Sami Belhadj @sbelhadj - Tumblr Blog

GPT-5.6: OpenAI Ships a Family, Not a Model

On July 9, OpenAI moved the GPT-5.6 family to general availability, two weeks after a limited preview. There is no single “GPT-5.6.” There are three: Sol, Terra, and Luna. If you write code for a living, this release is worth ten minutes of your attention, and a bit of skepticism. Why it matters The headline is not a benchmark. It is economics. OpenAI’s pitch is that GPT-5.6 gets more useful…

Where to invest: a maturity model for your team

A 5-part series — Part 5/5 The most common mistake isn’t picking the wrong discipline. It’s investing in a layer above where your team actually sits. People want to build a sophisticated harness when they don’t even have a test bench for their prompts. They stack autonomous agents on foundations that are taking on water. After four parts of definitions, here’s the part that’s useful at work:…

Harness engineering: the model is the smallest part

A 5-part series — Part 4/5 – Harness engineering When the Claude Code source leaked in late March 2026, people could finally measure something practitioners had suspected without proof: across roughly 512,000 lines of TypeScript, the model interaction itself is a tiny fraction. Everything else, the vast majority, is the harness. That sums up the outermost layer of our diagram. The model sits at…

Context engineering: the window is a scarce resource

A 5-part series — Part 3/5 Here’s the most counterintuitive idea in the whole series: giving a model more information can make it worse. It’s tempting to picture the context window as a big bag where you cram everything that might help. In practice it’s a budget. Every token you add spends another, and past a certain point the model degrades. Context engineering is the discipline that manages…

Prompt engineering: what still holds up in 2026

A 5-part series — Part 2/5 We buried prompt engineering a little too fast. The discipline isn’t dead, it’s been demoted. It now sits at the innermost layer of the diagram from Part 1: how you phrase a request, inside a single call. That’s small. It’s also the first place everything can go wrong. In this part I separate what actually works from what’s folklore. Because between 2022 and now, we…

Prompt, context, harness: the three ages of LLM engineering

A 5-part series — Part 1/5 Three years ago, knowing how to “talk” to a model was a rare and slightly magical skill. People traded screenshots of prompts that worked and passed around phrasings like cooking recipes. Andrej Karpathy summed up the era in a line that stuck: the hottest new programming language was English. Today nobody serious introduces themselves as a prompt engineer. The word…

Prompt, Context, Harness : les trois âges de l'ingénierie des LLM

Série en 5 épisodes — Épisode 1/5 Il y a trois ans, savoir « parler » à un modèle était une compétence rare et un peu magique. On s’échangeait des captures d’écran de prompts qui marchaient, on se refilait des tournures comme des recettes de cuisine. Andrej Karpathy a résumé l’époque d’une formule devenue culte : le langage de programmation le plus en vogue, c’était l’anglais. Aujourd’hui, plus…

Claude Fable 5 : Le premier modèle « Au-dessus d'Opus » que vous pouvez réellement utiliser aujourd'hui

Anthropic a lancé Claude Fable 5 hier, et je veux être prudent sur la façon dont j’en parle, car la ferveur marketing entourant les lancements de modèles est désormais fixée en permanence sur « cela change tout ». La plupart du temps, ce n’est pas le cas. Celui-ci a plus de chances d’avoir une réelle importance, et pas entièrement pour les raisons mises en avant dans l’article de…

Claude Fable 5: The First "Above-Opus" Model You Can Actually Use Today

Anthropic shipped Claude Fable 5 yesterday, and I want to be careful about how I talk about it, because the marketing temperature on model launches is now permanently set to “this changes everything.” Most of the time it doesn’t. This one is closer to actually mattering, and not entirely for the reasons the launch post leads with. Here is the short version for busy people: Fable 5 is the first…

Beyond the Benchmark: 5 Surprising Truths About Anthropic’s New Claude Opus 4.8

1. Introduction: The Post-Launch Reality Check In the hyper-accelerated cadence of the AI industry, “new” is beginning to feel synonymous with “incremental.” With Claude Opus 4.7 having arrived only a month prior, many developers initially greeted the May 28, 2026, release of Claude Opus 4.8 with a measure of release fatigue. At first glance, another flagship update so soon feels like a…

Building a Second Brain with Claude and Obsidian

A complete step-by-step tutorial on Andrej Karpathy’s LLM Wiki method Karpathy’s LLM Wiki gist on GitHub has been starred over 5,000 times and forked just as many. Most of those people will never actually build the thing. This tutorial exists for the people who will. Original reference: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f Why this isn’t another RAG tutorial When…

Architectural Implementation Blueprint: Deploying the Multi-Layer Claude Code Framework

1. The Paradigm Shift: From Chatbot to Senior Teammate Deprecate the “chatbot” mental model in favor of the stateful runtime primitive. In high-performance engineering environments, viewing Claude Code as a conversational assistant is a Level 1 bottleneck. To achieve Level 4 architectural maturity, you must treat Claude Code as a comprehensive agent development kit (ADK) capable of executing…

I Tested the New GPT-5.5 Model for You (for 2 Hours) and This is My Take

The “Spud” Has Landed: Beyond the Poem The AI “jagged frontier” just moved again, and this time it didn’t come with a flashy keynote. OpenAI surprise-dropped GPT-5.5 (codenamed “Spud”), and the technical community is scrambling to figure out if this is a true step-change or just another incremental polish. I’ve spent the last two hours ignoring the hype to focus on the signal. We are officially…

How to Never Hit Claude's Rate Limits

It’s 11 AM and Claude just told you to come back later. You’re mid-draft on a client deliverable, or halfway through debugging a function that was almost working. Sound familiar? Here’s the thing: the limit isn’t the problem. Your habits are. Most people default to the biggest model for everything, leave token-hungry features running in the background, and treat every chat like a never-ending…

Software 3.0 : The Evolution of Software in the Era of AI

From Code to Weights to Prompts: A Practitioner’s Guide to Software 3.0 By Sami BELHADJ Introduction: The Machine Is Finally Listening In the past three years, software development has hit an inflection point unlike anything the industry has seen since the transition from mainframes to personal computing. Large language models have exploded in popularity at a pace that caught even their…

Conséquences économiques de la Circulaire BCT n°2026-4 sur l'économie tunisienne

Analyse des impacts à court, moyen et long terme du durcissement des conditions de financement des importations non prioritaires Analyse consolidée — 4 avril 2026 sur https://tech-tech.life Par Sami Belhadj (https://www.linkedin.com/in/sbelhadj) Résumé exécutif La circulaire aux intermédiaires agréés n°2026-4 du 26 mars 2026 de la Banque Centrale de Tunisie (BCT) impose un changement radical…

Specification-Driven Development (SDD) : Why the Best Code You’ll Ever Write Isn't Written in Code

1. The Perpetual Gap: The Documentation Tax We have long suffered the “Documentation Tax”—that inevitable decay where the Product Requirements Document (PRD) becomes a tombstone for original intent the moment the first line of code is written. In the traditional cycle, we treat specifications as “discarded scaffolding”: temporary structures used to erect the “real” building (the code) and then…

Trending Blogs

Recently Viewed Blogs

Sami Belhadj