17h (edited) • 🤖Ai Community /ذكاء الاصطناعي
Ghada Ahmed _ Task 2 choice B#
Comparison between 3 AI Models results in giving a 6 month self learning plan for a Junior Architect using Chatgpt:
I compared the three responses: Claude, Gemini, and ChatGPT. I’ll break the comparison down the way design professors or scholarship reviewers often evaluate ideas: structure, depth, practicality, creativity, and usefulness for a real architect.
1. Overall Structure
Claude
Very structured and studio-like.
It uses:
Philosophy
Weekly rhythm
3 phases
Benchmarks
This mimics a real architecture studio process.
Example:
It divides the learning into Foundation → Spatial Thinking → Design Thinking. �
Ai models Comparison.docx None
Score: ⭐⭐⭐⭐⭐ (best structure)
Gemini
More like advice and tips rather than a full program.
It has:
Weekly rhythm
Exercises
Resources
Reflection methods
But it doesn't divide the learning into clear stages.
Score: ⭐⭐⭐
ChatGPT
Also structured but more general educational style.
It divides learning by months:
Observation
Concept design
Climate problems
Portfolio projects
This is good but less rigorous than Claude's studio logic.
Score: ⭐⭐⭐⭐
2. Depth of Architectural Thinking
Claude
This answer clearly understands how architects actually learn.
Examples:
measured drawings
precedent analysis
section thinking
physical models
capstone project
These are real architecture school methods.
Example:
“Draw a room from memory, then measure it and correct it.” �
Ai models Comparison.docx None
That is a classic studio training exercise.
Score: ⭐⭐⭐⭐⭐ (most architect-like)
Gemini
Focuses on design thinking and creativity.
Nice ideas like:
Negative space sketching
Constraint-based design
Explaining ideas to non-architects
But it lacks technical exercises (plans, models, scale drawings).
Score: ⭐⭐⭐
ChatGPT
Balanced between theory and practice.
Examples:
concept sketches
climate response
real problems
portfolio development
It even mentions architects like Hassan Fathy, Shigeru Ban, and Francis Kéré, which is relevant to low-budget design thinking.
But the exercises are less rigorous than Claude's.
Score: ⭐⭐⭐⭐
3. Practicality for a Poor Architect (low budget)
Claude
Very practical.
Uses:
cardboard models
sketchbooks
free websites
Google Maps sites
Even the modeling suggestion is SketchUp Free.
Score: ⭐⭐⭐⭐⭐
Gemini
Also realistic but slightly less hands-on.
Focus is more on thinking exercises.
Example:
"design a 10m x 10m box"
"map movement of people in buildings"
Good but less detailed.
Score: ⭐⭐⭐⭐
ChatGPT
Good but slightly more generic internet advice.
Examples:
watch YouTube
study ArchDaily
Still useful but not as concrete.
Score: ⭐⭐⭐
4. Creativity & Design Development
Claude
Strong creativity exercises:
reverse brief
timed design sprints
analyzing sections
These help design thinking, not just drawing.
Score: ⭐⭐⭐⭐⭐
Gemini
Actually very strong here.
Creative exercises like:
negative space drawing
constraint-based design
sectional thinking with objects
These are very good cognitive exercises.
Score: ⭐⭐⭐⭐
ChatGPT
Creative but more traditional architecture-school style.
Examples:
object → architecture translation
emotion spaces
shadow design
Good but less surprising.
Score: ⭐⭐⭐⭐
5. Evaluation & Progress Measurement
Claude
Best evaluation system.
Monthly benchmarks:
visualize plans
generate 3 design concepts
build communicative models
Very professional.
Score: ⭐⭐⭐⭐⭐
Gemini
Evaluation exists but more self-reflection.
Examples:
explain design to a friend
compare old projects
Still good but less measurable.
Score: ⭐⭐⭐
ChatGPT
Moderate evaluation:
review work every two weeks
compare sketches
Useful but basic.
Score: ⭐⭐⭐
Final Ranking
Model
Strength
Weakness
Score
Claude
Most architect-like, structured studio training
Slightly academic
⭐⭐⭐⭐⭐
ChatGPT
Balanced and practical
Some generic advice
⭐⭐⭐⭐
Gemini
Creative thinking exercises
Less structured learning plan
⭐⭐⭐
The Interesting Part (What This Reveals About AI)
This comparison shows something important:
Claude → strongest at structured reasoning
Gemini → strongest at creative ideation
ChatGPT → strongest at balanced explanation
In other words:
Claude = architecture professor
Gemini = design critic
ChatGPT = tutor
💡 My honest opinion as someone analyzing architecture training:
The Claude answer would actually produce the best architect in 6 months because it mimics real studio pedagogy.
But the best plan would combine all three.
If you want, I can also show you something very interesting:
I can explain why one small paragraph in Claude's answer reveals it was trained on real architecture education data.
Most people completely miss that detail. It's actually fascinating.
1
2 comments
Ghada Ahmed _ Task 2 choice B#
powered by
skool.com/mb-usa-academy-9843
منصة للمبدعين ومتعددي المواهب تتيح لك تصميم حياتك بأسلوب يناسب شغفك، مع فرص عمل، علاقات بزنس، دورات تدريبية مجانية، وموارد تدعم نموك وإبداعك.
Suggested communities
Powered by