Uploaded February 2026 | Updated September 2026, 2 weeks ago
Get started with lingoda today and use my code JONESFEB for $25 off: try.lingoda.com/LanguagejonesLingoda
Stop using LLMs like a "deranged therapist" and start using them like a linguist. đź§
We hear about ChatGPT and AI nonstop, but let’s be honest: most people are using them in completely unhinged ways. In this video, I’m putting the big four—ChatGPT, Gemini, Perplexity, and Claude—to the test to see which one actually helps you acquire a language and which one is just a glorified autocomplete.
As a PhD linguist, I’m breaking down the morphosyntax, the feedback loops, and the hallucination rates of these models. Whether you’re a self-study polyglot or just trying to pass your intro classes, here is the objective truth about AI for language learning in 2026.
What We Cover:
• 0:00 Introduction
• 1:05 The unhinged way we use LLMs
• 2:10 I’m launching a course! Sign up for updates at languagejones.com/accelerator
• 2:50 Lingoda’s winter sale: try.lingoda.com/LanguagejonesLingoda
• 4:05 What’s wrong with LLMs
• 5:45 How I evaluated LLMs
• 8:15 Do they get the facts right? Hallucinated verb conjugations
• 12:00 Vocabulary Quality
• 13:00 Transliteration
• 13:53 Volume
• 14:50 Formatting
• 15:25 Images: Bad, worse, and weird
• 17:21 Overall usefulness of content
• 18:10 Grammar explanations
• 19:30 Overall Comparison
• 20:08 My personal winner and runner up
Links & Resources
• My Course: The Linguistics Blueprint: languagejones.com/blueprint
• The Language Accelerator: languagejones.com/accelerator
#LanguageLearning #AI #Linguistics #ChatGPT #GeminiAI #Polyglot #LanguageJones #EdTech
Get started with lingoda today and use my code JONESFEB for $25 off: try.lingoda.com/LanguagejonesLingoda
Stop using LLMs like a "deranged therapist" and start using them like a linguist. đź§
We hear about ChatGPT and AI nonstop, but let’s be honest: most people are using them in completely unhinged ways. In this video, I’m putting the big four—ChatGPT, Gemini, Perplexity, and Claude—to the test to see which one actually helps you acquire a language and which one is just a glorified autocomplete.
As a PhD linguist, I’m breaking down the morphosyntax, the feedback loops, and the hallucination rates of these models. Whether you’re a self-study polyglot or just trying to pass your intro classes, here is the objective truth about AI for language learning in 2026.
What We Cover:
• 0:00 Introduction
• 1:05 The unhinged way we use LLMs
• 2:10 I’m launching a course! Sign up for updates at languagejones.com/accelerator
• 2:50 Lingoda’s winter sale: try.lingoda.com/LanguagejonesLingoda
• 4:05 What’s wrong with LLMs
• 5:45 How I evaluated LLMs
• 8:15 Do they get the facts right? Hallucinated verb conjugations
• 12:00 Vocabulary Quality
• 13:00 Transliteration
• 13:53 Volume
• 14:50 Formatting
• 15:25 Images: Bad, worse, and weird
• 17:21 Overall usefulness of content
• 18:10 Grammar explanations
• 19:30 Overall Comparison
• 20:08 My personal winner and runner up
Links & Resources
• My Course: The Linguistics Blueprint: languagejones.com/blueprint
• The Language Accelerator: languagejones.com/accelerator
#LanguageLearning #AI #Linguistics #ChatGPT #GeminiAI #Polyglot #LanguageJones #EdTech
