Rules for What We Know, AI for What We Don’t 

Posted on by lechuck park

This image presents a practical guide on how to effectively integrate Artificial Intelligence, specifically Large Language Models (LLMs), into software systems. The overarching theme is “Rules for What We Know, AI for What We Don’t,” which emphasizes using reliable, traditional computing for hard facts and reserving AI for complex reasoning and interpretation.

1. Don’t Prompt What You Can Query

This principle warns against using AI to retrieve exact data. Because LLMs generate responses based on probabilities, they can sometimes guess incorrectly or hallucinate. If you need a verified fact—like a user’s bank balance—you should use a standard database search to fetch that exact number. Once you have the accurate data, you can then pass it to the AI to draft a natural, polite response.

2. Connect the Certain, Compute the Complex

This section suggests building a hybrid approach to problem-solving. You should establish a strict, rule-based foundation (the “certain”) using traditional logic, math, or physics. Once that solid framework is in place, you let the AI operate on top of it to handle creative or flexible tasks (the “complex”). For example, use traditional software to ensure a building is structurally safe, and then use AI to design creative interior layouts within those safe boundaries.

3. LLM is the Engine, Not the Database

This final point clarifies the true role of an LLM: it is a processor, not a storage drive. You shouldn’t try to force an AI to memorize massive amounts of raw data, like a 10,000-page company manual. Instead, use a search system to find the exact page you need, and then feed just that relevant text into the LLM. The AI acts as the “engine” to read, understand, and summarize that specific information for you.

Summary

To build reliable AI applications, rely on traditional databases and strict logic for factual retrieval and structural constraints. Use LLMs strictly as reasoning and processing engines to interpret context, draft text, and solve complex problems based on the hard facts you provide them.

#AIArchitecture #LLM #ArtificialIntelligence #SoftwareEngineering #DataScience #PromptEngineering #GenerativeAI

Always Energy

Posted on by lechuck park

This infographic contrasts the way human knowledge has been accumulated with how modern Artificial Intelligence (AI) operates, focusing on energy consumption and processing structure.

1. Left: The Trajectory of Human Intelligence (Ultra-low Power, Time, and Connection)

  • 20 Watt Icon: Represents the biological limit and astonishing efficiency of a single human brain, consuming only 20W—roughly the energy needed to power a dim lightbulb.
  • Network of Brains: Accompanied by the phrase “Through an immense network of human brains,” the interconnected 20W icons illustrate that while individual intelligence is limited by its biology, a massive web of knowledge was formed through collective intelligence and communication.
  • Timeline: The clock icon, the phrase “Over vast stretches of time,” and the long green arrow stretching to the right emphasize that this knowledge wasn’t built overnight. It was gradually and painstakingly accumulated over the long course of human history.

2. Center: The Transfer of Knowledge (Accumulation and Technology)

  • Inside the large yellow transition arrow, there are icons of books (accumulated knowledge) and a microchip (computing technology).
  • This symbolizes the bridge where humanity’s vast knowledge, built by 20W brains over countless generations, meets modern semiconductor technology and transitions into the realm of machines.

3. Right: The Era of AI (Ultra-high Power and Massive Parallel Processing)

  • 1000+ TWh Icon: Visualizes the astronomical power consumption (over 1000 Terawatt-hours) of global AI and data centers. Placed in stark contrast to the human “20W,” it highlights just how energy-intensive AI technology truly is.
  • Artificial Neural Network Structure: Along with the phrase “Massive Parallel Processing,” it shows a structure where numerous nodes process massive amounts of data simultaneously.
  • While humans processed and passed down information over a “long period,” this illustrates that AI reduces time and achieves unprecedented performance by pouring in “massive power” to compute everything simultaneously (in parallel).

💡 Overall Review

“Humanity built civilization with a mere 20W of energy through time and connection, whereas modern AI operates on massive parallel processing, consuming over 1000+ TWh of immense energy.”

#ArtificialIntelligence #HumanIntelligence #AIvsHuman #CollectiveIntelligence #NeuralNetworks

With Gemini

PI-DLinear(Physics-Informed DLinear)

Posted on by lechuck park

PI-DLinear (Physics-Informed DLinear)

The provided image is a structured infographic slide titled “PI-DLinear (Physics-Informed DLinear).” It visually organizes the model’s core features into four distinct, color-coded columns:

1. Physics-Informed Loss Function (Blue Column)

This section focuses on how physical laws are integrated into the model’s learning process.

  • #Hybrid Objective: It explains that the model integrates data fidelity with physical governing equations.
  • #Physical Constraints: It states that the model penalizes thermodynamically impossible predictions (e.g., violating energy conservation or heat transfer laws).
  • #Mathematical Formulation: It provides the core equation for the loss function: Ltotal = Ldata + Lphysic.

2. Harness Engineering & Safe Control (Purple Column)

This column emphasizes the safety and control aspects for AI operations.

  • #Operational Scaffolding: It describes the model as acting as a strict guardrail for autonomous AI-driven agents.
  • #Boundary Adherence: It guarantees that forecasts and control actions remain within safe, predefined physical boundaries, completely preventing critical hallucinations.

3. Robust OOD (Out-of-Distribution) Extrapolation (Green Column)

This section highlights the model’s reliability during unexpected scenarios.

  • #Anomaly Resilience: It notes that the model maintains highly rational trajectories during unprecedented emergencies (like sudden chiller failures) where pure data-driven models would collapse.
  • #Predictive Diagnostics: It points out that the model delivers accurate fault propagation forecasting, which directly enables a drastic reduction in MTTR (Mean Time To Repair).

4. Structural Simplicity & Computational Efficiency (Red Column)

The final column outlines the architectural benefits of the model.

  • #Linear Decomposition: It explains that the model splits time-series into trend and remainder components using highly interpretable linear layers, bypassing heavy attention mechanisms.
  • #High-Throughput Inference: It emphasizes that the model is exceptionally lightweight and fast, making it optimal for real-time DevOps, edge deployments, and multi-center scaling.

Summary

The infographic effectively presents PI-DLinear as a powerful hybrid model for time-series forecasting. By combining the computational speed and simplicity of linear architectures with the strict mathematical boundaries of physical laws, it creates a highly reliable AI tool. It is specifically designed to handle unexpected anomalies safely and efficiently, making it ideal for critical infrastructure management where AI hallucinations cannot be tolerated.

#PIDLinear #PhysicsInformedAI #TimeSeriesForecasting #AIOps #MachineLearning #SafeAI #PredictiveMaintenance #HarnessEngineering

With Gemini

Sag & Swell

Posted on by lechuck park

The image provides a clear, side-by-side comparison of two major power quality issues: Voltage Sag (or Dip) and Voltage Swell. It looks like a great summary graphic prepared for your tech blog at eeumee.net, particularly because it sharply highlights how these electrical phenomena specifically impact AI Data Centers (AI DC).

1. Voltage Sag / Dip

  • Definition: A sudden, momentary decrease in voltage.
  • System Impact: It causes immediate service and system disruption. If the voltage drops too low, servers can suddenly power off or reboot.
  • AI DC Relevance: Noted as “Very high on AI DC.” The risk and frequency are elevated in AI environments.
  • Root Cause: This is primarily driven by sudden load or workload changes. When thousands of GPUs simultaneously spin up for intensive AI training or inference tasks, they draw massive amounts of current in an instant, causing the voltage to dip.

2. Voltage Swell

  • Definition: A sudden, momentary increase in voltage.
  • System Impact: Unlike a sag, a swell might not cause an immediate outage, but it forces overvoltage through the components, leading to equipment stress and degradation.
  • AI DC Relevance: It carries a significant cumulative impact. The hardware damage builds up over time, eventually leading to premature component failure.
  • Root Cause: Typically triggered by power system or control abnormalities, or when a massive electrical load is suddenly dropped from the grid.

💡 Core Insight

This slide captures why power dynamics in AI Data Centers are vastly different from traditional IT environments. The extreme, dynamic power fluctuations inherent to AI workloads make rigorous power quality monitoring (via DCIM) and the implementation of highly responsive, advanced power architectures—such as Battery Energy Storage Systems (BESS)—absolutely critical to maintaining uptime and protecting expensive hardware.

#AIDataCenter #PowerQuality #VoltageSag #VoltageSwell #DataCenterInfrastructure #TechBlog #GPUWorkloads #ServerCooling

With Gemini