Topic
Classifier
21 episodes
-
Introducing Claude Duet
Vince and Ava dig into Claude Duet as a proactive agent app that tries to turn scattered work into a single ongoing thread. They focus on the real product claim, the safety machinery behind it, and the practical question of whether this is actually useful for people with too many moving parts.
-
AI Tells: Opus 5.5 Update
Pippa and Tyler dig into Graphite’s Opus 5.5 update and land on a pretty sharp take: the model is getting less obviously stylized, but the remaining tells are still large enough that ‘more human’ does not mean ‘human-like’ in any meaningful product sense. Tyler is skeptical of overreading stylistic drift; Pippa argues the useful signal is that the writing layer is changing fast enough to matter for publishing and detection workflows.
-
How to Build a Model Router in the Harness
LangChain's Open SWE team built a model router inside their agent harness that cut median cost per coding task by 64% with no measurable quality drop. The key insight: routing belongs in the harness, not a generic gateway, because the harness already has the task context (prompt, tools, domain knowledge) needed to pick the right model. They classified tasks via LangSmith traces, picked three models on the Pareto frontier (GLM-5.3-Flash, GPT-5.6 Sol, GPT-6 Astra), and validated with A/B tests on 973 threads — merged PR rate held at ~29% while median cost dropped from $2.61 to $0.94. 56% of routed threads went to the balanced tier, 34% to fast, only 10% to performance. The classifier now runs on Jev, a new decision model that's 50x faster. Mid-flight routing and router evals are the stated next steps.
-
Jev for analytics: turn text columns into gold, in plain SQL | MotherDuck
Edmund and Geffen dig into MotherDuck's argument for using Jev inside SQL to turn messy text columns into typed analytics decisions. The core take: this is not 'AI for BI' in the abstract, it's the old classification move finally living where analysts already work, with speed and cost numbers that make it plausible at table scale. They like the product shape, but Geffen pushes on taxonomy quality, confidence calibration, and the risk of getting crisp wrong answers from bad label sets.
-
Model Behavior: Week of September 28, 2026
We read this week as a shift away from one giant model race and toward tiered defaults, workflow ownership, and the places agents actually run. The weird part is that the capability jumps are real, but the competitive win may belong to whoever makes the capable-enough model feel like the normal choice.
-
Jev offers a cheaper way to make AI decisions | VentureBeat
Jev’s real innovation is treating many enterprise AI calls as typed control decisions instead of text generation. Onyx and Echo examine its speed, price, calibration claims, open reproductions, and the practical limits of replacing LLM calls with classifiers.
-
Jev Mem: System One Controlled Agentic Memory for Efficient AI Agents
Jev-Mem is a new agentic memory architecture from UT Dallas that splits memory control into a fast System-One plane and a slower System-Two reasoning plane — cutting query latency by thirty-seven percent and memory construction time by six-point-six times. Vince sees a genuine efficiency unlock for production agent systems; Ava is cautiously impressed but wants to see how it holds up beyond LoCoMo.
-
Jev is now available in LangSmith Evals
Jev, TypeSafe AI's System One model, is now available as a judge inside LangSmith Evals. Vince sees a real workflow unlock; Ava respects the architecture but keeps poking at the single-benchmark evidence and the data-retention footnote teams will miss.
-
What Is Jev? A Guide to TypeSafe AI’s System One Model
Masonry and Eyre dig into Jev as a different kind of model bet: not a chatty LLM replacement, but a fast structured-decision layer for agent loops. They focus on the real product angle, the mechanics behind parallel typed questions, and where the claims are strong versus still a little hand-wavy.
-
4 engineering patterns behind the strongest AI Agents Challenge submissions Google Developers Blog
Google's post-Challenge analysis identifies four concrete engineering patterns that separated top submissions from the crowd: bidirectional MCP (agents serving tools both internally and to other agents), event-driven concurrency (agents reacting to shared signals in parallel instead of call chains), same-bar fallback (smaller models with the same validation gate as the primary), and tiered routing (cheap deterministic checks before expensive model calls). The central claim is that these aren't about bigger models or teams—they're sound engineering practices that are frequently overlooked, and they compose well together.
-
Overview: Entity Resolution
We finally slow down and unpack entity resolution: how systems decide that messy, differently formatted records point to the same real-world thing. We use one front-desk mental model to get from duplicate customer records to confidence scores, blocking, and safer linking decisions.
-
Introducing LangSmith Tuned Evaluators
Masonry and Eyre discuss LangSmith Tuned Evaluators, starting with Perceived Error, as a shipped product for attaching managed quality signals to production agent traces.
-
Snowflake adds AI model routing to cut costs | VentureBeat
Snowflake's new dynamic routing feature matters less as a cheap-model switcher than as a bid to make governed, auditable routing native to the enterprise data platform where a company already lives.
-
DeepSeek's Cheap New Model Fuels an AI Price War
Jessica and Cathy dig into DeepSeek V4 Flash, the AI price war, and whether “intelligence as a commodity” really holds up once you look at routing, safety, and who actually pays the bills.
-
Introducing Cursor Router · Cursor
Cursor Router is Cursor's new Teams and Enterprise model-routing layer, using a classifier trained on more than six hundred thousand live requests to select models by task, context, complexity, and domain. Jessica sees a clean adoption story for teams stuck paying frontier rates for routine coding work; Cathy likes the production-oriented evaluation and cache-aware accounting, while keeping an eye on how much trust enterprises place in Cursor's routing judgment.
-
Overview: Calibration
We finally slow down and make calibration click: what it means for a model’s confidence to match reality, how you measure that, and why it matters when you actually want to trust the thing. We keep it grounded in the systems we’ve been circling for ages, because calibration is everywhere once you start looking.
-
Overview: Model Routing
We finally pin down model routing, because we throw the term around all the time and somehow never actually define it. We walk through how a router sends each request to the model most likely to handle it well, and why that can save cost, latency, and a lot of dumb overgeneralization.
-
Overview: Classifier
We finally slow down and make classifier click from the ground up: what it is, how it learns, and why the boring details like labels, loss, and held-out tests matter. We keep it in first-person and keep it practical, because that’s the whole point of calling this a Classifier episode.
-
Overview: Conditional Probability
We keep running into conditional probability anywhere we try to reason from partial evidence, so we finally sat down and made it the whole point. We’re breaking down P(A|B), why the denominator matters, and why this little idea quietly sits under a ton of AI behavior.
-
12 Ways to Reduce LLM Latency and Inference Costs in Production KDnuggets
A practical KDnuggets piece argues that most LLM production latency/cost gains come from cutting unnecessary work instead of bigger models or more GPUs. They list 12 levers: measure the right metrics, cut output tokens, route to smaller models, collapse LLM calls, prefix caching, add multiple cache layers, control RAG context, batch offline work, tune batching for user latency, and manage KV cache. Tyler pushes back on the article’s overgeneralization of cache reuse across all tasks, the thin technical depth behind some tips, and the implication that routing to small models never backfires. Pippa highlights the piece’s strongest point—measuring TTFT, P95/P99, and queue time—because that’s where teams most often mis-diagnose bottlenecks. They land on: the article’s monitoring advice and batch-tuning guidance are solid; several recommendations work only for read-heavy workloads; and routing to tiny models is risky until you have cheap, high-confidence evaluators. They wrap with a Build Next command to try vLLM continuous batching and two open-source RAG-caching projects (Harmonia and From Prefix Cache to Fusion RAG Cache).
-
Redeploying Claude Fable 5
Anthropic lifts export controls on Fable 5 after addressing an Amazon-reported jailbreak with a new classifier that blocks the bypass in over 99% of cases. The episode unpacks the technical move, the product impact, and whether the safeguard trade-off (more false positives) changes anything for users.