AI, voice agents & platform engineering
Long-form posts on voice AI, WhatsApp automation, RAG, and building production-grade customer platforms.
18 posts
12 min readAI Hardware Beyond GPUs: The 2026 Accelerator Landscape
NVIDIA dominates the AI accelerator market with approximately 80% share. But dominance invites competition, and 2026 is the year that competition became credible. Google, Amazon, AMD, Cerebras, and a wave of startups are shipping chips that challenge NVIDIA on specific dimensions — training throughp…
Read more5 min readKnowledge Graphs vs Vector RAG: When to Use Which in 2026
RAG is the standard pattern for grounding LLMs in private data. The default uses vector search. Knowledge graphs offer a different approach with different trade-offs. How Vector RAG Works Chunk documents, embed them, store in a vector database, retrieve by semantic similarity, and inject into the pr…
Read more
5 min readGPT-5.5 vs Claude 4: A Head-to-Head Comparison in 2026
In 2026, the two most-discussed frontier models are OpenAI's GPT-5.5 family and Anthropic's Claude 4 series. Both are capable. The difference is in how they work, what they cost, and what they are best suited for. The Model Families GPT-5.5: Instant (latency and cost), Pro (balanced), Thinking (exte…
Read moreGPT-5.5 Thinking vs Instant: When to Use Each
OpenAI's GPT-5.5 line ships in two main flavors plus a Pro tier: Instant, Thinking, and Pro. They are not three different models in the old sense — they are three different reasoning modes over the GPT-5.5 family. Picking the right one is the difference between snappy answers, deep analysis, and bur…
Read moreMoE vs Dense Models in 2026: Which Architecture Wins
The architecture wars are mostly settled in 2026 — but not in the way 2024's debates predicted. Mixture-of-Experts dominates the 100B+ flagship class: DeepSeek V4, Llama 4 Maverick, Qwen 3.5 397B-A17, Mistral Large 3 — all sparse MoE. Meanwhile, dense holds the mid-tier: Mistral Medium 3.5 at 128B i…
Read moreLangGraph vs OpenAI Agents SDK: Which to Pick
The agent-framework landscape consolidated faster than most people expected. By mid-2026 two names dominate production stacks: LangGraph 1.x from the LangChain team and the OpenAI Agents SDK, released in March 2025 as a production-grade replacement for the experimental Swarm framework. They solve th…
Read moreAgent Evaluation Frameworks: Braintrust, Inspect, Langfuse, and DIY
The hardest question in agent engineering is not "how do I build it?" — frameworks have solved that. It is "is the new version better than the old one?" Without a credible answer, every prompt change is a vibe-check and every model bump is a coin flip. By 2026 the evaluation tooling has matured enou…
Read moreAutonomous Coding Agents in 2026: Claude Code, Codex, Vibe
Two years ago "autonomous coding agent" meant Devin's first demo and a wave of skepticism. By April 2026 the field has consolidated to a handful of production-grade options — Claude Code, Cursor, OpenAI Codex, Replit Agent 3, and Devin — each with a distinct opinion about how much autonomy is approp…
Read moreStructured Output vs Tool Use: Which When
By 2026 the "JSON parsing with regex" era is over. Both major model APIs offer constrained-decoding paths that produce schema-valid output, and tool use is mature enough that one or the other handles 90% of structured generation workloads. The remaining question is which to reach for — and the answe…
Read moreOllama vs LM Studio: Running LLMs Locally
Local LLM runtimes have stopped being a niche hobby in 2026. With 70B-class models running comfortably on a 24GB GPU and 32B-class models running on Apple Silicon laptops, "the model is on my machine" is now a mainstream deployment shape. The two tools that anchor this category are Ollama and LM Stu…
Read moreCursor vs Claude Code vs GitHub Copilot: 2026 Showdown
The "AI coding tools" market has consolidated. By mid-2026 there are three tools that almost every working developer either uses or has tried: Cursor, Claude Code, and GitHub Copilot. They are not the same shape — one is an IDE, one is a terminal-native agent, one is a multi-IDE extension — and the …
Read moreAI Code Review Tools in 2026
The promise of AI code review is simple: a bot that reads every PR, surfaces real bugs, and lets human reviewers focus on architecture and intent. The reality in 2026 is messier — the good tools meaningfully reduce time-to-merge on routine PRs, the bad ones flood reviewers with noise, and the differ…
Read moreVector Databases in 2026: Pinecone, Qdrant, Weaviate, pgvector
The vector database market has consolidated. By mid-2026 four products account for the overwhelming share of production RAG and embedding-search workloads: Pinecone, Qdrant, Weaviate, and pgvector. Each represents a distinct philosophy — fully managed serverless, OSS-first with a managed tier, hybri…
Read moreEmbedding Models in 2026: OpenAI vs Cohere vs Open Source
The choice of embedding model shapes everything downstream in a RAG system — retrieval quality, storage cost, query latency, and ceiling on hybrid-search performance. In 2026 the field has narrowed to a clear set of contenders: OpenAI's text-embedding-3 family, Voyage AI's voyage-3 / voyage-3-large,…
Read morevLLM vs TGI vs SGLang: Inference Engines Compared
If you self-host an LLM, the inference engine is the single highest-leverage piece of infrastructure you choose. By 2026 the decision has narrowed: most teams pick vLLM, some pick SGLang for prefix-heavy workloads, and TGI has entered maintenance mode. Here is the picture. TGI: end of an era Hugging…
Read moreFine-Tuning vs RAG: The 2026 Decision Framework
"Should we fine-tune or do RAG?" is a question that has lost most of its drama. By 2026 the field has settled on a clear answer: they do different things, and most production systems use both. The interesting question is no longer "which one?" but "what belongs in which?" The single most useful ment…
Read moreSpeech-to-Text in 2026: Whisper, Deepgram Nova, Saaras V3, and the Real-Time Race
For most of 2024 and 2025, the speech-to-text question was simple: "Whisper, or one of the latency-tuned commercial APIs?" In 2026 the picture is more interesting. The leading models now diverge sharply by use case — real-time vs. batch, English vs. multilingual, accent-tolerant vs. literal — and pi…
Read moreTTS Showdown 2026: ElevenLabs vs. Cartesia vs. OpenAI vs. Sesame
Text-to-speech got good somewhere in late 2024. By 2026, "good enough to fool a casual listener" is table stakes for every major vendor. The interesting differences now are at the edges: latency under 100ms, instructable emotion, self-hostability, and the long tail of accents and languages. Here is …
Read more