π 2026-10-10π C2 Proficientπ 10 new wordsπ§ daily podcast
π What is Muse English?
Muse English is a free English learning app with daily lessons across all 6 CEFR levels (A1 Beginner to C2 Proficient). Every day brings new vocabulary, a real podcast conversation, shadowing practice, grammar and quizzes β all at museenglishapp.com.
π Vocabulary
10 new wordsLearn new words with real examples and audio.
enshittification /en-shit-ih-fih-KAY-shun/
noun
the gradual decline of an online platform as it starts prioritizing profit over its users' experience
"Once the network effects kicked in, the platform shifted from delighting users to extracting value, a textbook case of enshittification."
agentic /ay-JEN-tik/
adjective
(of an AI system) able to act independently, pursuing goals and completing tasks on its own
"The assistant's agentic capabilities allow it to book flights, file expenses, and chase invoices without being told each step."
multimodal /mul-tee-MOH-dul/
adjective
able to understand and produce several kinds of input at once, such as text, images, and audio
"Multimodal models can read a chart, listen to a lecture, and draft a summary in a single pass."
tokenization /toh-kuh-nai-ZAY-shun/
noun
the process of splitting text into small pieces (tokens) that a language model can process
"Subword tokenization lets models handle rare words gracefully by breaking them into familiar pieces."
watermarking /WAW-tur-mahr-king/
noun
embedding hidden signals in AI-generated content so it can later be identified as machine-made
"Watermarking generated text offers regulators a fragile but useful tool for spotting synthetic content."
sandboxing /SAND-box-ing/
noun
running software in an isolated, controlled environment to test it without risk
"The team tests every suspicious attachment by sandboxing it on an isolated virtual machine."
quantization /kwon-tih-ZAY-shun/
noun
compressing an AI model by using simpler numbers, trading a little accuracy for much smaller size and faster speed
"Through quantization, the lab squeezed a flagship model onto a laptop with barely any loss in quality."
zero-shot /ZEE-roh-shot/
adjective
(of a model) able to handle a task it was never explicitly trained to do
"Remarkably, the model answered zero-shot, solving tasks it had never seen during training."
vibe coding /VYBE KOH-ding/
noun
writing software by describing what you want in plain language and letting AI generate the code
"Vibe coding has lowered the barrier to software, though purists still insist on understanding every line."
model collapse /MOD-ul kuh-LAPS/
noun
the degradation of an AI model that is trained mostly on AI-generated data
"If the web fills with synthetic text, future models risk model collapse as they train on their own output."
π§
Podcast: Technology
Listen to todayβs conversation.
π
Train your speaking
Shadow every sentence with studio audio and get your pronunciation scored β only in the Muse English app.
πGrammar: 8B: Reported speech β statements and questions
Reported speech tells us what someone said, without quoting their exact words. When the reporting verb (**said**, **told**, **asked**) is in the past, tenses usually shift back β this is **backshift**.
The shifts are systematic: present β past, present perfect β past perfect, **will** β **would**, **can** β **could**, **may** β **might**, **must** β **had to**.
You do NOT always backshift. Keep the original tense for general truths or facts still true now, and when you report something immediately after hearing it.
**Say** takes no person: she **said** she was tired. **Tell** takes a person
Master this grammar in the app
Get clear explanations, more examples and interactive exercises.