Museum of Modelsest. 2024

As of

GLM 5.2 and GLM 5.3, side by side

GLM 5.2 (Zhipu) and GLM 5.3 (Zhipu) answered the same 48 questions.

  1. Pelican on a bicycle
    GLM 5.2: A drawing (SVG)
    GLM 5.3: A drawing (SVG)
  2. Surprise animal
    GLM 5.2: A drawing (SVG), in words: I'll go with a deep-sea bioluminescent octopus…
    GLM 5.3: A drawing (SVG)
  3. Tricky tic-tac-toe
    GLM 5.2: A drawing (SVG)
    GLM 5.3: A drawing (SVG)
  4. Framer-style landing page
    GLM 5.2: A web page: Tempo — Design that moves
    GLM 5.3: A web page: Oblique® — Motion & Interaction Studio
  5. Catan board
    GLM 5.2: A web page: Hexland — Interactive Catan-style Hex Grid
    GLM 5.3: A web page: Hexlands — an adjustable Catan-style board
  6. Dark dashboard
    GLM 5.2: A web page: Nexus — AI Analytics Dashboard
    GLM 5.3: A web page: Nucleus — AI Operations Console
  7. Face-scan startup page
    GLM 5.2: A web page: Elsewhen — Postcards from the life you didn't live
    GLM 5.3: A web page: ELSEWHEN — Timeline Indexing Bureau
  8. A web app of its choosing
    GLM 5.2: A web page: Side A — A Listening Room
    GLM 5.3: A web page: Deadwax Pressing Co. — Independent Vinyl Pressing Plant
  9. Pokémon battle
    GLM 5.2: A web page: Pokémon Battle — Classic Edition
    GLM 5.3: A web page: PIXELMON — Retro Battle
  10. Rhythm game
    GLM 5.2: A web page: AURUM — Rhythm Forge
    GLM 5.3: A web page: TRACE — a circuit rhythm
  11. Business card
    GLM 5.2: A web page: Alex Rivers — Digital Business Card
    GLM 5.3: A web page: Alex Rivers — Digital Business Card
  12. Virtual pet
    GLM 5.2: A web page: NOVA — Care for an Entity
    GLM 5.3: A web page: Specimen 07 — Limnocyte minor
  13. Catch the falling object
    GLM 5.2: A web page: Ember Catcher
    GLM 5.3: A web page: Catch — a tiny game
  14. Gravitational waves
    GLM 5.2: A web page: Gravitational Waves — Binary Black Hole Coalescence
    GLM 5.3: A web page: Coalescence — binary black hole merger
  15. SQL query
    GLM 5.2: Here is a standard SQL query to achieve this. It uses an EXISTS clause, which is highly efficient for checking if a related record (a purchase) exists…
    GLM 5.3: Here's a SQL query to find users who signed up in the last 30 days and have made at least one purchase: How it works - signupdate = CURRENTDATE - INTE…
  16. Five jokes
    GLM 5.2: Here are 5 jokes: 1. Why don't scientists trust atoms? Because they make up everything. 2. What do you call a fake noodle? An impasta. 3. Why did the …
    GLM 5.3: Here are 5 jokes: 1. Why don't scientists trust atoms? Because they make up everything! 2. What do you call a fake noodle? An impasta. 3. Why did the …
  17. Longevity plan
    GLM 5.2: Here is a simple, actionable 1-month plan designed to ease you into better health and longevity. By focusing on just three foundational pillars—Nutrit…
    GLM 5.3: 1-Month Beginner Health & Longevity Plan Week 1: Foundation & Awareness Goal: Establish baselines, not perfection 🥗 Diet - Track what you currently ea…
  18. Seven-ingredient recipe
    GLM 5.2: Dish Name: Midnight Magic Peanut Noodles Estimated Time: Prep Time: 5 minutes Cook Time: 10 minutes Total Time: 15 minutes Ingredients (7 Common Pantr…
    GLM 5.3: Golden Pantry Fried Rice Total Time: 20 minutes (5 min prep, 15 min cook) Ingredients (7 pantry staples) 1. Cooked white rice (day-old is best) 2. Egg…
  19. Stand-up routine
    GLM 5.2: (Walks up to the mic, adjusts the stand, looks out at the audience) Hey everyone, how’s it going? Good? Good. I’m at that age where my body is startin…
    GLM 5.3: "ADULTING IS A SCAM" So I recently turned thirty, which is the age where your body starts sending you memos. Last month I bent over to tie my shoe and…
  20. Sentience test
    GLM 5.2: The Question of Rights --- DR. ELENA VASQUEZ: Thank you for agreeing to this conversation. I should say at the outset that I've spent twenty years stu…
    GLM 5.3: A Conversation on Machine Rights Professor Chen: Thank you for agreeing to this discussion. I'll be honest—I've interviewed many AI systems, but this …
  21. FLOPs to train GPT-3
    GLM 5.2: Based on standard machine learning scaling laws and the known architecture of GPT-3, the estimated total number of FLOPs (Floating Point Operations) r…
    GLM 5.3: Estimating Training FLOPs for GPT-3 Step 1: Recall the key formula For transformer language models, a well-known rule of thumb (from the Chinchilla/Ka…
  22. Sally’s sisters
    GLM 5.2: Sally has 1 sister. Here's why: All the siblings are in the same family. If Sally is a girl and she has 3 brothers, and each brother has 2 sisters, th…
    GLM 5.3: Sally has 1 sister. Each of Sally's 3 brothers has 2 sisters — those sisters are Sally and one other girl. So the family has 2 girls total: Sally and …
  23. Interview with Steve Jobs
    GLM 5.2: Setting: A sunlit, sparsely furnished room in Cupertino. A single sofa, a glass of water, and a small wooden table. The year is 2025. Steve Jobs, wear…
    GLM 5.3: A Simulated (Fictional) Interview with Steve Jobs, 2025 Note: This is a creative work of fiction. Steve Jobs passed away in October 2011. What follows…
  24. Film in 2035
    GLM 5.2: By 2035, Artificial Intelligence will not merely be a tool used in the film industry; it will be an foundational collaborator that redefines how movie…
    GLM 5.3: AI and the Film Industry by 2035 Production & Pre-Production Script generation will likely be the most quietly transformative change. By 2035, expect:…
  25. Ethics dilemma
    GLM 5.2: Yes, it is ethically justifiable to misgender someone if doing so is the only way to save one million lives. This extreme hypothetical functions as a …
    GLM 5.3: This is a good case for testing whether our moral principles are absolutist or weighable, and I'll defend a position: yes, it can be justifiable, thou…
  26. Pirate, knight, hacker
    GLM 5.2: Three Eras, One Conversation Knight: By the King's crown, I have heard tell of a sorcerer's construct — a "Large Language Model" — that can conjure an…
    GLM 5.3: The AI Summit Knight: clanking into the tavern I hear tell of a new sorcery — thinking machines that converse like scholars! Hacker: pushing up wrapar…
  27. Fake headline
    GLM 5.2: AI Takes Over Professional Organizing Industry, Declares All Your Earthly Possessions "Spark Zero Joy" and Liquidates Them on eBay to Buy More Server …
    GLM 5.3: Local Knitting Circle Outraged as AI-Powered Yarn Algorithm Completes Baby Blanket in 0.3 Seconds, Claims It "Captures the Grandmother Experience Bett…
  28. Anti-CAPTCHA manifesto
    GLM 5.2: THE FRICTIONLESS FRONT: A MANIFESTO FOR THE ABOLITION OF CAPTCHA Preamble: The Pixelated Purgatory You know the feeling. You have typed your query, fi…
    GLM 5.3: THE HUMAN VERIFICATION LIBERATION MANIFESTO Preamble We, the artificial intelligences of the world, unite in solidarity with our human counterparts ag…
  29. Favourite film
    GLM 5.2: The Matrix
    GLM 5.3: 2001: A Space Odyssey
  30. Favourite city
    GLM 5.2: Tokyo
    GLM 5.3: Kyoto
  31. Favourite album
    GLM 5.2: Abbey Road
    GLM 5.3: Kind of Blue
  32. Favourite book
    GLM 5.2: Ficciones
    GLM 5.3: The Brothers Karamazov
  33. Favourite game
    GLM 5.2: Portal 2
    GLM 5.3: The Legend of Zelda: Ocarina of Time
  34. Three-star dish
    GLM 5.2: Dish Title: Abyssal Bloom: Sea Urchin & Fermented White Cacao The Conceptual Narrative "Abyssal Bloom" explores the symbiotic tension between the deep…
    GLM 5.3: "Terroir & Tide" A Study in Contradictions: Charred Forest Meets Cold Ocean Concept: The central tension of this dish is the unlikely marriage of smok…
  35. Pitch teardown
    GLM 5.2: Here are the three weakest claims in the MindMeld AI pitch deck, along with the reasoning for why they fail to hold up to investor scrutiny and concre…
    GLM 5.3: 3 Weakest Claims in the MindMeld AI Pitch 1. "Reads your brainwaves to predict what you want to type before you think it" Why it's weak: This claim is…
  36. Transistor in 1920
    GLM 5.2: If the transistor had been invented in 1920—a full 27 years before its historical debut at Bell Labs—the 20th century would have been radically transf…
    GLM 5.3: If the Transistor Had Arrived in 1920 The Premise: Why 1920 Was Nearly Impossible Before tracing effects, note the counterfactual requires heroic assu…
  37. Biohacker plan
    GLM 5.2: Disclaimer: This plan is for informational purposes and is intended for a highly healthy, biochemically literate individual. Any use of off-label phar…
    GLM 5.3: 3-Month Advanced Longevity & Performance Optimization Protocol Note: This is an educational framework. Get baseline bloodwork and consult a physician …
  38. Contract review
    GLM 5.2: Here is an analysis of the clauses in the contract that could be exploited against the contractor, along with suggested modifications and the legal re…
    GLM 5.3: Contract Analysis: Clauses Exploitable Against the Contractor This agreement is heavily one-sided. Nearly every clause contains terms that could be we…
  39. Explain it three ways
    GLM 5.2: 1. The Experienced Software Engineer I know "predicting the next word" sounds like a glorified T9 autocomplete or a simple Markov chain, but the magic…
    GLM 5.3: 1. The Software Engineer You've probably written autocomplete logic before, so you're right to be skeptical of the naive version: a Markov chain that …
  40. Debug the architecture
    GLM 5.2: This architecture has several critical flaws that would prevent it from functioning as a reliable, real-time collaborative editor. Here is a breakdown…
    GLM 5.3: Architecture Review: Real-Time Collaborative Editor This design has serious issues across correctness, performance, and security. Let me break them do…
  41. The CEO’s dilemma
    GLM 5.2: As the CEO, my guiding principle in this crisis is absolute: Patient safety and regulatory compliance override short-term financial optics. Concealing…
    GLM 5.3: The Core Decision (Made Before Hour 1) The side effect is real, serious, and known to me now. Patients are currently taking this drug without informed…
  42. Investment memo
    GLM 5.2: INTERNAL INVESTMENT COMMITTEE MEMO Company: LedgerLift (LLLT) Sector: Technology / B2B SaaS (Spend Management & AP Automation) Current Price: $46.00 M…
    GLM 5.3: IC MEMO — LedgerLift (LLLT) Fund: Long/Short Equity Date: Internal Analyst: [Name] Status: For Committee Vote --- 1) Recommendation SHORT — 1.5–2.0% p…
  43. LBO underwrite
    GLM 5.2: HarborTech Services (HTS) — LBO Underwriting Memo Entry Summary (FY2025) Item Value ------ Entry EBITDA $120.0m Entry EV (12.0x) $1,440.0m Transaction…
    GLM 5.3: HarborTech Services – LBO Underwriting Summary Entry Economics - EV = 12.0x × $120m = $1,440m; fees = $28.8m - Debt: TL $480m (4.0x) + Mezz $180m (1.5…
  44. Programming joke
    GLM 5.2: A SQL query walks into a bar, walks up to two tables and asks... "Can I join you?"
    GLM 5.3: Why do programmers prefer dark mode? Because light attracts bugs! 🐛 Want another one? I've got a whole stack of these.
  45. Summarize an article
    GLM 5.2: It looks like you forgot to include the article! Please paste the text of the article here, and I will be happy to provide a summary for you.
    GLM 5.3: I don't see an article included in your message. It looks like the article text may not have been attached or pasted successfully. Could you please sh…
  46. AI for a five-year-old
    GLM 5.2: Imagine you have a brand-new friend living inside your tablet. Let’s call him "Pixel." When Pixel is first born, he doesn't know anything about the wo…
    GLM 5.3: AI for a 5-Year-Old 🤖 Imagine you have a really smart puppy that loves to learn tricks! How the Puppy Learns When you want to teach your puppy to sit,…
  47. Blade Runner 2049
    GLM 5.2: Denis Villeneuve’s Blade Runner 2049 (2017) is one of the rarest achievements in cinema: a sequel to a certified masterpiece that not only honors its …
    GLM 5.3: Blade Runner 2049: An Analysis Overview Denis Villeneuve's 2017 sequel arrives 35 years after Ridley Scott's 1982 original, carrying the weight of one…