# durability (blogs) — RSS Amplifier Recent posts from the 5 feeds in the RSS Amplifier directory that cover durability. Page: Feed: --- ## [The Model Was the Same. The Serving Configuration Was Not.](https://sekoudoumbouya.com/blog/qwen38-one-model-four-configurations/) _2026-08-14 · Sekou M. Doumbouya - Technical Insights_ I ran official Qwen3.8 27B BF16 and FP8 side by side, then changed one serving knob at a time. The fastest result was not the whole story. ## [What Wins in a Benchmark Is Not What You Ship](https://sekoudoumbouya.com/blog/what-wins-in-a-benchmark-is-not-what-you-ship/) _2026-08-04 · Sekou M. Doumbouya - Technical Insights_ The best long-context profile I measured passed every synthetic gate and crashed twice under real traffic. Here is why promotion has to be a guarded transaction, not a config edit. ## [A Card That Made Me Sign a Letter](https://sekoudoumbouya.com/blog/deepseek-0731-and-my-ai-system-upgrade/) _2026-08-03 · Sekou M. Doumbouya - Technical Insights_ Why I added a second RTX PRO 6000, upgraded my CPU, and then watched DeepSeek 0731 land on the exact card I just bought. ## [The Task Is Not the Unit of Work](https://sekoudoumbouya.com/blog/the-task-is-not-the-unit-of-work/) _2026-07-11 · Sekou M. Doumbouya - Technical Insights_ A 41-task graph ran for 41 hours, produced 115 reviews, and merged nothing. Then I pivoted mid-flight to milestones, and the same session shipped five PRs in half the time. ## [Structurally Valid Is Not Good Enough: Why My Model Router Measures Quality](https://sekoudoumbouya.com/blog/anvil-serving-local-where-its-proven/) _2026-07-02 · Sekou M. Doumbouya - Technical Insights_ anvil-serving routes coding-agent work to local models only where measured quality has earned it, verifies local output, and escalates up the tier ladder the moment verification fails. ## [The Model Name Is Just a String](https://sekoudoumbouya.com/blog/the-model-name-is-just-a-string/) _2026-07-02 · Sekou M. Doumbouya - Technical Insights_ How wiring two mismatched GPUs into my coding harnesses turned into intent-based routing, and why the measurement, not the transport, became the product. ## [A Billion Tokens to Ship One Feature. Forty-Six Million Were Real.](https://sekoudoumbouya.com/blog/a-billion-tokens-forty-six-million-real/) _2026-06-27 · Sekou M. Doumbouya - Technical Insights_ Two days of autonomous agents shipped a release and moved a billion tokens. Only about five percent was generated work. That five percent is the number that sizes a local box. ## [I Bet on Owning My Own Substrate](https://sekoudoumbouya.com/blog/owning-my-own-substrate/) _2026-06-27 · Sekou M. Doumbouya - Technical Insights_ Why I built a dual-GPU workstation and stood up local model serving, and why owning the durable layer is the bet I am most glad I made. ## [I Said 'Report in the Morning' and Went to Sleep. It Shipped a Release.](https://sekoudoumbouya.com/blog/report-in-the-morning/) _2026-06-24 · Sekou M. Doumbouya - Technical Insights_ I handed an agent a goal and one sentence, then slept. 23.7 hours of unattended runtime later, Anvil v0.3.0 was on PyPI. Here is what it woke me for, and what it shouldn't have. ## [Status Is Downstream of Proof: The Architecture of Anvil](https://sekoudoumbouya.com/blog/status-is-downstream-of-proof/) _2026-06-24 · Sekou M. Doumbouya - Technical Insights_ An agent reported a test it never wrote. That gap is the whole reason Anvil exists, now open source as anvil-state on PyPI. Here is the architecture, and the one part still unsolved. ## [Dissecting Series A Rounds (Sponsored)](https://crawlproof.com/a/dlAa8LhlM3Wu) _2026-06-23 · **Sponsored**_ Two real Series A cases: timelines, investor sequencing, and common founder mistakes. ## [I Rebuilt My Crew on a New Runtime. Only the Ledger Survived.](https://sekoudoumbouya.com/blog/only-the-ledger-survived/) _2026-06-21 · Sekou M. Doumbouya - Technical Insights_ I ported a nine-agent crew from one harness to another. The orchestration had to be rebuilt. The ledger moved untouched. That gap is the whole product. ## [I Didn't Need a Bigger Computer. I Needed Room to Think.](https://sekoudoumbouya.com/blog/usable-context/) _2026-06-21 · Sekou M. Doumbouya - Technical Insights_ A single 5090 could run three local agents. It just couldn't give them room to work. The story of the VRAM wall, the three ways out I priced, and the upgrade I actually made. ## [Volume Goes Local. Stakes Go to the Cloud.](https://sekoudoumbouya.com/blog/volume-local-stakes-cloud/) _2026-06-21 · Sekou M. Doumbouya - Technical Insights_ The question that routes an agent task is not which model is best. It is what the task can afford to get wrong. How I pinned a nine-agent crew across a cloud frontier tier and a local GPU. ## [I Bought a GPU to Run My Own Agents. Then I Red-Teamed the Plan.](https://sekoudoumbouya.com/blog/i-bought-a-gpu-to-run-my-own-agents/) _2026-06-20 · Sekou M. Doumbouya - Technical Insights_ Maxing my AI plans and fighting rate limits, I split the spend across providers, bought a local GPU for overflow, and designed a pull queue for my agents. Then I red-teamed the plan. ## [Everyone Grew the Same Eye: A Field Guide to the Spec-Driven Wave](https://sekoudoumbouya.com/blog/field-guide-to-the-spec-driven-wave/) _2026-06-16 · Sekou M. Doumbouya - Technical Insights_ Spec Kit, Kiro, Tessl, task-master, BMAD, Agent OS, OpenSpec: seven projects, twelve months, one idea. A field guide to who's building what, and why they converged. Part 2. ## [The Newest Idea in AI Coding Is Forty Years Old](https://sekoudoumbouya.com/blog/spec-driven-development-older-than-the-hype/) _2026-06-16 · Sekou M. Doumbouya - Technical Insights_ Spec-driven development feels like a 2025 invention. It isn't: it's literate programming, MDA, TLA+, BDD, and Terraform's desired state in a hoodie. Part 1 of a series. ## [Where the Wave Breaks: The Layer Nobody Wants to Build](https://sekoudoumbouya.com/blog/where-the-spec-driven-wave-breaks/) _2026-06-16 · Sekou M. Doumbouya - Technical Insights_ Every spec-driven tool stops at the same place: a great document handed to an agent, and trust. The layer they skip (durable state, leases, evidence) is the hard part. Part 3. ## [The Anvil, Not the Hammers: Where fakoli-state Goes Next](https://sekoudoumbouya.com/blog/fakoli-state-the-anvil/) _2026-06-10 · Sekou M. Doumbouya - Technical Insights_ I had an agent system audit my agent system against the field. The verdict: the orchestration is being commoditized. The moat is the boring, durable state layer. ## [Reading Other People's Bug Trackers: The Bets Behind fakoli Are Getting Called](https://sekoudoumbouya.com/blog/reading-other-peoples-bug-trackers/) _2026-06-10 · Sekou M. Doumbouya - Technical Insights_ I mined four hundred GitHub issues from the tools next to mine. The field's pain validated the bets — and Anthropic shipping agent teams natively confirmed the biggest one. ## [The Fakoli Style: An Operating Model for Building With Agents](https://sekoudoumbouya.com/blog/the-fakoli-style/) _2026-05-01 · Sekou M. Doumbouya - Technical Insights_ Three plugins, four invariants, one principle. What I built isn't a toolkit — it's an operating model for serious development with AI coding agents. ## [Follow Releases, Get Reminders (Sponsored)](https://crawlproof.com/a/JBtyD4wIJgj8) _2026-04-30 · **Sponsored**_ Follow a genre or a name and receive reminders on release day — free. ## [Plan, Claim, Apply: Building fakoli-state](https://sekoudoumbouya.com/blog/building-fakoli-state/) _2026-04-24 · Sekou M. Doumbouya - Technical Insights_ A local-first project-state layer in SQLite, with claims that have heartbeats and evidence captured by hooks. Here's how it was built and which trade-offs I'm still living with. ## [Crash-only confidence, part 1: from random kills to exact boundaries](https://aalhour.com/posts/beachdb-crash-testing-part1/) _2026-04-23 · Ahmad Alhour · aalhour_ A crash test that cannot explain itself is mostly theater. ## [State Is the Product: What Agentic Workflows Were Missing](https://sekoudoumbouya.com/blog/state-is-the-product/) _2026-04-19 · Sekou M. Doumbouya - Technical Insights_ Workflow orchestration alone won't save you. The thing that breaks first when you run real multi-agent work is the state — and almost nobody is building that layer. ## [Making the on-disk state real: Sorted-String Tables (SSTables)](https://aalhour.com/posts/beachdb-sstables-v1/) _2026-04-05 · Ahmad Alhour · aalhour_ Files are easy; files you can trust, inspect, and recover from are the real step. ## [Building the Engine: Architecture of a Durable Agentic Execution System](https://sekoudoumbouya.com/blog/baara-building-the-engine/) _2026-04-05 · Sekou M. Doumbouya - Technical Insights_ A 10-package monorepo, 27 MCP tools, checkpoint-based durability, and three sandbox types. The full blueprint for Baara Next. ## [Intent Should Be Durable: The Missing Layer Between Agent SDKs and Production Systems](https://sekoudoumbouya.com/blog/baara-intent-should-be-durable/) _2026-04-05 · Sekou M. Doumbouya - Technical Insights_ Agent SDKs give you tools and reasoning. They don't give you queues, checkpoints, or crash recovery. That gap is where your intent dies. ## [From Pressing Buttons to Intent-Driven Flow](https://sekoudoumbouya.com/blog/intent-driven-agentic-flow/) _2026-04-04 · Sekou M. Doumbouya - Technical Insights_ How I went from managing AI agent context windows by hand to building an orchestration system that lets me describe what I want and walk away. ## [Teaching AI Agents to Work Like a Team (Not a Crowd)](https://sekoudoumbouya.com/blog/fakoli-crew-agent-archetypes/) _2026-03-25 · Sekou M. Doumbouya - Technical Insights_ What I learned building specialized agent archetypes that coordinate through file ownership, wave execution, and structured handoffs. ## [From Cloud Infrastructure to AI Infrastructure: What Transfers and What Doesn't](https://sekoudoumbouya.com/blog/infrastructure-engineer-to-ai-architect/) _2026-03-10 · Sekou M. Doumbouya - Technical Insights_ After twenty years building cloud platforms, here's what I've learned about bridging into AI infrastructure, and which skills matter more than you'd expect. ## [Why I Rebuilt My AI Assistant From Scratch (And What the Analysis Told Me)](https://sekoudoumbouya.com/blog/building-a-personal-ai-gateway/) _2026-03-05 · Sekou M. Doumbouya - Technical Insights_ How a structured competitive analysis of my own code convinced me that a clean-room rewrite was the right call, and what I learned building Nexus. ## [Privacy-first AI-native browser (Sponsored)](https://crawlproof.com/a/62ZwzmZZOrDw) _2026-03-04 · **Sponsored**_ No telemetry or ads — built on Ungoogled Chromium with a built-in AI sidebar and CLI. ## [The Memtable: where writes go to wait](https://aalhour.com/posts/beachdb-memtable-v1/) _2026-02-21 · Ahmad Alhour · aalhour_ Before the disk gets involved, memory has to learn the rules. ## [The syscall I forgot: directory fsync](https://aalhour.com/posts/beachdb-the-syscall-i-forgot/) _2026-02-15 · Ahmad Alhour · aalhour_ Sometimes the missing durability bug is one directory above where you were looking. ## [Documentation Is Infrastructure (Just Made of Text Instead of Routers)](https://sekoudoumbouya.com/blog/documentation-as-infrastructure/) _2026-02-15 · Sekou M. Doumbouya - Technical Insights_ Why the blueprint document I wrote for our AWS organization might be the highest-leverage thing I shipped that year. ## [Durability is a promise you can't hand-wave: Write-Ahead Log (WAL)](https://aalhour.com/posts/beachdb-wal-v1-milestone/) _2026-02-11 · Ahmad Alhour · aalhour_ Once bytes leave your process, the comforting story gets complicated. ## [The Multiplier Effect: Why I Keep Mentoring as a Core Output](https://sekoudoumbouya.com/blog/multiplier-effect-mentorship/) _2026-01-20 · Sekou M. Doumbouya - Technical Insights_ The only sustainable scaling strategy is multiplying competent decision-makers. Here's how I think about mentorship as a senior IC. ## [Calculator Unboxing 10](http://www.ssaarcasfpaa.com/unboxing/calculator-unboxing-10) _2025-11-19 · Brady Haran · South Surrey And Associated Regions Calculator Appreciation Society for Professionals and Amateurs_ Bitoon BT-7839 BITOON BT-7839 Display: 5 Functionality: 4 Accuracy: 3 Ergonomics: 7 Cred: 9 Durability: 5 Total: 33 ## [Platform Engineering Is Product Management for Infrastructure](https://sekoudoumbouya.com/blog/platform-engineering-philosophy/) _2025-11-15 · Sekou M. Doumbouya - Technical Insights_ Why treating your internal platform like a product changes how infrastructure teams operate, and how golden-path solutions reduce cognitive load. ## [Questioning Architectural Defaults: A Transit Gateway Cost Story](https://sekoudoumbouya.com/blog/transit-gateway-cost-lesson/) _2025-09-20 · Sekou M. Doumbouya - Technical Insights_ Why periodically questioning inherited infrastructure decisions matters, illustrated through an AWS Transit Gateway to VPC peering migration. ## [The Player-Coach Model: Why I Stay 40% Hands-On](https://sekoudoumbouya.com/blog/player-coach-leadership/) _2025-07-10 · Sekou M. Doumbouya - Technical Insights_ Why staying hands-on at the principal engineer level creates better outcomes for your team, your architecture, and your credibility. ## [Start using Object.hasOwn()](https://yuval.hashnode.dev/start-using-objecthasown) _2025-05-07 · Yuval Aloni · devdreams_ Above - screenshot from MDN’s Object.hasOwn() documentation. If you so far waited to make the switch because of “lack of wide support”:Since April 2025 - even the Android native browser and Android Firefox support it by now (and there has been wide... ## [Designing Multi-Region Infrastructure: Patterns That Scale](https://sekoudoumbouya.com/blog/multi-region-architecture-patterns/) _2025-05-01 · Sekou M. Doumbouya - Technical Insights_ A taxonomy of region types, spec-driven provisioning, and phased orchestration for predictable multi-region expansion. ## [Making DOM Text Nodes Reactive](https://yuval.hashnode.dev/making-dom-text-nodes-reactive) _2024-10-05 · Yuval Aloni · devdreams_ Introduction One of the main aspects of popular client-side frameworks is reactivity, which means a “binding” between two variables: whenever one value changes, the other is automatically adjusted. This binding is usually one-way in the context of th...