Ben Felix Framework vs Durable Sessions: Which to Use?
// TL;DR
These two frameworks solve completely unrelated problems—one is personal finance, the other is AI product engineering—so there is no crossover use case. If you are making a financial decision (investing, rent vs. buy, goal-setting), use the Ben Felix Evidence-Based Financial Decisions Framework. If you are building or fixing the streaming architecture of an AI chat product, use the Christensen Durable Sessions AI UX Framework. Pick based on the problem in front of you; they never compete.
// HOW DO THEY COMPARE?
| Dimension | Ben Felix Evidence-Based Financial Decisions Framework | Christensen Durable Sessions AI UX Framework |
|---|---|---|
| Best For | Anyone making personal financial decisions (investing, housing, saving, insurance, goal-setting) | Engineers and product designers building or auditing AI chat/agent streaming architectures |
| Domain | Personal finance and behavioural economics | AI product engineering, real-time streaming infrastructure |
| Complexity | Low to moderate — designed for non-experts; no technical prerequisites | High — requires understanding of SSE, WebSockets, pub/sub, and agent architectures |
| Time to Apply | 1–3 hours for a full financial audit; 15 minutes for a single decision like rent vs. buy | Days to weeks for a full architecture redesign; 1–2 hours for an initial audit |
| Prerequisites | Basic knowledge of your own finances (income, savings, debts, goals) | Working knowledge of streaming protocols, backend architecture, and AI agent systems |
| Output Type | Prioritised personal action list (e.g., max tax-advantaged accounts, set up index funds, write a will) | Architecture redesign plan with specific failure-mode fixes and a session-layer blueprint |
| Creator Background | Ben Felix — portfolio manager, CFA, host of the Rational Reminder podcast, known for academic-research-backed finance education | Mike Christensen — engineering leader at Ably, specialist in real-time infrastructure and AI UX delivery |
| Core Principle | Investing is solved (low-cost index funds); the hard part is behaviour and avoiding common mistakes | AI product quality depends on the delivery layer, not the model; decouple agents from clients via Durable Sessions |
| Number of Workflow Steps | 10 steps covering diagnosis, goal-setting, housing, investing, tax, insurance, spending profile, human capital, and action plan | 10 steps covering architecture audit, session layer design, transport swap, multi-agent flattening, and validation |
| Who Should NOT Use It | Someone looking for speculative stock picks, crypto trading strategies, or software architecture guidance | Someone looking for personal financial advice, investing principles, or goal-setting frameworks |
What does the Ben Felix Evidence-Based Financial Decisions Framework do?
The Ben Felix framework is a structured, 10-step system for auditing and improving any major personal financial decision using academic research. It covers investing, rent vs. buy analysis, goal-setting, tax planning, insurance, estate planning, and spending psychology.
Its core thesis is radical in its simplicity: the investment question is solved. Low-cost, broadly diversified index funds beat the vast majority of alternatives. The real challenge is behavioural — avoiding the psychological traps (checking your portfolio too often, picking individual stocks, ignoring opportunity costs) that erode the returns a disciplined investor would earn.
The framework walks users through diagnosing which of the "Top 10 Financial Mistakes" apply to them, setting high-quality goals using the PERMA model from positive psychology, applying the 5% Rule to rent-vs-own decisions, auditing tax-advantaged account usage, and building a prioritised action plan. It is designed for anyone — no finance degree required.
What does the Christensen Durable Sessions AI UX Framework do?
The Christensen framework is a technical architecture guide for engineers building AI-powered chat or agent products. It diagnoses why AI product experiences break under real-world conditions — network drops, multi-device usage, concurrent agent activity — and prescribes a specific architectural pattern called Durable Sessions to fix them.
The core insight is that most AI products use direct HTTP streaming (typically SSE), which couples the health of the AI response stream to the health of a single client connection. When that connection drops, the stream dies. This is the "Single-Connection Trap," and it prevents three capabilities that separate a fragile demo from a production-quality AI product: Resilient Delivery (surviving disconnections), Continuity Across Surfaces (session follows the user across devices), and Live Control (steering or stopping an agent mid-generation).
The solution is a Durable Sessions layer — a persistent, shared channel between agents and clients where agents publish events and clients subscribe. Neither side holds a direct pipe to the other. This unlocks all three capabilities simultaneously and dramatically simplifies agent code.
How do they compare?
These two frameworks occupy entirely different domains and solve entirely different problems. There is zero overlap in their use cases, audiences, or outputs.
The Ben Felix framework is a personal decision-making tool for anyone with money questions. It requires no technical skills. Its output is a prioritised action list for your financial life — things like "max your TFSA," "switch to index funds," or "write a will."
The Christensen framework is a software architecture tool for engineers building AI products. It requires deep familiarity with streaming protocols, pub/sub systems, and agent topologies. Its output is an architecture redesign plan — things like "introduce a Durable Sessions layer," "replace SSE with WebSockets," or "flatten your orchestrator relay."
The only meaningful similarity is structural: both are 10-step diagnostic workflows that start with an audit, identify failure modes, and produce a prioritised fix list. Both are opinionated — they tell you what to do, not just what to consider. And both argue that the hard problem is not where most people think it is: Ben Felix says the hard part of investing is behaviour, not stock selection; Christensen says the hard part of AI products is infrastructure, not model quality.
Which should you choose?
Choose the Ben Felix framework if you are making any significant personal financial decision or want to audit your overall financial health. It is the right tool if you are asking questions like: Should I buy or rent? How should I invest my savings? Am I making financial mistakes? What should my goals even be? It works for beginners and experienced investors alike.
Choose the Christensen Durable Sessions framework if you are an engineer or product designer building an AI chat or agent product and your streaming experience breaks on disconnect, does not work across devices, or does not support a stop button or mid-generation steering. It is the right tool if your AI product works in demos but fails in production.
If you are a software engineer who also wants to get their personal finances in order, use both — they address completely separate areas of your life. There is no scenario where you would choose one instead of the other.
// FREQUENTLY ASKED QUESTIONS
Can I use the Ben Felix framework and the Durable Sessions framework together?
Yes, but they solve completely different problems. The Ben Felix framework is for personal financial decisions. The Durable Sessions framework is for AI product architecture. An engineer could use both — one for their finances, one for their product — but they never overlap or substitute for each other.
Which framework is better for beginners?
The Ben Felix framework is explicitly designed for non-experts and requires no technical background — just basic knowledge of your own financial situation. The Durable Sessions framework requires significant software engineering knowledge (streaming protocols, WebSockets, pub/sub systems) and is not accessible to beginners outside of engineering.
Does the Ben Felix framework tell me what stocks to buy?
No. Its core principle is that stock picking is a mistake. The framework recommends low-cost, broadly diversified index funds as the default for virtually all investors. The focus is on avoiding behavioural errors, using tax-advantaged accounts, and setting evidence-based goals — not selecting individual investments.
Does the Durable Sessions framework work with any AI model or just specific ones?
It is model-agnostic. The framework addresses the delivery and connectivity layer between AI agents and clients, not the AI model itself. It works with any LLM or agent system — OpenAI, Anthropic, open-source models, or custom agents — because the session layer sits between the agent and the client.
How long does it take to apply the Ben Felix 5% Rule for rent vs. buy?
About five minutes. Multiply the home purchase price by 5%, divide by 12, and compare the result to your monthly rent for a comparable property. If rent is lower, renting is the better financial decision. The framework also provides guidance on when to adjust the 5% figure for taxes or maintenance.
What is the biggest mistake the Durable Sessions framework prevents?
The Single-Connection Trap — where your AI product's response stream dies whenever a client's network connection drops. This is the root cause of most AI chat UX failures in production. The framework eliminates it by decoupling agent output from client connections via a persistent session layer.
Is the Ben Felix framework only for Canadian investors?
No. While Ben Felix is Canadian and references Canadian accounts like TFSAs and RRSPs, the framework's principles — index fund investing, the 5% Rule, PERMA goal-setting, opportunity cost thinking — are universal. It includes prompts to identify jurisdiction-specific tax-advantaged accounts for any country.
Do I need to switch from SSE to WebSockets if I use the Durable Sessions framework?
Only if you need Live Control — the ability for users to send steering messages or cancel signals while the agent is generating. SSE is one-way, so it cannot distinguish between a network disconnect and a user pressing stop. For Resilient Delivery and Cross-Surface Continuity alone, a Durable Sessions layer on top of SSE can work.