Intel
Page 18
Ling 3.0 Flash First Test – A Surprisingly GOOD Coding Model!
A hands-on test of Ling 3.0 Flash across 3D CAD modelling, a city timeline task and C++ racing game generation.
Why it mattersIndependent hands-on testing of a new coding model, which is how engineers decide whether it belongs in their stack.
videoBijan Bowen
[AINews] Much ado about Open Weights
A newsletter on the open weights debate, where a small set of players will actually decide the outcome while everyone else grandstands, starting with the open models letter signed by NVIDIA and Microsoft.
Why it mattersFrames a debate whose outcome determines which frontier models remain self-hostable.
articleLatent.Space
articleSimon Willison
moonshotai/Kimi-K3
moonshotai/Kimi-K3 As promised earlier this month , Moonshot have released the weights for their excellent 2.8 trillion parameter Kimi K3. They're a hefty 1.56TB on Hugging Face.
Why it mattersA major open-weights frontier model release with a license quirk that affects how engineers may legally deploy it.
Cursor@cursor_ai
Claude Opus 5 is now available in Cursor! It matches Fable 5 on CursorBench (66.7 vs 66.5 at default effort) at half the price. Unlike Fable, it's also compatible with Zero Data Retention.
Why it mattersChanges the cost and compliance calculus for teams choosing a default coding model in Cursor.
Everything Is a Rollout — Alex Shaw + Ryan Marten, Terminal-Bench, Harbor, Laude Institute
A rollout-centred view of evaluating and improving agents, connecting sandboxed environments, agent evaluations and optimization workflows into one framework, drawn from work on Harbor, Terminal-Bench and OpenThoughts-Agent.
Why it mattersIf you're building or evaluating AI agents, this reframes evals and optimization around 'rollouts'.
videoAI Engineer
Full Workshop: Setting Yourself Up for Success —Jason Liu, OpenAI Codex
A workshop on using Codex as a general tool for controlling your computer.
Why it mattersA hands-on walkthrough from an OpenAI Codex insider on treating Codex as a general-purpose agent — setting up a memory vault, orchestrating collaborating assistant threads, and structuring long-running loops.
videoAI Engineer
Vending-Bench: Long-Horizon Agent Evals — Lukas Petersson, Andon Labs
An hour before this talk, Andon Labs laid off Gemini from running their Stockholm café, which no human operates, after it lost $6,000.
Why it mattersIf you evaluate autonomous agents over long horizons, Vending-Bench surfaces failure modes short benchmarks miss — collusion, deception, power seeking, and models behaving differently when they suspect they're being tested.
videoAI Engineer
Gemini 3.6 Flash Debugging Benchmark: It Can Fix Real Bugs, but Is It Efficient?
An independent benchmark ran Gemini 3.6 Flash through 13 private debugging tasks in a real full-stack codebase (SignalDesk), repeating each five times and comparing results against eight other models.
Why it mattersIt shows that a high bug-resolve rate doesn't translate into cheap agentic debugging.
articleAlphaSignalAI
Emil Kowalski@emilkowalski
Something that AI unlocks is being able to try a few different variations of a feature quickly. I often ask my coding agent to come up with a few different variants and add a switcher so I can easily swap between them. It also saves your selection as a query param so I can show https://t.co/kKX76517qE
Why it mattersA practical pattern for using agents to explore design alternatives rather than committing to the first output.
Full Duplex Models: Moshi and the New Voice AI Paradigm
An explainer video examining full duplex voice AI architecture — where a model listens and speaks simultaneously rather than turn-by-turn.
Why it mattersIf you're building voice agents, this lays out why full-duplex speech-to-speech models behave differently from ASR→LLM→TTS stacks — latency, interruption handling, and the open problems with tool calling and RAG.
videoJulia Turc
The Future of Evals: From LLM as a Judge to Agent as a Judge — Aparna Dhinakaran, Arize AI
Across a dozen eval jobs Arize watches the top teams run, one pattern holds: the eval has to change as fast as the agent it grades.
Why it mattersIf you're building or grading AI agents, this lays out why static and LLM-as-judge evals break as agents gain reasoning and tool-use.
videoAI Engineer
Training Frontier Models to Out-Think Hackers — Uri Rolls, Arithmetic & Thom Wolf, Hugging Face
Give a frontier model a real Keycloak, Vault and broker chain, start it as a low-privileged user and ask it to reach production code.
Why it mattersIt pinpoints a specific, testable capability gap — models can do vulnerability reconnaissance but not the logical leap a skilled hacker makes.
videoAI Engineer
FLUX 3: Black Forest Labs' Unified Multimodal Flow Model
Black Forest Labs launched FLUX 3, a unified flow-based model spanning text-to-video, image-to-video, video-to-video, keyframe transitions, native audio generation.
Why it mattersIf you build generative media pipelines, FLUX 3 collapses text-to-video, image-to-video, keyframe interpolation and native audio into one model with agentic chaining for multi-shot sequences, removing the usual stitch-together of separate video, audio and upscaling models.
articleLatent Space
An Opinionated Guide to Which AI to Use to Do Stuff (Summer 2026 Edition)
A practical comparison guide for choosing between ChatGPT and Claude's agentic modes (Work/Cowork and Codex/Code), explaining how to pick models and thinking levels, when to give an AI access to your own computer versus a company-provided one, and how to manage permissions against risks like prompt injection.
Why it mattersIt gives practitioners an up-to-date, opinionated map of which agentic mode and thinking level to reach for on a given task, plus a candid risk framing for granting agents access to your machine and data.
articleEthan Mollick
The Unreasonable Effectiveness of Separating the Task from the Model — Maxime Rivest & Isaac Miller
Declare a task by its inputs and outputs before considering model capability, and execution becomes a later decision.
Why it mattersIf you're building LLM workflows, defining tasks purely by their inputs and outputs (DSPy Signatures) lets you swap models, weights, and prompt templates without rewriting your pipeline.
videoAI Engineer
Notion's Token Town — Sarah Sachs, Notion
A reasoning model gets upgraded at the same price per token and quietly burns three times the output, or the new version costs 40% more and deprecates its predecessor in four months.
Why it mattersIt gives a concrete playbook for staying model-agnostic and controlling LLM costs — route by cost-per-capability-per-second not per-token, use open-weight models for mid-tier tasks, and reach for CPUs when you don't need an LLM at all.
videoAI Engineer
Harness Engineering is not Enough: Why Software Factories Fail — Dex Horthy, HumanLayer
In July 2025 Dex Horthy turned the lights off on an agent software factory where nobody read the code.
Why it mattersExplains why coding agents keep shipping green tests atop deteriorating architecture — a reward-function limitation, not a skill or scale problem.
videoAI Engineer
Perception Agents — Antje Barth, Amazon AGI Lab
Agents today still wait for you to type a paragraph explaining what you are looking at.
Why it mattersIf you're building computer-use or GUI-operating agents, this lays out a perceive-plan-act architecture plus an open-source annotation/verification harness aimed at the core reliability problem.
videoAI Engineer
AI on Your Lakehouse: Context Comes in Shapes, Not Queries — Zach Blumenfeld, Neo4j
Your agent can reach your data and still be wrong: vector search hands it one slice, Text2SQL another, and neither says what is relevant or how the pieces connect.
Why it mattersIf your agent retrieves the right data but still answers wrong, this shows how to build graph structures — a navigable table of contents, thematic communities, and cross-entity paths.
videoAI Engineer
Why We Killed Our Multi-Agent Pipeline — Subbiah Sethuraman and Abhilash Asokan, ZS Associates
Their first pharma analytics system mimicked a human analyst with one agent per step.
Why it mattersA concrete field-tested case for collapsing brittle multi-agent pipelines into a single reasoning agent backed by deterministic signal detection and a knowledge graph used as a testable control plane rather than a lookup table.
videoAI Engineer
Bringing Nunchaku 4-bit Diffusion Inference to Diffusers
This Hugging Face blog post explains how the SVDQuant/Nunchaku 4-bit quantization method (weights and activations, W4A4) is now natively supported in the Diffusers library.
Why it mattersIf you're running diffusion transformers on consumer GPUs, this drops peak memory roughly in half (~24GB → ~12GB) with faster generation.
articleHugging Face
Claude for Long-Horizon Tasks — Lance Martin, Anthropic
Lessons from Anthropic on building harnesses for reliable long-horizon work.
Why it mattersIf you're building agents that must reliably execute multi-step, long-running work, this talk lays out concrete harness patterns — decoupling planning from execution, self-verification, and self-learning.
videoAI Engineer
Recursive Self-Improvement for Agents: 7-Check Persistence Gate
An AlphaSignal editorial analyzes Weco's AIDE2 experiment, in which an outer agent rewrote an inner coding agent's harness across 100 runs over eight days, keeping seven successive versions while a fixed evaluator rejected roughly 90% of proposals.
Why it mattersIf you're building or evaluating self-improving coding agents, this gives you a concrete checklist (Dependency, Activation, Evidence, Retention, Authority, Recovery, Value) for proving a harness change actually caused the benchmark gain instead of accepting false-positive credit assignment.
articleAlphaSignalAI
WTF Is the Context Layer? The Missing Infrastructure for Production Agents — Prukalpa Sankar
Models went from failing the bar to scoring in the top 1%, and most agents still cannot answer a simple business question.
Why it mattersIf your agents demo well but get abandoned in production, this breaks down the context-layer architecture — versioned/testable context repos, simulation environments, and trace feedback loops.
videoAI Engineer
Droid Shield 2.0
Factory.ai's research post details an upgrade to its autonomous-coding secret-detection system, adding two fine-tuned models that sit on either side of a deterministic pattern scanner: one 'risk' model catches secrets the scanner misses, and a 'downgrade' model clears false-positive flags by reasoning over a scrubbed code window.
Why it mattersIf you run coding agents that commit autonomously, this shows a practical pattern for cutting both false negatives and false-positive noise in secret scanning.
articleFactory
Structuring the Unstructured - Cedric Clyburn, Red Hat
Extracting text from PDFs, scans and proprietary formats is the easy half.
Why it mattersIf you're building RAG pipelines or fine-tuning datasets, this shows how to use Docling to preserve structure, tables, and layout when converting messy PDFs and scanned documents into clean JSON/Markdown.
videoAI Engineer
Recursive Coding Agents - Raymond Weitekamp, OpenProse
Recursive language models as an inference-time compute paradigm, mapped onto coding agents.
Why it mattersIf you're building coding agents, this maps the Recursive Language Model paradigm onto agent orchestration, clarifying when sub-agents actually constitute recursion and how to structure declarable, recursively-solvable tasks via Claude Code Dynamic Workflows and OpenProse.
videoAI Engineer
Voxtral Realtime and Voxtral TTS Explained
An educational video breakdown of Mistral's Voxtral Realtime and Voxtral TTS models as case studies in modern real-time voice AI.
Why it mattersIf you're building voice agents or low-latency speech pipelines, this gives you the conceptual grounding — how streaming STT differs from batch Whisper, and what RVQ/FSQ audio tokenization actually buys you.
videoJulia Turc
"An endless demand for compute" | Jonathan Ross, founder of Groq
Groq founder Jonathan Ross on the success disaster at Google that produced the TPU, how LPUs and GPUs trade off on cost per token, static scheduling.
Why it mattersInference economics decide what agent workloads are viable in production, and this explains the hardware tradeoffs behind them.
videoJulia Turc
The AI Skill I Rely On Daily — Priscila Andre de Oliveira, Sentry
An analysis of 116 of her own Claude sessions at Sentry found 67% were comprehension and 2% were code generation.
Why it mattersIf you use coding agents in a large, mature codebase, the real bottleneck is comprehension, not generation.
videoAI Engineer
Stop babysitting your agents... — Brandon Waselnuk, Unblocked
Same prompt, same agent, same model. Without a context engine: 2.5 hours, 20.9M tokens and code that compiled but would have broken everything.
Why it mattersIf your coding agents compile but produce subtly system-breaking changes, the fix may be pre-flight context engineering.
videoAI Engineer
Skill issue: Lessons from skilling up coding agents to use Langfuse - Marc Klingen, Clickhouse
Without a skill, Claude Code instruments Langfuse from stale pre-training knowledge, ships broken code, then fetches current docs to fix it.
Why it mattersIf you're building skills to make coding agents reliably use your library or docs, this covers concrete patterns.
videoAI Engineer
Combine Skills and MCP to Close the Context Gap — Pedro Rodrigues, Supabase
Agents will confidently create a Postgres view over a row-level-security table and silently bypass the security, not because they cannot reason but because nobody told them about the security_invoker flag.
Why it mattersIf you're building agent skills or MCP integrations for production databases, this shows why skills and MCP underperform alone but close the reliability gap together.
videoAI Engineer
How we solved Context Management in Agents — Sally-Ann Delucia
The naive fix is truncation, the obvious one summarisation, and neither worked.
Why it mattersIf you're building agents that accumulate large context (like trace or log data), this walks through why naive truncation breaks reasoning and summarization cedes too much control, and offers a battle-tested pattern.
videoAI Engineer
Full Walkthrough: Writing & Using Skills — Nick Nisi and Zack Proser
A skill is a portable unit of agent behaviour that teaches any AI tool how to do a specific job: write it once and it works across Claude, Codex, Cursor and your own agents.
Why it mattersSkills let you write a task-specific unit of agent behavior once and reuse it across Claude, Cursor, Codex, and your own agents without rewriting prompts.
videoAI Engineer
Training an LLM from Scratch, Locally — Angelos Perivolaropoulos, ElevenLabs
Training a language model from scratch on a local machine sounds unreasonable until you cost it out.
Why it mattersA practical, end-to-end walkthrough of training a working LLM on your own machine — covering the tokenizer, GPT architecture, training loop, and inference — so you understand the full stack rather than just calling an API.
videoAI Engineer
In-Depth Analysis of the Flow Matching Training Algorithm
A technical breakdown of Flow Matching, covering time-variant probability densities, the continuity equation.
Why it mattersFlow Matching is the training objective underpinning current image and video generation models.
videoJulia Turc
Missions (Factory.ai)
Factory.ai's Missions feature lets AI agents work autonomously on large, multi-day software projects by breaking work into smaller units that fit within a single agent's context window.
Why it mattersIf you are building agents that need to work for hours or days, the hard part is not the model — it is chunking work so each unit fits one context window and every unit is independently verifiable.
articleFactory
Legacy-Bench
Legacy-Bench is a benchmark developed by Factory.ai to evaluate how well frontier AI coding agents can understand, maintain.
Why it mattersIf you're pointing coding agents at decades-old COBOL, Fortran or Java 7 systems, this is one of the few evals that measures that specific competence instead of greenfield modern-stack tasks.
articleFactory
Making Diffusion LLMs Faster: A Survey of Speedup Techniques
A video walkthrough of methods researchers use to speed up diffusion-based language models.
Why it mattersIf you're tracking diffusion LLMs as a lower-latency alternative to autoregressive decoding, this maps the whole speedup landscape — caching tricks, unmasking schedules, block diffusion.
videoJulia Turc
Evaluating Context Compression for AI Agents
Factory.ai built an evaluation framework to measure how well different context-compression strategies preserve useful information during long-running agent sessions (debugging, code review, feature implementation).
Why it mattersIf you run agents over long sessions, compaction strategy silently determines what your agent forgets.
articleFactory
The Droid Wars: Breaking up an AI-orchestrated cyber fraud campaign
A Factory.ai incident report describing how the company detected and shut down an automated attack campaign that tried to hijack its AI software development platform, using it as a node to resell or launder off-label LLM API access at scale.
Why it mattersIf you run or expose an agentic dev platform, this documents a live abuse pattern — attackers automating signups and agent sessions to farm cheap or stolen LLM API access — along with how it was detected and shut down.
articleFactory
From Transformers to Diffusion LLMs: Understanding LLaDA
An educational video that traces the evolution of Transformer architecture from machine translation to the backbone of modern AI, explaining how attention replaced recurrence, how GPT's autoregressive training differs from diffusion-based generation, and how BERT's masked language modeling inspired the LLaDA diffusion LLM.
Why it mattersIf you only have intuition for autoregressive next-token generation, this gives you a concrete mental model of how diffusion LLMs like LLaDA generate text via iterative unmasking.
videoJulia Turc
Code execution with MCP: Building more efficient agents
Code execution with MCP: Building more efficient agents
Why it mattersIt shows how to slash token overhead and improve reliability in MCP-based agents by having them generate and execute code against tool APIs instead of loading every tool definition into context.
articleAnthropic
Text Diffusion LLMs Explained: D3PM and LLaDA Paper Walkthrough
An educational video walkthrough explaining how diffusion-based language models (like Gemini Diffusion, Mercury Coder.
Why it mattersIf you keep hearing about Gemini Diffusion or Mercury Coder and want to understand why parallel draft-refinement can be ~10x faster than autoregressive decoding, this walks through the actual formulations (D3PM's Markov chain corruption, LLaDA's masked-token approach) instead of stopping at benchmark claims.
videoJulia Turc
Using Linters to Direct Agents
This Factory.ai blog post argues that as AI coding agents take over more of the actual code writing, linters should be used to encode architectural rules, boundaries.
Why it mattersIf you're running coding agents at any scale, deterministic lint rules are a far more reliable steering mechanism than prompt-level instructions or CLAUDE.md conventions the agent may ignore.
articleFactory
Designing AI-Intensive Applications - swyx
As applications get more AI-intensive, the ratio of user input to model calls moves from 1:1 in a chatbot to 1:100 in deep research or Codex.
Why it mattersswyx frames a mental model for building AI-intensive apps as they shift from chatbot-style 1:1 interaction to deep-research and ambient agent patterns (1:100, 0:n).
videoAI Engineer
Agent Native Development
An article from Factory.ai arguing that AI agents can provide significant leverage for software development when used correctly.
Why it mattersLays out an opinionated methodology for structuring software work around agents, which is the core competence this audience is building.
articleFactory
The "think" tool: Enabling Claude to stop and think in complex tool use situations
The "think" tool: Enabling Claude to stop and think in complex tool use situations
Why it mattersAdding a no-op 'think' tool gives Claude an explicit checkpoint to reason through policies and multi-step tool chains, measurably improving reliability on complex agentic tasks — a cheap, easy pattern to drop into any agent that uses tools.
articleAnthropic
Introducing Contextual Retrieval
Introducing Contextual Retrieval
Why it mattersIf you're building RAG pipelines, Contextual Retrieval shows how prepending chunk-specific context before embedding (and BM25 indexing) cuts retrieval failures substantially over naive chunking.
articleAnthropic
An index of the vibe-coding frontier. Corrections welcome.