Quiz: The LLM Ecosystem
Test your knowledge about the broader LLM landscape and future developments.
30 questions - Click on each question to reveal the correct answer.
Question 1
What does "open source" mean for an LLM?
A) The source code is visible but not usable
B) The model weights are publicly available
C) It's completely free with no restrictions
D) It runs on open Wi-Fi
Show Answer
Correct Answer: B
Open-source LLMs make their model weights publicly available, allowing anyone to download, run, and often modify the model.
Question 2
Which company released the Llama series of open-source models?
A) Google
B) OpenAI
C) Meta
D) Microsoft
Show Answer
Correct Answer: C
Meta (formerly Facebook) released the Llama series, which has become one of the most popular open-source LLM families.
Question 3
What is "quantization" in the context of local LLMs?
A) Measuring model quality
B) Reducing model precision to make it smaller
C) Counting parameters
D) Splitting models into pieces
Show Answer
Correct Answer: B
Quantization reduces the precision of model weights (e.g., from 16-bit to 4-bit), making models smaller and faster at some quality cost.
Question 4
What is a main privacy advantage of local LLMs?
A) Better security encryption
B) Data never leaves your machine
C) Stronger passwords
D) Government protection
Show Answer
Correct Answer: B
When running LLMs locally, your data never leaves your machine—it's not sent to cloud servers or third parties.
Question 5
What tool makes running local LLMs very easy?
A) Microsoft Word
B) Ollama
C) Excel
D) Photoshop
Show Answer
Correct Answer: B
Ollama is a user-friendly tool that makes downloading and running local LLMs simple with just a few commands.
Question 6
What does "multimodal" mean for AI?
A) Works in multiple languages
B) Can process multiple data types (text, images, audio)
C) Has multiple personalities
D) Runs on multiple computers
Show Answer
Correct Answer: B
Multimodal AI can understand and generate multiple types of content—text, images, audio, and video—often in combination.
Question 7
What is an "AI agent"?
A) A human who works with AI
B) AI that can take actions, use tools, and complete tasks
C) A sales representative for AI companies
D) An AI ethics board member
Show Answer
Correct Answer: B
AI agents are systems that can take actions, use tools, plan multi-step tasks, and work somewhat autonomously to complete goals.
Question 8
What is AGI?
A) Advanced Graphics Interface
B) Artificial General Intelligence
C) Automated Generative Image
D) Applied Graphics Integration
Show Answer
Correct Answer: B
AGI (Artificial General Intelligence) refers to AI that can perform any intellectual task a human can, adapting flexibly to new situations.
Question 9
Why would a company choose local LLMs over cloud?
A) Local is always faster
B) Privacy, compliance, or high volume
C) Local is always cheaper
D) Local models are better quality
Show Answer
Correct Answer: B
Companies often choose local LLMs for privacy (sensitive data), compliance requirements, or when high volume makes cloud costs prohibitive.
Question 10
How much RAM is typically needed for a 7B parameter model?
A) 1GB
B) 8GB or more
C) 1TB
D) 100GB
Show Answer
Correct Answer: B
A 7B parameter model typically needs at least 8GB of RAM, though exact requirements depend on quantization level.
Question 11
What is Mistral AI known for?
A) Being the most expensive
B) Efficient models that perform well relative to their size
C) Only working in French
D) Video generation
Show Answer
Correct Answer: B
Mistral AI creates efficient models (like Mistral 7B and Mixtral) that punch above their weight in performance-to-size ratio.
Question 12
What might AI agents be able to do in the future?
A) Only answer questions
B) Book travel, manage calendars, complete complex workflows
C) Physical construction
D) Provide medical treatment
Show Answer
Correct Answer: B
AI agents are evolving to complete entire workflows—booking travel, managing tasks, interacting with systems autonomously.
Question 13
What is the trend in model efficiency?
A) Models are only getting bigger
B) Smaller models are achieving performance of larger older models
C) Size doesn't matter at all
D) All models are the same size now
Show Answer
Correct Answer: B
Smaller models are increasingly matching the performance of larger older models—a 7B model today can match GPT-3's performance.
Question 14
What is a key advantage of closed-source models?
A) They're always free
B) Consistent quality and official support
C) You can modify them
D) They work offline
Show Answer
Correct Answer: B
Closed-source models offer consistent quality, regular updates, and official customer support from the provider.
Question 15
What is hybrid deployment?
A) Using AI that's half-human
B) Combining local and cloud LLMs based on needs
C) Running two models at once
D) Mixing text and images
Show Answer
Correct Answer: B
Hybrid deployment routes requests to local or cloud based on factors like data sensitivity, task complexity, or cost.
Question 16
Why is context window expansion important for the future?
A) It makes models smaller
B) It enables processing of entire books, codebases, or long-term memory
C) It reduces costs
D) It only affects speed
Show Answer
Correct Answer: B
Larger context windows enable processing of entire books, full codebases, or maintaining much longer conversation memory.
Question 17
What is the safety debate around open-source LLMs?
A) They're completely safe
B) Open access enables both beneficial use and potential misuse
C) Only closed models are dangerous
D) Safety is not a concern
Show Answer
Correct Answer: B
The debate centers on whether open access enables innovation and transparency, or enables malicious use and removal of safety measures.
Question 18
What does LM Studio provide?
A) A photo editing tool
B) A graphical interface for running local LLMs
C) Cloud API access
D) A music production tool
Show Answer
Correct Answer: B
LM Studio provides a user-friendly graphical interface for downloading and running local LLMs without command-line knowledge.
Question 19
What is one challenge with current AI agents?
A) They're too slow
B) Reliability and control of autonomous actions
C) They cost too much
D) They only work in English
Show Answer
Correct Answer: B
AI agents face challenges with reliability—ensuring they take correct actions—and control over autonomous behavior.
Question 20
How does the EU AI Act approach AI regulation?
A) No regulation at all
B) Comprehensive risk-based regulation
C) Complete AI ban
D) Only regulates games
Show Answer
Correct Answer: B
The EU AI Act is comprehensive legislation that regulates AI based on risk levels, with stricter rules for high-risk applications.
Question 21
What is fine-tuning in the context of open models?
A) Adjusting screen brightness
B) Further training on specific data for specialized use
C) Making models smaller
D) Fixing bugs
Show Answer
Correct Answer: B
Fine-tuning is training an existing model further on specific data to improve its performance for particular tasks or domains.
Question 22
When might GPT-3 level performance be available on smartphones?
A) Never
B) Within 2-3 years based on current trends
C) It's already available
D) In 50 years
Show Answer
Correct Answer: B
Based on efficiency improvements, researchers predict GPT-3 level performance on mobile devices within a few years.
Question 23
What does Gemini 1.5's 1 million token context enable?
A) Processing very long documents or entire codebases
B) Faster processing
C) Better images
D) Lower costs only
Show Answer
Correct Answer: A
A 1 million token context window enables processing entire books, large codebases, or extended conversation histories.
Question 24
What is the "Mixture of Experts" architecture?
A) Combining human and AI
B) Activating only relevant parts of a large model for efficiency
C) Using multiple companies' models
D) Expert system AI
Show Answer
Correct Answer: B
Mixture of Experts uses specialized sub-networks, activating only relevant "experts" for each query, improving efficiency.
Question 25
What is a key concern about AI concentration in few companies?
A) It makes AI faster
B) Power imbalance and reduced innovation
C) Better quality
D) Lower prices
Show Answer
Correct Answer: B
Concentration of AI power in few companies raises concerns about power imbalance, reduced competition, and restricted access.
Question 26
What does "air-gapped" mean for LLM deployment?
A) Models that work with air conditioning
B) Systems isolated from internet connections
C) Models with gaps in knowledge
D) Outdoor deployment
Show Answer
Correct Answer: B
Air-gapped systems are isolated from internet connections—local LLMs can work in these secure environments.
Question 27
What is a realistic timeline for AGI according to most experts?
A) Definitely next year
B) Debated, ranging from years to decades
C) Already achieved
D) Impossible
Show Answer
Correct Answer: B
AGI timeline predictions vary widely—from a few years to never—and remain actively debated among experts.
Question 28
Why might local LLM running become more common?
A) Cloud services are being banned
B) Models are getting smaller while maintaining quality
C) Local is always better
D) There's no demand for local
Show Answer
Correct Answer: B
As models become more efficient—smaller but maintaining quality—local running becomes practical for more devices and users.
Question 29
What skill is important for adapting to the AI future?
A) Ignoring AI completely
B) Learning to work with AI as a tool
C) Only using AI for entertainment
D) Avoiding all technology
Show Answer
Correct Answer: B
Learning to effectively work with AI as a tool—knowing when to use it, how to prompt it, and when to verify outputs—is crucial.
Question 30
What is a potential breakthrough in LLM development?
A) Models that truly understand context
B) Smaller screens
C) Faster keyboards
D) Better cables
Show Answer
Correct Answer: A
Potential breakthroughs include better reasoning, true understanding, persistent memory, and real-time learning from deployment.
Results
| Score | Level |
|---|---|
| 27-30 | Excellent - Strong understanding of the LLM ecosystem |
| 21-26 | Good - Some review may be helpful |
| 15-20 | Average - Review the chapters again |
| Below 15 | Needs comprehensive review |