Forwarded from AddiTech (Alireza Ghiasvand)
Which AI Model Has the Best Reasoning Skills? 🤖
🧩 The Challenge:
We tested five top-tier Large Language Models (LLMs) with a mind-bending logical puzzle—a 3-digit code mystery hidden within five ambiguous and contradictory clues! 💡
🔥 LLMs in the Battle:
🟢 GPT-4.0
🟣 GPT-4.03
🔴 Claude 3.7 Sonnet
🔵 Grok 3
🟡 DeepSeek R1
@AddiTech
🧩 The Challenge:
We tested five top-tier Large Language Models (LLMs) with a mind-bending logical puzzle—a 3-digit code mystery hidden within five ambiguous and contradictory clues! 💡
🔥 LLMs in the Battle:
🟢 GPT-4.0
🟣 GPT-4.03
🔴 Claude 3.7 Sonnet
🔵 Grok 3
🟡 DeepSeek R1
📌 Results:
✅ GPT-4.03
Final Answer: ✅ Correct! (832) 🎯
Reasoning Power: 🌟 Exceptional (identified contradictions & resolved them!)
Processing Speed: ⚡️ Moderate (42 sec)
❌ GPT-4.0
Final Answer: ❌ Incorrect (382)
Reasoning Power: 🔥 Decent but flawed
Processing Speed: ⚡️ Moderate (35 sec)
❌ Claude 3.7 Sonnet
Final Answer: ❌ Incorrect (378)
Reasoning Power: 🚀 Fast but lacked depth
Processing Speed: ⚡️ Super Fast (28 sec)
❌ Grok 3
Final Answer: ❌ Incorrect (584)
Reasoning Power: 🚀 Quick but superficial
Processing Speed: ⚡️ Lightning Fast (23 sec)
❌ DeepSeek R1
Final Answer: ❌ Completely Wrong (5482 & 582)
Reasoning Power: 🐌 Struggled with logic
Processing Speed: ⚡️ Moderate (40 sec)
@AddiTech
👍4
Forwarded from AddiTech (Alireza Ghiasvand)
Alireza Ghiasvand
Photo
Scientific Analysis & Insights:
💬 What do you think? Which model do you prefer, and why? 🤔🚀
✍️Author:
@Ghiasvand_Engineering
@AddiTech
Only GPT-4.03 successfully used Chain-of-Thought reasoning to detect and resolve contradictions. 🔥
Claude 3.7 & Grok 3 had high-speed processing but lacked deep analytical skills. 🏃♂️
DeepSeek R1 had the worst performance due to major flaws in logical processing & maintaining information coherency. 🚨
🔬 How AI Models Process Logical Problems:
✅ Text Understanding (NLU & Encoder)
✅ Working Memory Retention
✅ Step-by-Step Deduction (Chain-of-Thought Processing)
✅ Contradiction Resolution (Logical Inference & Conflict Handling)
🚀 Final Verdict:
🥇 GPT-4.03 is the clear winner! This model dominated logical reasoning, accurately solved the puzzle, and even pointed out an inconsistency in the clues! 👏💡
📌 Pro Tip: If you need a model for high-stakes logical analysis and problem-solving, GPT-4.03 is the best pick! 🔥
✨ What’s Next?
The future of LLMs depends on enhancing deep logical reasoning & improving the balance between speed and accuracy! 🌐🤖
💬 What do you think? Which model do you prefer, and why? 🤔🚀
✍️Author:
@Ghiasvand_Engineering
@AddiTech
👍3
Forwarded from AddiTech (Alireza Ghiasvand)
ContentStudio_video.mp4.gif
2.5 MB
🔹 What is Generative Design?
Generative design is an AI-driven engineering approach that automatically generates multiple optimized design alternatives based on user-defined inputs like materials, constraints, and performance goals. It mimics nature’s evolution process to find the best possible design.
🔹 How Can It Solve Our Challenges?
✅ Reduces Weight & Material Waste → Optimized structures with less material but higher strength
✅ Speeds Up Product Development → AI explores thousands of designs in hours, not weeks
✅ Cost-Effective Manufacturing → Creates ready-to-produce designs for 3D printing, CNC, and more
✅ Enhances Product Performance → Finds the strongest, lightest, and most efficient designs
🔹 Why is Generative Design the Future?
By automating complex design processes, companies can innovate faster, reduce costs, and create breakthrough products that were previously impossible to design manually.
@AddiTech
Generative design is an AI-driven engineering approach that automatically generates multiple optimized design alternatives based on user-defined inputs like materials, constraints, and performance goals. It mimics nature’s evolution process to find the best possible design.
🔹 How Can It Solve Our Challenges?
✅ Reduces Weight & Material Waste → Optimized structures with less material but higher strength
✅ Speeds Up Product Development → AI explores thousands of designs in hours, not weeks
✅ Cost-Effective Manufacturing → Creates ready-to-produce designs for 3D printing, CNC, and more
✅ Enhances Product Performance → Finds the strongest, lightest, and most efficient designs
🔹 Why is Generative Design the Future?
By automating complex design processes, companies can innovate faster, reduce costs, and create breakthrough products that were previously impossible to design manually.
@AddiTech
👍3😍2👏1
Forwarded from AddiTech (Alireza Ghiasvand)
This media is not supported in your browser
VIEW IN TELEGRAM
𝗔𝗩𝗢𝗜𝗗 𝟯𝗗 𝗣𝗥𝗜𝗡𝗧𝗜𝗡𝗚 𝗘𝗥𝗥𝗢𝗥𝗦 𝗪𝗜𝗧𝗛 𝗦𝗜𝗠𝗨𝗟𝗔𝗧𝗜𝗢𝗡 🖥
At Danish Technological Institute, experts use simulations to optimize additive manufacturing and achieve 'first-time-right' results 📈
Using Simufact software, AM design specialist Andreas Weje Larsen predicts errors like deformation, cracks, shrink lines, and recoater contacts in titanium (Ti6Al4V). Aluminium (AlSi10Mg) is next ⚙️
Below is a test build for AMSIS GmbH, where shrink lines were predicted and later confirmed in the printed part 😊
The goal: Predict and fix errors before printing 👍
@AddiTech
At Danish Technological Institute, experts use simulations to optimize additive manufacturing and achieve 'first-time-right' results 📈
Using Simufact software, AM design specialist Andreas Weje Larsen predicts errors like deformation, cracks, shrink lines, and recoater contacts in titanium (Ti6Al4V). Aluminium (AlSi10Mg) is next ⚙️
Below is a test build for AMSIS GmbH, where shrink lines were predicted and later confirmed in the printed part 😊
The goal: Predict and fix errors before printing 👍
@AddiTech
👍4🔥1😍1
Forwarded from AddiTech (Alireza Ghiasvand)
This media is not supported in your browser
VIEW IN TELEGRAM
⚡️ Ansys Fluent Meets #SimAI: Revolutionizing Simulation ⚡️
🚀 What makes SimAI powerful?
1) Lightning-fast predictions: Evaluate performance in under a minute — design cycles reduced by 10-100X.
2) Massive design exploration: Quickly test and compare countless design iterations.
3) No-code AI experience: Designed for engineers and designers — no deep learning expertise needed.
One impressive example: SimAI's drag prediction on a new SUV geometry takes less than 1 minute, with an error of less than 0.5% compared to CFD, while maintaining accurate skin friction and wake topology predictions.
🔁 Left: Traditional Fluent CFD | Right: AI Prediction from #SimAI
Our LinkedIn Page
@AddiTech
Ansys SimAI, a cloud-enabled generative AI platform, is transforming the engineering landscape by delivering ultra-fast and accurate performance predictions across physics domains like fluid dynamics.
🚀 What makes SimAI powerful?
1) Lightning-fast predictions: Evaluate performance in under a minute — design cycles reduced by 10-100X.
2) Massive design exploration: Quickly test and compare countless design iterations.
3) No-code AI experience: Designed for engineers and designers — no deep learning expertise needed.
One impressive example: SimAI's drag prediction on a new SUV geometry takes less than 1 minute, with an error of less than 0.5% compared to CFD, while maintaining accurate skin friction and wake topology predictions.
🔁 Left: Traditional Fluent CFD | Right: AI Prediction from #SimAI
Our LinkedIn Page
@AddiTech
❤1🔥1
Forwarded from AddiTech (Alireza Ghiasvand)
This media is not supported in your browser
VIEW IN TELEGRAM
🧠 AI-Powered Crash Predictions by NAVASTO | Powered by Autodesk 🕓
Crash simulations that once took hours can now be performed in seconds thanks to the AI technology developed by NAVASTO, a company backed by Autodesk.
Originally built for sensitivity analysis, this AI model now enables real-time crash predictions. A striking example: a Toyota vehicle crashing into a wall, fully simulated with AI, showing accurate deformation and impact behavior all happening in real time.
This is a major step forward in engineering workflows, making simulation faster, smarter, and more accessible.
Our LinkedIn Page
@AddiTech
Crash simulations that once took hours can now be performed in seconds thanks to the AI technology developed by NAVASTO, a company backed by Autodesk.
Originally built for sensitivity analysis, this AI model now enables real-time crash predictions. A striking example: a Toyota vehicle crashing into a wall, fully simulated with AI, showing accurate deformation and impact behavior all happening in real time.
Why it matters:
• From hours to seconds: real-time crash analysis
• Fast evaluation of multiple design options
• Supports early decision-making
• Reduces physical testing needs
This is a major step forward in engineering workflows, making simulation faster, smarter, and more accessible.
Our LinkedIn Page
@AddiTech
🔥2
Forwarded from AddiTech (Alireza Ghiasvand)
🧠 Conceptual Design with Generative AI and CFD on AWS ☁️
AWS is transforming early-stage product development by combining Generative AI with high-performance CFD simulations in the cloud. This integration enables engineers and designers to rapidly generate, evaluate, and refine complex geometries based on performance targets all within a scalable, cloud-native environment.
This cloud-based workflow empowers R&D teams to move from idea to validated design faster and more efficiently, accelerating innovation across industries such as aerospace, automotive, and energy.
Our LinkedIn Page
@AddiTech
AWS is transforming early-stage product development by combining Generative AI with high-performance CFD simulations in the cloud. This integration enables engineers and designers to rapidly generate, evaluate, and refine complex geometries based on performance targets all within a scalable, cloud-native environment.
Using tools like Amazon SageMaker, NVIDIA Modulus, and Ansys Fluent, design teams can:
• Generate optimized geometry concepts in minutes
• Run CFD simulations at scale using AWS ParallelCluster and HPC infrastructure
• Apply AI-driven surrogate models for rapid performance prediction
• Significantly reduce iteration cycles in the conceptual phase
This cloud-based workflow empowers R&D teams to move from idea to validated design faster and more efficiently, accelerating innovation across industries such as aerospace, automotive, and energy.
Our LinkedIn Page
@AddiTech
😱2❤1
Forwarded from AddiTech (Alireza Ghiasvand)
🚗💡 Generative Design Boosts Formula Student Performance!
The application of Generative Design is revolutionizing how student teams develop their Formula Student race cars 🏎
By using topology optimization, engineers can find the ideal material distribution within components — making them lighter, stronger, and perfectly adapted to real-world loads and design constraints ⚙️✨
📍 A great example:
💡 Why it matters:
Reducing weight means less mass to accelerate — translating to:
✅ Faster acceleration
✅ Improved handling
✅ Lower energy consumption
🚀 Smarter design = better performance on the track!
Our LinkedIn Page
@AddiTech
The application of Generative Design is revolutionizing how student teams develop their Formula Student race cars 🏎
By using topology optimization, engineers can find the ideal material distribution within components — making them lighter, stronger, and perfectly adapted to real-world loads and design constraints ⚙️✨
📍 A great example:
The Elbflorace Formula Student Team from TU Dresden applied Generative Design to optimize their rock shafts.
🔧 Through additive manufacturing in titanium, they’ve cut the component’s weight by a massive 50% since the first iteration! 💪
💡 Why it matters:
Reducing weight means less mass to accelerate — translating to:
✅ Faster acceleration
✅ Improved handling
✅ Lower energy consumption
🚀 Smarter design = better performance on the track!
Our LinkedIn Page
@AddiTech
👍2❤🔥1👏1