Where the AI market is talking

Zuckerberg says AI agent development going slower than expected

The last two years have been perfect for accumulating tech debt. 2023 you would have probably implemented your Agents with LangChain and RAG 2025 you'd use MCP and OpenAI/Anthropic Agent SDK. 2027 you will use a workspace frameworks (Amazon, Microsoft) sensor libraries and world models. Agents are a fantastic generational technologies, but in mid-2026 the environment they are operating in is quickly changing. The only way forward is to stay agile, understand model and vendor risk.

OpenWiki: CLI that writes and maintains agent documentation for your codebase

it's langchain, what did you expect

OpenWiki: CLI that writes and maintains agent documentation for your codebase

I think it's own sub-dir OpenWiki serves as an example of the sort of output you might get: https://github.com/langchain-ai/openwiki/blob/main/openwiki/... I agree with others this seems somewhat over-engineered; you can get similar results with a good prompt/skill; I guess the rest of the implementation here is intended as an agent-maintainer.

Ask HN: Who wants to be hired? (July 2026)

Location: Bucharest, Romania & San Francisco, CA Remote: Yes (preferred) Willing to relocate: Yes Technologies: Python, Django, Svelte / TypeScript, Shell, Linux, PostgreSQL, Redis, Docker, Hugging Face, OpenAI, LlamaIndex, LangChain, AWS, GCP, Unity (C#), Git, Github Actions, Software Design & Architecture, Devops, Product, UX, Sales Résumé/CV: Available on request Email: hi [at] lucianlazar [dot] io GitHub: https://github.com/xucian hey, I'm Lucian, have 13y of exp as a generalist swe, started as a gamedev then expanded into devops, ai/ml, trading, now working on a pretty diverse range of projects, from games, gamedev courses, ai tools, saas templates etc. more info on my profile looking for a fractional cto role, ideally, but I'm flexible

Ask HN: Who wants to be hired? (July 2026)

Location: Chicago, IL Remote: Yes Willing to relocate: No but open to relcoation occasionally like 1 or 2 times a year. Technologies: - AI Engineering & LLM Systems — LLMs (OpenAI GPT-4/5, Claude, Gemini, Llama, Mistral), Retrieval-Augmented Generation (RAG), AI Agents, Multi-Agent Systems, LangChain, LangGraph, LlamaIndex, MCP, Tool Calling, Prompt Engineering, AI Evaluation Frameworks, AI Safety & Guardrails, Conversational AI, Voice AI, Multimodal AI, Embedding Pipelines, Semantic Search - Software Engineering — Python, FastAPI, C#, ASP.NET Core, .NET Core, Node.js, TypeScript, JavaScript, React, Next.js, REST APIs, GraphQL, Microservices, Distributed Systems, Event-Driven Architecture, WebSockets - Cloud & Infrastructure — AWS, Azure, GCP, AWS Bedrock, Docker, Kubernetes, Terrafo…

Ask HN: Who wants to be hired? (July 2026)

Location: Greater New York (US Citizen) Remote: Yes Willing to relocate: No About: I am a Generative AI Engineer who also happens to be pretty handy across the full stack. Most of my recent work has been building RAG pipelines, AI agents, and natural language search — using LangChain, LangGraph, and vector databases like Pinecone and ChromaDB — with OpenAI, Anthropic, and Gemini APIs under the hood. I care deeply about shipping GenAI features that hold up beyond the demo stage. Beyond GenAI, I am working on the full stack — React/TypeScript on the frontend, Node.js or ASP.NET Core APIs on the backend, and AWS/Azure for infra. I like owning things end-to-end and building software that's simple and reliable. Technologies: • Generative AI: LLMs, LangChain, LangGraph, Agentic RAG, P…

Ask HN: Who wants to be hired? (July 2026)

Location: Rawalpindi, Pakistan Remote: Yes Willing to relocate: Yes if assisted. Technologies: LangGraph, RAG, FastAPI, Qdrant, AWS Bedrock, LLM agentic pipelines Résumé/CV: muhammadsamar.vercel.app Email: samarshehzad598@gmail.com

Show HN: I built an AI agent to yell at me about my ADHD

Thanks! You made me realize I completely forgot to mention memory in my post . For running memory, hex maintains a 10 message full-length context window, with the rest being summarized when the threshold is being crossed (15 messages) by a separate LLM call and stored in vector storage (Qdrant). Aside from that, observational memories get constantly stored in Qdrant as well for convenient retrieval and embeddings are calculated locally via Qdrant's FastEmbed. Preferences, facts and things like that are store in SQLite.

Ask HN: Who wants to be hired? (July 2026)

Location: Oslo, Norway (CET) Remote: Yes (EU/US time zones) Willing to relocate: Yes. Technologies: Rust, Python (PyTorch, NumPy/Pandas, Flask, Django), PyO3 Databases/infra: Postgres, ClickHouse, Qdrant, Redis, Kafka, Docker, Kubernetes (k8s) ML and systems: vLLM, TensorRT, Ray, edge ML inference, gRPC/Protobuf, systems profiling (perf, flamegraphs, Valgrind) Rust focus: Async services (Tokio, Axum/Actix), data/compute (Polars/Arrow), database proxies, performance-critical pipelines. Domain expertise: - Academic: PhD in mathematics (topological data analysis, representation theory), MSc finance, MSc neuroscience. - Applied: database proxies/networking, quant dev and backtesting engines, ML/DL for banking & finance (risk/fraud), LLMs and a…

Ask HN: Who is hiring? (July 2026)

Starbridge | Senior AI Engineer (Python) | NYC or Remote | Full-time | Starbridge is building an AI platform that turns large-scale public and enterprise data into reliable sales insights. We are early, moving fast, and building from zero to one, so this role will have huge ownership and product impact. AI Engineer: Applied AI plus strong software engineering. You will build, evaluate, and deploy LLM-driven features like deep document analysis and interactive chat, working with models from OpenAI, Anthropic, and Gemini. Expect hands-on Python, ML system design, experimentation, and production reliability. Bonus for RAG depth and frameworks like LangChain, LlamaIndex, or Hugging Face. We're looking to build our in-person team in NYC but also open to remote! Apply: https://starbri…

Show HN: PulseDB – Unified Redis and Pinecone Alternative in Python

Show HN: Shikhu – Understand the code your agents write

Agents write code really well, but it can get hard to understand what they've written and why. Shikhu is a CLI tool and Agent Skill that facilitates learning your code through self-quizzing, transcript analysis, and validation flows. I'm Arjun, I'm currently a developer advocate at Pinecone, and I use agents to code a lot! I enjoy using agents to code, but I've been feeling like I've lost some conceptual learning and understanding that comes with writing the code yourself. I wanted to make a tool that would make it easy to re-build habits around learning, while complementing workflows that use agents. I found inspiration for the tool from reading a research paper on skill formation from Anthropic, and Shikhu is the result of my attempt to build a tool around that: https://arxiv.…

[dead]

I found verified, live credentials leaking from Microsoft, Google, Red Hat, Grafana, and LlamaIndex in public GitHub repositories.

LlamaIndex integration for SynapCores (RAG, GraphRAG, and hybrid retrieval)

We built a LlamaIndex integration for SynapCores, an AI-native database that combines SQL, vector search, graph traversal, and AI functions in a single engine. The integration supports RAG, GraphRAG, hybrid retrieval, metadata filtering, and SQL-based querying via a single backend. We'd appreciate feedback on the API, developer experience, and any features you'd like to see.

LlamaIndex integration for SynapCores (RAG, GraphRAG, and hybrid retrieval)

Ask HN: What are some rock solid open source vector databases

Hi timvisee, Thanks for your work on Qdrant! I actually don't really have any complaints at this time. I'm using it in conjunction with Qwen3-VL to autocaption/describe images and then place them in a large vector Qdrant db so its easy to semantically search through ~100k worth of image assets.

[dead]

Hi HN, I created AccuVideo because I couldn't find a privacy-respecting way to query and search inside my video library without uploading gigabytes of data to third-party clouds. AccuVideo runs on your local machine to index, transcribe, and let you chat with your video content. The indexed points are vectors are stored in Qdrant (cloud). We are launching today on Product Hunt and I'm offering free trials (generating licenses manually every hour). I would deeply appreciate the HN community's feedback on how to improve the local indexing pipeline. Also in product hunt today. Thanks for your time.

Mistral OCR 4

(jerry from llamaindex here) we're gonna benchmark on ParseBench and report the results!

Ask HN: What are some rock solid open source vector databases

Dev of Qdrant here. What would change that to 'very happy' in your use case?

Ask HN: What are some rock solid open source vector databases

I run a dockerized version of qdrant for RAG related stuff and have been pretty happy with it. https://github.com/qdrant/qdrant