Artificial Intelligence
How to install and tune oMLX on M5 Max 128 GB
Tested May 2026 recipe: oMLX 0.3.8 on Mac M5 Max with 128 GB, TurboQuant at 3.5-bit, Qwen 3.6 35B-A3B model stack, Claude Code wiring and real benchmarks.
Tag
llm
Architecture
Hybrid RAG in 2026: the patterns that keep winning
Hybrid RAG in 2026 combines dense and lexical search fused with RRF, cross-encoder reranking over the top-50 candidates, structure-aware chunking, and continuous evaluation with Ragas or TruLens. It is the pattern that survives in serious production systems three years after the initial embeddings boom.
Artificial Intelligence
FinOps on agent tokens: the invoice that surprises
The first invoice for a production agent usually runs double or triple the estimate. This article walks through five real levers, in priority order, caching, routing, context control, batching, and telemetry, to cut cost without touching perceived quality.
Artificial Intelligence
Runtime-generated UI: the first serious year
The idea of UI generated on the fly instead of pre-built reached production in 2025. After a year of real-world use, the balance is more nuanced than the initial enthusiasm suggested.
Artificial Intelligence
DPO and alternatives to RLHF: practical state in 2026
Direct Preference Optimization (DPO) and its variants, IPO, KTO, and SimPO, have displaced RLHF as the preferred alignment method for language models: they drop the separate reward model, cut training cost, and are easier to reproduce. RLHF still has an edge only for frontier models with very large budgets.
Artificial Intelligence
Synthetic training data in 2026: when it works
Synthetic data has moved from a precarious substitute for real data to a central component of modern model training: the most reliable pattern expands a real core of 500 examples with thousands of synthetic paraphrases, provided you validate diversity, correctness, and distribution, and keep at least 30% real data to avoid model collapse.
Artificial Intelligence
LLM red teaming: a practical playbook
LLM red teaming has gone from an esoteric activity to a mandatory practice. With the OWASP Agentic Top 10 and the CSA Agentic AI Red Teaming Guide converging on shared vocabulary, this is the operational playbook any team deploying agents needs to have.
Artificial Intelligence
Prompt Engineering: From Trick to Mature Discipline
Prompt engineering has moved from viral tricks to a discipline with reproducible patterns: few-shot, chain-of-thought, and structured output with function calling. Teams treating prompts like code (versioned, tested, and monitored) get consistently better results than those who improvise.
Artificial Intelligence
FinOps for AI workloads in 2026: the real pain
La factura de IA en las empresas ha dejado de ser anecdótica. Entre tokens de modelos frontera, GPUs reservadas que nadie usa y pipelines RAG con cachés mal configuradas, muchos equipos pagan diez veces lo que deberían. Guía de FinOps específico para IA sin relatos promocionales.
Artificial Intelligence
Knowledge graph renaissance with LLMs
Knowledge graphs spent two decades waiting for their moment. With LLMs now bridging free text and formal ontology, and the GraphRAG pattern already mature, the technology is back in the spotlight. Time to look at why it finally fits and where it actually pays off.
Artificial Intelligence
LLM-generated documentation: when it helps and when it gets in the way
Large language models have spent two years promising effortless documentation for code, APIs and architecture. After watching dozens of projects try it, clear patterns emerge for where it works and where it just becomes more debt.
Artificial Intelligence
LLM guardrails: frameworks and their real cost
Guardrails frameworks promise to filter language-model inputs and outputs to block data leaks, harmful content, or hallucinations. After evaluating four of the most popular ones in production, I cover what they actually do, what latency and billing cost they add, and when they pay off over simpler controls.
Artificial Intelligence
AI agent observability: what to instrument first
Agents that chain calls to models, tools and memory are hard to debug without instrumentation designed for them. After a long year running agents in production, I cover what to measure first, which standards are consolidating, and which costly mistakes are avoided by getting the traces right from the start.
Architecture
LLM caches: saving tokens without dropping quality
A caching proxy in front of a language model can cut the token bill significantly, but it introduces subtle risks if the design is not careful. Which cache types work in production, where the usual traps sit, and how to add them without degrading the experience.
Architecture
Inference routers: choosing a model based on the request
Un enrutador de inferencia decide qué modelo atiende cada petición en función de coste, latencia y complejidad. Bien diseñados reducen la factura de tokens sin que el usuario perciba degradación; mal diseñados introducen fallos sutiles difíciles de depurar.
Artificial Intelligence
Testing with AI: the determinism problem
Probar sistemas que incluyen modelos de lenguaje rompe la primera regla del testing: la misma entrada da la misma salida. Analizo las estrategias que han funcionado tras un año largo integrando IA en productos reales, por qué los tests deterministas tradicionales no bastan y cómo plantear un cinturón de pruebas que capture regresiones sin bloquearse en la varianza.
Architecture
Agent OS: the concept shaping the new stack layer
The term Agent OS has spent a year gaining traction across research and product circles. It describes a layer that goes well beyond an agent library: request scheduling, context management, persistent memory, and isolation. A look at the real state of that concept.
Architecture
Model Context Protocol in 2025: from announcement to ecosystem
Model Context Protocol turns ten months old since Anthropic's announcement, and it is no longer just a proposal: hundreds of servers, cross-vendor implementations and a public registry now back it. A look at what has worked, what is still weak, and why 2025 marks the shift from curiosity to basic infrastructure.
Artificial Intelligence
GPT-5: public availability and early impressions
After months of rumors, OpenAI released GPT-5 in early August. The first weeks of real-world use show a picture less spectacular than the marketing suggested and more useful than many expected. It is worth separating what is genuinely new from what is merely incremental.
Artificial Intelligence
SLMs at the industrial edge: when the small model is better
Small language models have become genuinely useful. Phi-3.5, Gemma 2, and Llama 3.2 fit on modest hardware and solve bounded tasks without reaching the cloud. A look at where they fit on the factory floor and when skipping the large model pays off.
Artificial Intelligence
RAG 2.0: knowledge graphs, vectors, and hybrid
El RAG de 2023 era búsqueda vectorial con un LLM detrás. El de 2025 es un sistema híbrido que combina vectores, búsqueda léxica y grafos de conocimiento. Qué ha cambiado, dónde funciona cada pieza y qué decisiones marcan la diferencia entre un RAG útil y uno decepcionante.
Artificial Intelligence
Gemini 2.5: context scaling and multimodality
Google publicó Gemini 2.5 Pro en vista previa en marzo y la versión general llegó en junio. El salto respecto a Gemini 2.0 no está solo en puntuaciones sino en dos frentes prácticos: ventana de contexto utilizable en serio y multimodalidad que deja de ser demostración para convertirse en herramienta.
Artificial Intelligence
The initial Claude 4 family: first quality tests
Anthropic presentó Claude Opus 4 y Claude Sonnet 4 el 22 de mayo de 2025, el primer salto grande de nomenclatura desde la serie 3.5. Un mes de uso real en código, documentación técnica y agentes para separar lo que ha mejorado de lo que sigue igual.
Artificial Intelligence
The knowledge graph era is reborn with LLMs
For a decade, knowledge graphs were an academic idea with few real use cases, held back by the cost of building and maintaining the schema. LLMs have changed that equation: they now extract entities automatically and help anchor answers, audit reasoning, and support agents without hallucinating.
Artificial Intelligence
Continuous evaluation of RAG: dashboards that actually matter
Un sistema RAG sin evaluación continua se degrada en silencio. Los índices cambian, los modelos se actualizan, los usuarios preguntan cosas nuevas. Este es un repaso práctico de qué métricas vigilar y cómo montar el cuadro de mando que avisa antes del incidente.
Artificial Intelligence
AI agents in the enterprise: from demo to measurable value
AI agents have moved from a lab curiosity to serious SDKs from three major providers. A reflection on moving from the flashy demo to an internal use case that shifts a real, measurable metric.
Architecture
Applying graph RAG to a real product
Desde que Microsoft abrió GraphRAG, el patrón de usar grafos sobre tus propios datos ha pasado de experimento académico a técnica con aplicaciones prácticas. Reflexión sobre cuándo compensa, cómo se monta y qué errores se repiten.
Artificial Intelligence
Claude 3.7 Sonnet: the intermediate step toward the 4 family
Anthropic publicó Claude 3.7 Sonnet a finales de febrero con pensamiento extendido opcional y un compañero de consola llamado Claude Code. Reflexión sobre qué cambia de verdad y qué queda para la próxima familia.
Artificial Intelligence
vLLM in 2025: the improvements that matter to LLM-serving teams
vLLM remains the reference engine for serving LLMs on GPU in 2025: automatic prefix caching sharply cuts latency for repeated prompts, speculative decoding speeds up large models, and multi-LoRA support lowers the cost of multi-tenant SaaS, though multi-GPU support and non-NVIDIA hardware remain weak points.
Architecture
Microsoft’s GraphRAG in enterprise: patterns that work
GraphRAG has been in real enterprise use for over a year: during indexing, an LLM builds a knowledge graph that answers global questions about a corpus well, precisely where classic RAG fails because no single chunk holds the full answer. Here I compare indexing costs, the cases where it pays off, and the hybrid pattern that teams have settled on.
Artificial Intelligence
Alignment evaluation: RLHF, DPO, and recent alternatives
Three years after RLHF became popular, the model-alignment field is far richer. A review of RLHF, DPO, and newer methods such as KTO and ORPO, with criteria for choosing between them.
Artificial Intelligence
Gemma 2: Google’s open model one year later
Google released Gemma 2 in mid-2024, and it has since seen real production use. A look at how it competes in the open-model ecosystem, which sizes actually make sense, and where its adoption has settled in.
Artificial Intelligence
o3 in public: the reasoning leap is confirmed
o3-mini, the first public release of OpenAI's o3 reasoning series, clearly improves logic, math, and complex code over GPT-4o, though it answers slower and still hallucinates facts. This analysis, based on weeks of real use, explains where it pays off and where it does not.
Artificial Intelligence
Gemini 2.0: integrated tools and agent mode
Google ha lanzado Gemini 2.0 con un énfasis claro en uso de herramientas y agentes. Repaso de qué aporta, dónde está por detrás de la competencia y en qué tipo de aplicaciones encaja mejor.
Artificial Intelligence
NPU in the PC: faster, cheaper local AI
Qualcomm, Intel and AMD Copilot+ processors have normalised the presence of an NPU in everyday PCs. A 40 TOPS NPU can run quantised Phi-3 Mini drawing just 5-10 W, versus 40-50 W for a laptop GPU doing the same task. What actually changes for running AI models locally, and when it is worth it.
Artificial Intelligence
LoRA and QLoRA: Efficient Fine-Tuning on a Single Laptop
LoRA cuts fine-tuning cost for large language models by training only small low-rank adaptation matrices instead of every parameter in the base model. QLoRA adds 4-bit quantization on top, cutting required GPU memory by 65-75%, with quality loss of just 1-3% versus full fine-tuning.
Artificial Intelligence
Claude 3.5 Sonnet: The Model That Rewrote Price-Quality Balance
Claude 3.5 Sonnet (Anthropic, June 2024) matches Claude 3 Opus quality at Sonnet pricing, with a 200k-token context window and 92% on HumanEval. It stands out in coding and complex instruction-following, and introduced the Artifacts workspace feature on Claude.ai.
Artificial Intelligence
Mistral Large: European Contender Against GPT-4
Mistral Large 2, released by French startup Mistral AI in July 2024, is a 123-billion-parameter model with a 128k-token context window that rivals GPT-4o and Claude 3.5 Sonnet on several benchmarks. Its EU data residency and its 3 EUR per million input tokens pricing make it the most serious European alternative to US providers.
Artificial Intelligence
GPT-4 Turbo: Long Context and More Reasonable Costs
GPT-4 Turbo, released in November 2023, expanded GPT-4's context to 128,000 tokens and cut the input price threefold, down to 10 dollars per million tokens. GPT-4o now beats it on price, speed and answer quality, but Turbo still holds up in stable production apps, contracts pinned to a specific version, and deterministic tests that depend on its exact behaviour.
Artificial Intelligence
Constrained Decoding for Structured LLM Outputs
Outlines, Guidance e Instructor obligan al modelo a emitir JSON válido en el propio paso de generación. Cuándo ganan frente a reintentos y function calling.
Artificial Intelligence
Mixtral 8x22B: Open and Powerful Mixture of Experts
Mixtral 8x22B is Mistral AI's Mixture of Experts model released in April 2024: 141B total parameters but only 39B active per token, an unrestricted Apache 2.0 licence, and multilingual performance ahead of Llama 3 70B in Spanish, French, Italian, and German. Production serving needs datacenter-class GPUs.
Artificial Intelligence
Claude 2: Anthropic’s Alternative to GPT-4
Claude 2, launched by Anthropic in July 2023, offers a 100,000-token context window and safety grounded in Constitutional AI. Against GPT-4 it wins on long-document analysis and wide-context code; GPT-4 remains ahead on complex mathematical reasoning and its tooling ecosystem.
Artificial Intelligence
LangChain: The Framework for Orchestrating LLM Applications
LangChain is a Python framework that unifies building LLM applications: prompt templates, retrievers over vector databases, function-calling agents, and conversational memory. It earns its keep in fast prototypes and multi-model systems, but for a single well-defined production use case, direct code usually stays more maintainable.
Artificial Intelligence
OpenAI Function Calling: Structuring Model Output
Function calling converts the LLM into a component that returns structured data. How it works, real use cases, and common production pitfalls.
Artificial Intelligence
GPT-4 In Depth: Real Capabilities vs Expectations
Five months after launch, GPT-4 excels at chained reasoning, technical writing, and medium-complexity code, but still fails at arithmetic, post-cutoff information, and cross-conversation consistency. Claude 2 wins on long context; LLaMA 2 wins on cost and privacy.
Artificial Intelligence
Bard and PaLM 2: Google’s Bet on Generative AI
Google launched Bard in February 2023 with PaLM 2 as its answer to ChatGPT, unveiling the model in May the same year in four sizes: Gecko, Otter, Bison, and Unicorn. PaLM 2 competes with GPT-3.5 and GPT-4 on benchmarks like MMLU and BIG-bench, but Google's real edge is Workspace integration, not the model itself.
Artificial Intelligence
LLM Fine-Tuning: When It’s Worth Training Your Own
Fine-tuning your own LLM pays off in three cases: you need a very specific style or voice, a rigid structured output format, or you want lower cost and latency from a small specialised model. LoRA and QLoRA have cut the GPU cost, but preparing data and running the model in production are still expensive. For everything else, RAG and prompt engineering are usually enough.
Artificial Intelligence
Cerebras-GPT: 7 Open-Source LLM Models Ready to Use
Cerebras-GPT is a family of 7 open-source language models, ranging from 111 million to 13 billion parameters, trained by Cerebras Systems on its CS-2 processors with the standard GPT-3 architecture. Released on Hugging Face and GitHub under the Apache 2.0 license, they suit fine-tuning, research, and local inference, though they understand only English.