Versioning an AI Agent's Tools: Schema Evolution and the Regression a Validator Can't Catch
You can change a tool's schema in a fully backward-compatible way and still break your agent. The contract has two consumers that version differently — your code, which you can pin, and the model, which you can't.
5 min
Semantic Caching for AI Agents: Why the 73% Cost-Cut Number Doesn't Apply to You
The headline savings from semantic caching are real — and they come from a workload your agent doesn't have. Two different things are both called 'caching,' and only one of them is safe to put around a tool call.
4 min
How to Redact PII Before It Reaches an LLM Without Breaking the Task
Replacing every name with "[PERSON]" tells the model John and Jane are the same person — and one-way masking means you can never put the real name back in its reply. Redaction is the easy half.
4 min
RadixArk: Why NVIDIA, AMD, and MediaTek All Wrote Checks for the Same Open-Source Inference Engine
SGLang's team spun out as RadixArk on a $100M seed at a $400M valuation. Read the cap table, not the press release: hardware rivals rarely fund the same software unless it threatens something they all share.
4 min
Making a Pydantic AI Agent Crash-Proof: Temporal vs DBOS vs Prefect vs Restate
Pydantic AI now speaks four durable-execution backends with near-identical code. That means the choice isn't about the framework — it's about which piece of infra you're willing to run.
4 min
OpenTelemetry for AI Agents: The Span Tree Is Stable, the Attributes Aren't
The GenAI semantic conventions are still 'Development' and change almost every release. That sounds like a reason to wait. It isn't — you just have to instrument the part that's holding still.
4 min
Omnigent: Databricks' Meta-Harness for Running Claude Code, Codex, and Cursor as One Layer
Databricks open-sourced a common orchestration layer over Claude Code, Codex, Cursor, and your own agents — swap the harness in one line of YAML. The interesting bet isn't portability. It's who reviews the code.
5 min
How Many GPUs Does Your AI Agent Need? NVIDIA's NeMo Agent Toolkit Sizes the Cluster by Load Test, Not Math
The sizing calculator in NVIDIA's NeMo Agent Toolkit profiles a multi-agent workflow under concurrency and extrapolates a GPU count. The quiet lesson: an agent's cost is emergent, not calculable.
5 min
Memora vs Wiki Memory: Two Bets on What Agent Memory Should Be
Microsoft and LangChain shipped agent-memory frameworks a day apart in June. They disagree on the one axis the benchmarks don't measure — whether you should be able to read what your agent remembers.
4 min
MCP Tool Annotations, Explained: What readOnlyHint, destructiveHint, and idempotentHint Actually Guarantee
The four booleans on an MCP tool look like a permission model. They aren't — they're a risk vocabulary for trusted servers, and wiring them into auto-approval is the mistake.
5 min
MCP Finally Has a Deprecation Policy: A 12-Month Guarantee That Stops at the Core
The 2026-07-28 spec's quietest change is the one that decides whether you can build a business on MCP — a formal feature lifecycle with a year of runway. The catch is where the guarantee ends.
4 min
MCP Caching Explained: ttlMs, cacheScope, and the One Word That Leaks User Data
The 2026-07-28 spec lets an MCP server tell clients how long a result stays fresh and whether it's safe to share. One of those two fields is a performance knob. The other is a security boundary people will read as a performance knob.
5 min
MCP Apps, Explained: How Servers Render Interactive UIs in Sandboxed Iframes
The 2026-07-28 spec ships MCP Apps as an official extension. The sandboxed iframe everyone points to is not the security boundary — the consent path is, and that changes what you should actually worry about.
5 min
LongCat-2.0: China's Biggest Model Yet Was Trained on Domestic Chips — and Meituan Won't Say Whose
Meituan's 1.6-trillion-parameter LongCat-2.0 claims end-to-end training on 50,000+ domestic accelerators, no NVIDIA involved. That claim is the story — and the fact that it names no chip vendor is the part worth reading closely.
4 min
How Reliable Is Your LLM Judge? That's Half the Question
Rerun the same eval and an LLM judge flips 1 in 7 of its verdicts — while its own scores show no real difference between the answers. Reliability and validity are two different axes, and the number most teams report can't see either one.
5 min
LangGraph Node Timeouts: run_timeout vs idle_timeout for Agent Nodes
LangGraph 1.2 shipped per-node timeouts with two knobs that look interchangeable and aren't. Pick the wrong one and you either kill healthy slow work or never catch the hang you added it for.
5 min
LangGraph's DeltaChannel: The Checkpoint Cost That Scales With Your Thread
Every superstep, the default channel re-serializes your entire message list into the checkpoint. On a long-running agent, that write cost grows with the conversation — and DeltaChannel is the fix that finally makes it linear.
6 min
How to Structure an Agent Skill: Progressive Disclosure vs. a Flat File
The same procedure, packaged two ways. A controlled study finds the layout of a skill changes what the agent actually does — not just how many tokens it burns.
5 min
How to Run Agent Evals in CI Without a Flaky Gate
A pull-request gate has to give a clean yes or no. Agent quality is graded and noisy. Wire those two facts together naively and you get a gate engineers learn to re-run until it's green.
4 min
How Many Tools Should an AI Agent Have? Your Retriever's Recall Can't Tell You
Retrieve 100 tools and the right one is 'in the list' 99% of the time — the same odds a random shortlist gives you. Two 2026 papers show why recall is the wrong number, and why fewer tools win.
5 min
gpt-realtime-2.1 and the Voice-Agent Latency Tail: Why the Fix Was Caching, Not a Faster Model
OpenAI cut p95 latency 25% across its Realtime voice models by improving prompt caching — and where that speedup lands tells you why your agent slows down as the call goes on.
4 min
ADK 2.0 Turns Agents Into Graph Nodes: Inside Google's Workflow Runtime
Google's Agent Development Kit shipped a graph-based execution engine — and quietly retired the org-chart of agent types that used to be its whole pitch against LangGraph.
4 min
Context Offloading for AI Agents: Writing Tool Results to Disk to Beat the Context Window
The counterintuitive fix for context bloat is to stop reading tool output. Offload the payload to a file, hand the model a pointer — and move the retrieval decision from write-time to read-time.
4 min
Ambient Agents and the Agent Inbox: The Bottleneck Isn't Autonomy, It's Review
The leap from chat agents to always-on, event-triggered ones gets framed as a question of how autonomous the agent can be. The harder, quieter constraint runs the other way.
4 min
When Agent Memories Contradict: Don't Let the LLM Decide Which One Is Fresh
Your agent stores the same fact twice with different values. The intuitive fix — ask the model which is newer — is the one 2026's benchmarks say to avoid.
4 min
How to Publish and Install an Agent Skill in 2026
The SKILL.md format takes five minutes to learn. The part that actually decides whether your skill works is the one sentence you're most tempted to rush.
4 min
The Quiet Default Flip: Agent Frameworks Now Ask Before They Act
In mid-2026 the three biggest agent frameworks converged on the same primitive — tool calls gated behind a human approval — and Microsoft made it the default for anything a skill brings in. It's the security fix sandboxing couldn't provide.
6 min
Agent Skills Have a Supply-Chain Problem, and the Sandbox That Saved npm Isn't Coming
Studies this year found prompt-injection patterns in roughly a quarter to a third of scanned agent skills. The scary part isn't the number — it's that the standard fix doesn't apply.
5 min
Why Your LLM Isn't Reproducible at Temperature 0 — and How to Fix It
Setting temperature to 0 doesn't make an LLM deterministic. The real culprit isn't sampling or 'random' GPU math — it's that your request's output depends on who else is in the batch.
4 min
Tenstorrent Built a CPU for the Agent Loop: Inside TT-Ascalon S
The AI-hardware story has been about matmul for a decade. Tenstorrent's new RISC-V core is a bet that the agentic bottleneck is quietly moving back onto the CPU's branch-heavy control plane.
4 min