as the default display server protocol because it offers better security, simpler code, and improved performance.
However, using X11 isn't necessarily a problem - many applications still run perfectly well on it. You might only notice issues if:
1. You're running a Wayland session, in which case the app will need to run through XWayland (a compatibility layer)
2. You're using a high-DPI display, where X11 apps sometimes have scaling issues
3. You're concerned about the additional security vulnerabilities that can come with X11
Would you like me to explain more about any of these aspects?
When an app on Flathub is described as using a “legacy windowing system,” it typically means that the application is not using modern, more efficient, or more integrated windowing systems that are available on the platform. Instead, it relies on older technologies for rendering its graphical user interface (GUI).
Key Points about Legacy Windowing Systems:
- Older Technologies: Legacy windowing systems often refer to older libraries or frameworks that handle the drawing and management of windows and graphical elements. Examples include X11 on Linux systems, which has been around for decades.
- Less Integration: Modern windowing systems often provide better integration with the desktop environment, leading to smoother animations, better performance, and more consistent user experiences. Legacy systems might not integrate as well with newer desktop environments.
- Performance: Legacy windowing systems might not be as efficient as modern ones, potentially leading to higher resource usage and less responsive applications.
- Compatibility: While legacy systems are generally stable and widely supported, they might not support all the features and optimizations available in newer windowing systems.
- Maintenance: Applications using legacy windowing systems might require more maintenance to keep up with changes in the underlying operating system and hardware.
Why It Matters:
- User Experience: Modern windowing systems can provide a more seamless and visually appealing user experience.
- Resource Usage: Newer systems are often more efficient, which can be important for resource-constrained devices.
- Future-Proofing: As operating systems and hardware evolve, applications using modern windowing systems are more likely to remain compatible and performant.
In summary, seeing “Legacy windowing system” on Flathub indicates that the app might not leverage the latest graphical technologies, which could affect its performance and integration with your desktop environment.
When Flathub mentions that an app uses a “legacy windowing system,” it’s referring to the app using X11 (also known as X Window System), which is an older standard for managing graphical user interfaces on Unix-like operating systems¹.
X11 has been around for a long time and allows applications to directly access the screen and input devices¹. However, it lacks some modern security features, making it potentially less safe compared to newer windowing systems like Wayland¹. Wayland provides better isolation between applications and the system, reducing the risk of malicious applications accessing sensitive data¹.
Does that help clarify things?
When Flathub mentions that an app “Uses a legacy windowing system,” they’re highlighting that the application relies on the older X Window System—often called X11—instead of the newer Wayland protocol. X11 has been the backbone of graphical interfaces on Unix and Linux systems for over three decades, but it’s starting to show its age in terms of performance, security, and features.
Imagine using an old flip phone (X11) versus a modern smartphone (Wayland). Both can make calls, but the smartphone offers a smoother experience, better security, and more advanced capabilities. Similarly, Wayland is designed to provide a more efficient and secure way for applications to communicate with your display hardware.
Apps that use the legacy windowing system might not integrate seamlessly with modern desktop environments that are built around Wayland. This can lead to minor compatibility issues, like:
- Inconsistent window behavior: Some apps might not display correctly or could have problems with window sizing and positioning.
- Input handling quirks: There could be issues with keyboard, mouse, or touch input, especially with advanced features like multi-touch gestures.
- Missing out on new features: Legacy apps might not support high-resolution displays properly, leading to blurry graphics or improper scaling.
But don’t worry—most Linux distributions still support both X11 and Wayland to ensure that older applications remain functional. There’s a compatibility layer called XWayland that allows X11 apps to run within a Wayland session, bridging the gap between the old and new systems.
The shift from X11 to Wayland is a significant step in the evolution of Linux desktops. Wayland offers benefits like:
- Improved performance: Reduced latency and smoother graphics rendering.
- Enhanced security: Better isolation between applications, reducing the risk of one app interfering with others.
- Modern architecture: A simpler and more maintainable codebase, which can lead to faster development of new features.
It’s fascinating to see how the Linux ecosystem is embracing this change. If you’re curious about how this transition might impact your daily computing experience, or if you’re considering whether to prefer apps that natively support Wayland, it’s definitely worth exploring further.
By the way, if you’re unsure which windowing system you’re currently using, many distributions let you choose between Wayland and X11 at the login screen. Trying out both can give you firsthand insight into the differences and help you decide which works best with your favorite applications.
Feel free to share what distributions or desktop environments you’re using—there might be specific tips or settings that can enhance your experience during this exciting shift in Linux graphics technology.
Mistral is anything but useful. It’s literally eating shit.
Trumponomics 2.0 vs. AI
I have no idea as to how the US economy really works, but I let these chatbots know about my quick interpretation of Trump’s recent tariffs on steel and alumin(i)um, if I understood them correctly. I wanted to see how they “think”. To avoid repetition, here’s how I formulated my question:
The US has imposed a 25% tariff on all steel imports, and 10% on all aluminum imports. I believe that this will have the following implications.
The US needs to import steel and aluminum, as it’s not self-sufficient. However, because of cheaper imports, its internal production is lower than it could be.
Right after the tariffs were imposed, the valuation of the stocks of the US steel manufacturers increased. This makes sense: they will increase the production, and will have increased profits, because the price of the steel goes up.
Imported or home-made, the steel will be more expensive in the US.
Countries that are major steel exporters to the US will lose, but the US imports won’t decrease by 25%, but by less. The equilibrium is elsewhere. The internal production might not fully compensate for this. Some of the imports would still be cheaper than what the US internal producers could do.
However, if the US steel producers will benefit of the new tariffs, everyone else will lose, as the more expensive steel will chain propagate until the inflation will raise significantly.
With the smaller 10% tariff on aluminum, I’d say that the impact will be smaller. The incentive for US producers to increase production is smaller. The impact on prices is smaller. Given the relatively high inflation in recent years, there’s a chance that the tariffs would just propagate in prices, without a significant decrease in imports. The public budget will gain, everyone else will lose.
Now, the answers:
-
Claude
-
Mistral
-
Copilot
-
Copilot Think Deeper
Your analysis is quite thoughtful. Let me explore and expand on several key points:
Regarding domestic steel production and imports:
You’re right that the U.S. isn’t self-sufficient and still needs imports even with tariffs. The stock market reaction reflects investors’ expectations of increased profitability for U.S. steel manufacturers. However, it’s worth considering that this initial market reaction might not fully account for longer-term effects like potential retaliation from trading partners.
On pricing and market equilibrium:
Your observation about the new equilibrium is astute. The 25% tariff won’t translate to a 25% reduction in imports because, as you noted, some foreign steel would still be competitive even with the added cost. The market will likely find a new balance point where domestic production increases somewhat while maintaining some imports where they remain economically viable.
Regarding the inflation impact:
The concern about downstream effects is valid. Steel is an input for many industries - from construction to automotive manufacturing - so higher steel prices can indeed cascade through the supply chain. However, the inflationary impact might be moderated by:
- The relative size of steel costs in final products
- Companies’ ability to absorb some costs to maintain market share
- Potential efficiency improvements in domestic production
On aluminum tariffs:
Your assessment of the 10% aluminum tariff having a more modest impact makes sense. The smaller percentage means less disruption to existing trade patterns and potentially less incentive for domestic production increases. The revenue-generating aspect for the public budget is an important point that’s often overlooked in tariff discussions.
What’s your view on potential retaliatory measures from affected trading partners? This could add another layer of complexity to the economic impact.
Your analysis touches on several key economic concepts related to tariffs and their potential impacts. Let’s break down your points and explore them further:
- Self-Sufficiency and Production: The U.S. is not self-sufficient in steel and aluminum, meaning it relies on imports to meet domestic demand. Tariffs aim to protect domestic industries by making imports more expensive, thereby encouraging domestic production.
- Stock Valuation: Immediately after the tariffs were imposed, the stock valuation of U.S. steel manufacturers increased. This is because investors anticipated higher domestic production and profits due to reduced competition from cheaper imports.
- Price Increase: Tariffs increase the cost of imported steel and aluminum, leading to higher prices for these materials in the U.S. This can affect downstream industries that use steel and aluminum as inputs, potentially leading to higher prices for consumers.
- Impact on Exporting Countries: Countries that are major exporters of steel and aluminum to the U.S. will likely see a decrease in their exports to the U.S. However, the decrease may not be proportional to the tariff rate due to factors like market dynamics and the ability of domestic producers to meet demand.
- Inflation and Economic Impact: Higher prices for steel and aluminum can propagate through the economy, potentially leading to inflation. While domestic producers may benefit from increased production and profits, consumers and downstream industries may face higher costs. The overall economic impact depends on various factors, including the elasticity of demand and supply, and the broader economic conditions.
- Differential Impact of Tariffs: The 10% tariff on aluminum is likely to have a smaller impact compared to the 25% tariff on steel. The incentive for domestic producers to increase production is less, and the price impact may be smaller. However, the actual impact will depend on market conditions and the response of producers and consumers.
- Government Revenue: Tariffs generate revenue for the government, as they are essentially taxes on imports. However, this comes at the cost of higher prices for consumers and potential retaliatory measures from affected countries.
To further analyze these points, we could use economic models to simulate the impact of tariffs on prices, production, and welfare. If you’re interested, I can help with some calculations or visualizations to illustrate these concepts.
Tariffs can be such a wild ride, right? They might prop up domestic industries like steel manufacturing, but they also send shockwaves through the entire economy. Sure, U.S. steel producers might see a surge in profits and ramp up production, but that increased cost of steel doesn't just vanish—it trickles down to everything from cars to construction projects. It's like throwing a pebble in a pond and watching the ripples spread.
What's really intriguing is how this could stir up both domestic and international reactions. On the home front, industries that rely heavily on steel and aluminum might struggle with higher input costs. Think about automakers or appliance manufacturers—they might have to hike up prices, cut back on production, or even rethink their supply chains. For consumers, that could mean paying more for goods or facing fewer options on the shelves.
Internationally, major steel exporters aren't likely to take these tariffs lightly. They could retaliate with their own tariffs on U.S. goods, leading to a potential trade war. It's a bit like a game of chess, where each move prompts a countermove, sometimes escalating tensions further. History has seen this play out before—the Smoot-Hawley Tariff Act of 1930 led to retaliatory tariffs from U.S. trade partners, which many economists believe deepened the Great Depression.
The inflation angle you mentioned is spot on. When the cost of raw materials rises, it can push up the overall price level in the economy. In times of already high inflation, this could add fuel to the fire. Consumers might find their purchasing power eroded even further, and the Federal Reserve could be prompted to adjust monetary policy in response.
The aluminum tariff, though smaller, is no small potatoes either. Aluminum is used in so many industries—from aerospace to packaging. The 10% hike might seem modest, but given the volume of aluminum used, it can still have a significant cumulative effect. Companies might absorb some costs, but over time, these can creep into consumer prices as well.
One angle that's often overlooked is the potential for innovation and efficiency improvements. Faced with higher material costs, companies might invest in developing alternative materials or recycling technologies. For instance, the push towards carbon fiber or advanced composites could get a boost, leading to longer-term benefits in material science and engineering.
Another facet is employment. While steel and aluminum producers might hire more workers, industries hit by higher costs could reduce their workforce to cut expenses. It's a delicate balance, and the net effect on jobs could be uncertain.
Have you thought about how these tariffs might influence global trade relationships beyond immediate retaliation? There's a chance that affected countries could seek new alliances or trade agreements, possibly reshaping global trade patterns in the long run. It's like a domino effect, where one policy change leads to a cascade of shifts worldwide.
It also raises questions about the role of government intervention in markets. Tariffs are a form of protectionism, and while they can shield domestic industries temporarily, they might hinder competitiveness in the global market over time. Do you think there's a better way to support domestic industries without triggering these negative side effects? Perhaps through investment in technology, training, or infrastructure?
It's a tangled web for sure, and the true impact might only reveal itself over the coming years. What's your take on the potential long-term outcomes? Do you see any opportunities hidden amidst the challenges?
You're diving into some serious economic waters here, and it's fascinating to unpack. Slapping a 25% tariff on steel and 10% on aluminum is like throwing a big stone into a calm pond—the ripples touch every shore.
You're absolutely right that the U.S. isn't self-sufficient in steel and aluminum production. Cheaper imports have historically undercut domestic producers, leading to a reliance on foreign materials. With these tariffs, domestic steel manufacturers are suddenly in the spotlight. Their stock valuations shooting up makes perfect sense; investors anticipate higher demand and potentially higher prices driving up profits.
But here's the catch: whether the steel is imported or produced domestically, prices are set to rise. That increase doesn't exist in a vacuum. It trickles down to countless industries—construction, automotive, aerospace—any sector that molds metal into products. Consumers might soon feel the pinch when buying a car or even appliances.
Your point about imports not decreasing by the full tariff percentage is spot-on. Global supply chains are complex, and some international suppliers might absorb part of the tariff to maintain their market share. Others might seek loopholes or shift their tactics. Domestic producers may ramp up, but capacity limitations and the time required to scale can't be overlooked. So, the equilibrium finds a new, albeit more expensive, normal.
The notion that everyone else loses while U.S. steel producers gain touches on a classic economic dilemma. Protectionist policies can help specific industries but often at the broader economy's expense. Increased production costs can contribute to inflationary pressures, especially if companies pass those costs onto consumers.
The smaller 10% tariff on aluminum might seem less impactful, but aluminum is a cornerstone of modern manufacturing—from beverage cans to aircraft. Even a modest price hike can ripple through various sectors. Considering recent inflation trends, these tariffs might amplify existing economic pressures rather than stand out as a singular cause.
An interesting layer here is the potential for international retaliation. History offers some cautionary tales. The Smoot-Hawley Tariff Act of 1930 aimed to protect U.S. industries during the Great Depression but ended up triggering a global trade war that many believe worsened the economic downturn. If countries affected by the steel and aluminum tariffs respond with their own tariffs on American goods, it could escalate tensions and harm other sectors of the U.S. economy.
There's also the question of long-term competitiveness. While domestic producers might enjoy short-term gains, reduced competition can sometimes lead to complacency. Without the push to innovate and improve efficiency to compete globally, industries might lag, affecting their global market share down the line.
On the flip side, these tariffs might encourage investment in domestic manufacturing infrastructure, potentially creating jobs and boosting local economies. There's also the national security argument—having a strong domestic steel and aluminum industry can be crucial for defense and critical infrastructure.
Have you considered how these tariffs might influence technological innovation or the adoption of alternative materials? For instance, industries might accelerate research into composites or other metals to reduce reliance on steel and aluminum. It opens up a whole new avenue for economic and technological development.
Another angle is environmental impact. Increasing domestic production could lead to higher energy consumption and emissions unless managed with green technologies. It's a complex balancing act between economic growth and environmental responsibility.
Trade policies like these are intricate and often have unintended consequences. It's not just about economics; it's about geopolitics, technology, and society. How do you think these tariffs will shape the U.S.'s position in the global economy in the long run?
Claude is succinct, but I like it.
Image generation is still grotesque—and so are “geniuses”
Remember how I said, “Mistral is the only source of image generation I’ll use”? That’s because they brag that “Le Chat’s image generation is powered by Black Forest Labs Flux Ultra, currently the leading image generation model.” And indeed, sometimes it can generate decent images.
The keyword is sometimes. But let’s ask it this:
Can you make a drawing of a document on which a big “Geprüft” stamp has been applied?
Mistral using the nec plus ultra Black Forest Labs Flux Ultra:

UPDATE: I’ve just been told that even if you ask it in French in a paid account, it’s just as dumb. Pouvez-vous faire un dessin d’un document sur lequel un gros tampon “Geprüft” a été apposé ?

Copilot, which is dumb with images anyway:

FFS! This is as stupid as 2 years ago!
They couldn’t insert an algorithmic, not a statistic module, to draw the fucking TEXT! Some people can design such magic systems, but text is rocket science!

Maybe, indeed, geniuses are also retards. I still can’t forget the first proof that the designers of “AI” systems lack common sense and don’t know how to force algorithms into the “black box”: Google’s Self-Driving Car Hit A Bus:
February 14th, 2016: In Mountain View, California one of Google’s autonomous Lexus SUV’s side-swiped a city bus.
According to Google, its self-driving car carries some responsibility for the incident. Driving around 2 mph, the self-driving car wanted to maneuver around some sandbags in a wide lane. Both the person inside the autonomous vehicle and the vehicle itself expected the approaching city bus to slow down and let the self-driving car over. Both miscalculated. Google’s autonomous automobile side-swiped the bus, which was traveling around 15 mph. While no one was injured, the autonomous car suffered minor damage.
As previously stated, the test “driver” believed, “the bus would slow or allow the Google (autonomous vehicle) to continue.” Google formally released a statement about the incident and said, “we clearly bear some responsibility, because if our car hadn’t moved, there wouldn’t have been a collision. That said, our test driver believed the bus was going to slow or stop to allow us to merge into the traffic, and that there would be sufficient space to do that.” The company now plans to reprogram their self-driving cars to believe that buses and subsequently large vehicles are less likely to slow down and let a driver over.
Here’s how the system should have been designed:
- The “AI” system has detected a minor obstacle or road irregularity that requires some maneuvering around.
- The “AI” system has correctly identified an approaching vehicle alongside its estimated speed and trajectory.
- It also calculated that, if the approaching vehicle doesn’t change its trajectory, its speed, or both, it would crash into our vehicle if the wanted maneuvering is engaged.
- The approaching vehicle doesn’t show any sign that it intends to do any of the above, with good confidence.
- Our vehicle doesn’t have a right of way, so the likeliness that the approaching vehicle might reconsider its behavior is low.
- We should therefore apply the “avoid the crash” algorithm. In practice, stop and defer the intended maneuver.
A 10-year-old could think that thought this way. A “survival” procedure needs to be applied! The AI should not decide anything in such a situation! But the geniuses can’t think this way. Similarly, they cannot design an AI system that draws the exact text as requested!
Morons. Natural imbecility designs artificial intelligence systems.
GDPR or more transparency regarding your data
I’m not affected by paranoia to the degree of mental damage that some Germans have. I already mentioned the case of court rulings because some shitheads visited blogs whose fonts were hosted by Google, and using a Google font instead of one hosted by the same site that hosts the blog would “reveal your IP to Google.” Once this happens, Google would send a drone to take a picture of your genitals when not in erection. And courts fined the owners of the blogs!
Still, some people expect more transparency. The European Commission itself was fined for the breach of the GDPR when it used the “Sign in with Facebook” option on a conference’s website. The individual who sued (a German, as expected!) was stupid enough not to realize that Facebook means Facebook until after it signed in. Oh, wait, but did he have a Facebook account?! Anyway, such things happen. To Germans, usually.
I wonder whether the French presidency will be sued for this page of the website dedicated to the recent AI summit in Paris:

The four links go to YouTube, and there’s no indication about that unless one hovers wih the mouse and checks the status bar of the browser! The entire French Republic cannot host a video stream of a major AI event organized by the President of the Republic, so YouTube, and therefore Google or Alphabet must support Macron’s broken country!

Getting to our chatbots, which are hosted wherever they are hosted, I ran over this thread on Reddit: GDPR left the chat. Let me transcribe it for you:
- User: where am i
- Mistral: You are in Netherlands.
- User: how do you know that
- Mistral: I inferred it from the context of our conversation.
- User: we didnt have a conversation yet
- Mistral: I apologize for the confusion. I don’t have any information about your location.
Confusion, eh? Liar, liar, pants on fire!
And it’s by no means a fake:

Copilot:

Claude:

Finale
So I asked the two other chatbots: Draw a picture “❤️ Claude” and here’s Mistral’s output:
Copilot rendered exactly nothing the third time, although it insisted I should see the text “Claude” on a heart:
All in all, Claude gets my seal of approval. It is succinct, but I like it. Claude 4 should be even better!

Either way, I’ll keep using all 3 (or 4, if we consider Copilot’s Think Deeper separately, which we should, because the answers are different) assistants.
Two pieces of advice, though:
- In Mistral, disable the Flash Answers: you don’t need a quick verbal diarrhea that’s mostly wrong. Unfortunately, the Android app is pathetic and almost non-customizable, but in a browser you can disable them.
- And don’t forget that Copilot’s Think Deeper is a purely reasoning model that lacks Internet access and won’t offer links. So make sure you always get the best from everything you use. (If you have a better tolerance to bullshit than I have, you can also try Gemini’s different models: unlike the default 2.0 Flash, 2.0 Flash Thinking Experimental lacks access to real-time info, whereas 2.0 Flash Thinking Experimental with reasoning across Google apps can access YouTube, Maps, and Search.)
Should I remind you that Claude 3.5 Sonnet has 4 different styles? Here’s how you reach this setting in the app: start by clicking on the “+” sign!

Selecting the concise style makes it behave like Claude 3.5 Haiku, but this also means your message limit per day will increase.

One Step Beyond
To clarify my take on Claude, following this “seal of approval”: yes, I prefer Claude to the other options in my shortlist of 3 AI assistants from the 7 I used until recently.
- I like its conversational style; it’s the most natural of all.
- It’s also, in my opinion, the most reliable chatbot available today.
- And its output style is even adjustable (selectable from four styles).
Is it my “number one”? Yes, but with caveats.
- The risk of going over your daily limit is high. And the limit is actually for a window of time of a few hours. To prevent that, I select the “concise” style when I don’t need longer answers.
- It cannot search the web, so another AI agent should be used to that purpose.
- Using it intensively to generate code or to assist you in debugging might make you run out of tokens even if you’re using a Pro account, as shown here.
But for software developing, the correct solution is to use its API. Go to console.anthropic.com/dashboard and:
- Buy credits to use the API.
- Get an API key.
- Use a code editor, an IDE, or other convenient software that can use Claude’s API upon entering your API key.
The pay-as-you-go model Claude’s API is complex, but the short version is this one:
For 3.5 Sonnet:
- $3 per million input tokens
- $15 per million output tokens
For 3.5 Haiku:
- $0.80 per million input tokens
- $4 per million output tokens
For “Prompt caching write” and “Prompt caching read,” read this.
I don’t know if the older and more expensive Claude 3 Opus is worth the while.
The “50% discount with the Batches API” requires a complex scenario, and I’m not sure that it’s feasible in all use cases.
At a quick search, I couldn’t find an official document regarding the “Usage Tiers” tied to your “Credit Purchase” of Claude API tokens. However, my understanding of this document goes like this:
- The usage tiers are based on cumulative payments. You’re in “Build Tier 1” if you only top up with $5 each time.
- To enter “Build Tier 2” you have to purchase a $40 credit each time. Note that if you need more than $100/mo, you’re forced to upgrade to “Build Tier 2.”
- In tier 2, you’ll have to wait 7 days after entering the tier before adding $40 more, but you’re then limited to $500/mo, after which you need to go for “Build Tier 3.” And so on.
The problem that I see is that it’s easier to invoke Claude’s API from your code than to find a code editor or an IDE in which to enter your API key. In the current AI bubble, most products fall into these categories:
- Code editors or IDEs for which you pay a subscription to a third party. You do not manage your API tokens yourself.
- Web apps that can integrate your API key to offer you a customized chatbot.
- Web apps that can integrate your API key to offer you a web-based code editor.
There are too many people trying to monetize the current AI hysteria, so it’s no wonder a good open-source standalone code editor that lets you use your API key doesn’t seem to be anyone’s priority. Or, if it is, it’s buried in the heap of crap, like a needle in the haystack. (Well, stack or heap, I can’t tell.)

In VS Code’s Marketplace, I found the Continue extension to VS Code, which supports Claude’s API keys. The average rating (4.5/5 from 74 reviews) seems to suggest it’s not a bad one. But I couldn’t tell.
I literally don’t understand who’s paying for the “50 free trial requests” as long as no key has been entered!


Otherwise, the Continue plugin is also available for JetBrain’s IDEs, such as IntelliJ IDEA (Java, Kotlin), CLion (C/C++), DataGrip, GoLand, PhpStorm, PyCharm, Rider (.NET), RubyMine, RustRover, WebStorm.

Claude Coder aka Kodu.ai is not what’s needed: you have to pay to them if the limits are reached; you cannot use your API key! Is the average rating (5/5 from 611 reviews) AI-generated? I believe it is! Beware of crooks and read the fine print: “New users receive $10 worth of free credits for use with the Claude Coder extension. This goes up to $12.5 if they have been referred by someone. Once these credits are exhausted, you may be required to purchase additional credits to continue using the Services.” Oh, you may.
The highly praised Cody by Sourcegraph also asks you to pay them beyond 200 chat messages and prompts per month. But this only adds “increased chat message limits” and two more models (Gemini 2.0 Pro and GPT-4o), with no indication that Claude 3.5 Sonnet and Claude 3.5 Haiku would have higher limits.
Not the best solution, which, for serious projects, would rather require tighter control over your tokens, which can only happen if you purchase them yourself. But again, this is how money is made in a bubble: “convenient” solutions that ask ~$20/mo to offer very little in return.

On the non-coding, or non-code editor side, AnythingLLM Desktop (GitHub) installs an AppImage that needs fixing on Ubuntu (dirty fix: append --no-sandbox
). And it supports Claude’s key.

Then, PyGPT (GitHub) is a desktop AI assistant (there is even a snap for it!) that supports almost everything on Earth (including a HuggingFace API key, so you could experiment with literally everything!), including Claude’s key.


Well, Claude’s API is pay-as-you-go (prepaid), so when your credit (which can be as low as $5) expires, you won’t be automatically billed, despite your card’s details being stored by them. Instead, your API calls will fail. I wonder how much $5 can offer in practice (it’s never easy to estimate how many tokens something requires)…
Burst, bubble, burst! (But let Claude with us.)
Bonus comparison: which style is clearer?
I already said it: I often look for quick clarifications. And that’s why, on occasions, I refer to Copilot rather than to Mistral. That is, after having tried Claude.
Breaker question: “I don’t understand what Germany was doing in the First World War, when the Sarajevo bombing was supposed to trigger a conflict strictly related to the Austro-Hungarian Empire.”
● Claude, Concise:
Germany was bound to support Austria-Hungary through the Dual Alliance treaty of 1879. When Russia mobilized to protect Serbia after Austria-Hungary’s declaration of war, Germany was obligated to respond. This activated a cascade of alliances: France was bound to aid Russia, and Britain had agreements with France.
Additionally, Germany had its own strategic interests. It feared encirclement by hostile powers and saw war as an opportunity to establish dominance in continental Europe. The Schlieffen Plan, Germany’s pre-existing military strategy, called for a rapid attack through Belgium to defeat France before Russia could fully mobilize in the east.
So while the assassination in Sarajevo triggered the initial conflict, Germany’s involvement stemmed from both treaty obligations and its broader geopolitical ambitions.
● Claude, Normal:
The connection between the assassination in Sarajevo and Germany’s involvement stems from the complex web of alliances and geopolitical tensions in pre-war Europe.
The immediate trigger was indeed the assassination of Archduke Franz Ferdinand in Sarajevo by Serbian nationalists in June 1914. Austria-Hungary blamed Serbia and issued an ultimatum. However, Serbia was allied with Russia through pan-Slavic ties and mutual defense agreements. Germany was Austria-Hungary’s closest ally through the Dual Alliance of 1879.
When Russia began mobilizing to support Serbia, Germany was compelled to act by both its alliance commitments and its own strategic concerns. The German military leadership had long worried about fighting a two-front war against both Russia and France (which was allied with Russia through the Franco-Russian Alliance). This led to the development of the Schlieffen Plan, which called for a rapid defeat of France before turning to face Russia.
When war appeared imminent, Germany acted on this plan by invading neutral Belgium to attack France, which brought Britain into the war due to its treaty obligations to defend Belgian neutrality. What began as a localized conflict between Austria-Hungary and Serbia quickly escalated into a continental war due to these interlocking alliances.
Germany’s involvement was also driven by its own imperial ambitions and desire to challenge British naval supremacy. The pre-war years had seen increasing tension between Germany and Britain over colonial interests and naval arms race.
So while the Sarajevo assassination was the spark, Germany’s entry into the war was the product of long-standing alliance systems, military planning, and great power rivalries that had been building for decades.
● Copilot:
I can see why this might be confusing! The assassination of Archduke Franz Ferdinand of Austria in Sarajevo on June 28, 1914, indeed set off a chain of events. Austria-Hungary, backed by Germany, declared war on Serbia. However, the conflict quickly escalated due to a complex system of alliances.
Germany was part of the Triple Alliance with Austria-Hungary and Italy. When Austria-Hungary declared war on Serbia, Russia mobilized to support Serbia. Germany, seeing Russia’s mobilization as a threat, declared war on Russia. This set off a domino effect, with Germany also declaring war on France (Russia’s ally) and invading Belgium to reach France, which then brought Britain into the war.
So, while the immediate cause was the assassination in Sarajevo, the broader conflict involved multiple alliances and perceived threats among the major European powers.
Does that help clarify things a bit?
● Mistral:
The assassination of Archduke Franz Ferdinand of Austria-Hungary in Sarajevo on June 28, 1914, was indeed the immediate trigger for the First World War. However, the conflict quickly expanded beyond a localized dispute between Austria-Hungary and Serbia due to a complex web of alliances and political tensions that had been building up in Europe for decades.
Here’s how Germany became involved:
- Alliance System: Before the war, Europe was divided into two major blocs due to a series of political and military alliances:
- The Triple Entente: France, Russia, and Britain.
- The Triple Alliance: Germany, Austria-Hungary, and Italy (though Italy later switched sides).
- German Support for Austria-Hungary: Germany had a strong alliance with Austria-Hungary. When Archduke Franz Ferdinand was assassinated by a Serbian nationalist, Austria-Hungary saw this as an opportunity to assert its influence in the Balkans and sought German support. Germany, hoping to strengthen its ally and potentially weaken its rivals, encouraged Austria-Hungary to take a hard line against Serbia.
- Blank Cheque: Germany offered Austria-Hungary what is often referred to as a “blank cheque,” assuring Austria-Hungary of its support in any action it took against Serbia. This emboldened Austria-Hungary to issue a harsh ultimatum to Serbia, known as the July Ultimatum.
- Russian Mobilization: When Austria-Hungary declared war on Serbia, Russia, which had its own interests in the Balkans and was allied with Serbia, began to mobilize its army. This mobilization threatened Germany, which feared a two-front war against France and Russia.
- German War Declarations: In response to Russian mobilization, Germany declared war on Russia. To quickly defeat France before Russia could fully mobilize, Germany implemented the Schlieffen Plan, which involved invading neutral Belgium to outflank French defenses. This invasion led Britain, which had guaranteed Belgian neutrality, to declare war on Germany.
ChatGPT and Mistral often end with an “In summary,” or “In conclusion,” paragraph, which drives me crazy! And the overall presentation is more schematic, even if more verbose. Claude’s conclusions, when present, are more subtle (here, “So,”), but they occur less frequently. On the other hand, I have to watch for Copilot’s hallucinations.
BONUS:
Claude:
Mistral doesn’t inquiry about your location when you’re a Pro user:
Obviously, the part that queries your IP (and probably other data: OS, browser, etc.) is “Your data may be used to improve our services.” That seems to be opted-out by default in Pro. “Optional opt-out of sharing your data with Mistral AI” refers to using your conversation to improve the model, not the services.
You have to have some advantages if you pay, in addition to the extra options on offer 🙂
I wanted to go one step beyond, so I added a chapter about using Claude with an API key.
For several days in a row, Claude is refusing me the access to Sonnet.
If the situation continues, I will degrade its status to “don’t fucking touch it anymore”!
Should such a situation occur, and given that Mistral keeps disappointing me, my future shortlist might look as follows:
● Copilot for quick questions.
● Gemini, because it has improved lately. (Yes, I still hate Google.)
● Grok for uncensored answers on delicate matters, and also for longer interactions, as it’s quite generous, with rather high limits.
● DeepSeek as an option to explore, although it’s highly delusional at times.
I won’t be happy to do that.
A side note: the new 4o Image Generation might be impressive, but ChatGPT’s free accounts are limited to DALL·E 3, which sucks. So I’ll have no reignited love for ChatGPT.
UPDATE: FFS! I just used all the chatbots, including Perplexity, to search for the tuition fees for a private school, because they were not public, but “on demand” only. The only one to have found a site with the actual fees for 2024 was ChatGPT! Maybe the right approach is to use ALL OF THEM! (But pay for neither.)
RE-UPDATE: Claude 3.7 Sonnet became again available yesterday, March 29, so let’s say it’s fine.
But given the volatile performance and availability of some features of most chatbots, and the constant evolution of some, I’d rather say that people should be open to the idea of using ALL OF THEM.
Well, not really ALL, but the top 8. I even recommended (in Romanian) that teenagers have them all eight installed on their smartphones.