AI Application Architecture
Patterns for building production AI applications, including orchestration, memory, tool use, state management, APIs, and system design.
Vertical AI Teams Need Domain Experts Who Own Quality Loops
Chris Lovejoy of Notius Labs argues that vertical AI companies increasingly fail or succeed on whether they can turn domain judgment into product quality, not simply on access to better models. He proposes three operating models for that expertise: an Oracle who both judges and changes outputs, an Evaluator who defines and measures quality while engineers implement fixes, and an Architect who designs systems that improve from use. His case studies of Granola, Tandem and Anterior show why the right model depends on whether quality is subjective, measurable, or too variable for manual iteration.
Context Graphs Make AI Decision Trails Queryable
Stephen Chin of Neo4j argues that enterprise AI systems need context graphs because retrieval alone can surface relevant facts while missing the relationships that make them usable. In his examples, a graph-augmented system can connect a patient’s emphysema care plan to smoking history or a credit decision to prior rejections, policies, margin trades and fraud signals. Chin’s case is that agents should preserve not only documents and answers, but the decision traces, tool calls, causal chains and outcomes that let humans inspect and reuse prior reasoning.
Economic Entanglement, Not Decoupling, Defines the New China Bargain
Salesforce CEO Marc Benioff joined the All-In hosts for a discussion that framed U.S.-China relations, enterprise AI, and the software selloff around the same question: when dependence is a stabilizer and when it becomes leverage. Benioff argued that more trade with China can lower conflict risk and that large software platforms remain valuable because AI still needs trusted customer data, cash-flowing distribution, and enterprise deployment. David Friedberg, Chamath Palihapitiya, and Jason Calacanis extended the argument across Taiwan, chips, AI assistants, El Niño-driven food risk, and private-market SPVs, where interconnection can either absorb shocks or transmit them.
AI Software Winners Will Own Context, APIs, or Outcomes
Tasklet chief executive Andrew Lee argues that AI software is consolidating toward a few horizontal agent platforms that hold context, connect tools, generate interfaces, and choose among models. In a discussion with Nathan Labenz, Lee says Tasklet has rewritten its agent stack around file-system memory, agentic search, and provider-specific context management because the chat transcript is no longer enough. He also frames Anthropic as both Tasklet’s critical supplier and a major competitor, making model neutrality central to Tasklet’s bid to survive the AI transition.
Legacy Infrastructure Is Slowing Enterprise Agentic AI Adoption
Kris Lovejoy, global strategy leader at Kyndryl, argues that enterprises are not being held back from agentic AI mainly by model capability or startup speed, but by the difficulty of running agents securely and reliably inside legacy infrastructure. In a conversation with Craig Smith, she says pilots are widespread but scaled deployments remain rare because agents need context, governance, compliance controls and modernized IT foundations before they can touch core systems. Her near-term prediction is narrower than much of the hype: by about 2031, agentic AI may handle roughly half of traditional line-one and line-two IT administration tasks, with humans still supervising the loop.
PFF’s Two-Engineer Agent Team Shipped 10x More Output
PFF CTO Mike Spitz argues that AI agents change the basic operating constraint of an engineering organization: the question is no longer how to make engineers faster, but how to make agents faster. In a three-month case study, he says two agent-heavy engineers shipped far more frequently than a ten-person team on the same codebase, with PFF measuring a 10x output gain per engineer and higher customer satisfaction. The result, in his account, was not the end of engineers but the removal of Scrum-era coordination rituals and a sharper split between agent-executed work and human judgment.
Supabase Says Skills and MCP Close the Agent Context Gap
Pedro Rodrigues of Supabase argues that agents fail on production systems less because they cannot reason than because they lack product-specific judgment. In a test using the same Postgres task, Supabase found that Claude with MCP alone created a view that could bypass row-level security, while MCP plus a Supabase skill added the required `security_invoker = true` flag. Rodrigues’s case is that MCP gives agents tools, but skills supply the rules, workflows, and current documentation paths needed to use those tools safely.
Intercom Doubled Engineering Throughput by Standardizing on Claude Code
Brian Scanlan, a senior principal engineer at Intercom, argues that the company doubled engineering throughput by treating AI coding as an internal platform strategy rather than an individual productivity tool. In his account, Intercom standardized on Claude Code, encoded recurring engineering work into agent-usable skills, connected agents to internal systems under existing controls, and made AI adoption an explicit expectation across R&D. The reported result was a doubling of pull-request throughput, including 17.6% of merged PRs approved by Claude, alongside new bottlenecks in review and CI.
Abridge Bets Clinical Conversations Can Become Healthcare’s Intelligence Layer
Abridge executives Janie Lee and Chaitanya “Chai” Asawa argue that the patient-clinician conversation is becoming healthcare’s core intelligence layer, not merely an input for automated notes. In a discussion with Redpoint’s Jacob Effron, they describe Abridge’s move from ambient documentation into clinical decision support, prior authorization and other workflows that depend on EHR data, payer rules, medical literature and local guidelines. Their case is that healthcare AI will be judged less by chatbot fluency than by whether it can deliver accurate, low-latency, privacy-preserving support inside clinical workflows without adding to clinicians’ alert burden.
Cerebras IPO Puts a Public Price on Fast AI Inference
TBPN’s John Coogan and Jordi Hays use Cerebras’s first day as a public company to frame a narrower AI hardware argument: the market is beginning to price low-latency inference as a product in its own right. Cerebras founder Andrew Feldman argues that fast inference will eventually consume demand for slow AI responses, while SemiAnalysis’s Doug O’Laughlin cautions that the company’s wafer-scale SRAM architecture may be limited by memory scaling and model size. The result is a public-market test of whether owning a valuable slice of the AI compute stack is enough.
Codex Is Moving From Code Generation to Delegated Knowledge Work
Codex is moving from a coding assistant toward an agent for delegated knowledge work, according to Thibault Sottiaux, OpenAI’s head of Codex. In an OpenAI Forum conversation with Chris Nicholson of OpenAI Global Affairs, Sottiaux argues that as models have become more reliable and better connected to workplace context, Codex is being used to research, organize information, create files and presentations, coordinate across tools, and run background tasks. That shift, he says, makes delegation, trust and access controls central as agents act across files, communications tools and company systems.
Choosing The Right Eval Matters More Than Tuning The Judge
Laurie Voss of Arize argues that agentic applications need the same engineering discipline as other production software: instrumentation, inspectable traces, targeted evals, and controlled experiments, not a handful of prompts that “look right.” In a hands-on workshop using a financial analysis agent, Voss shows how teams should read traces before writing evals, classify failures by root cause, and combine deterministic checks, LLM judges, custom rubrics, and human-labeled meta-evaluation. His central warning is that the choice of eval can dominate the result: the same agent scored 0 out of 13 on a correctness eval and 13 out of 13 on a faithfulness eval because the first judge was asking the wrong question.
Agent Observability Is Moving From Dashboards to Eval-Driven Optimization
Amy Boyd and Nitya Narasimhan of Microsoft argue that agent observability has to track the widening gap between what an AI agent is meant to do and what it actually does as models, prompts, tools and user behavior change. Their walkthrough of Microsoft Foundry frames observability as a loop of OpenTelemetry tracing, trace-linked evaluations, monitoring, optimization and red teaming. The central demonstration is an observe skill that can generate an evaluation dataset, run batch tests, optimize prompts, compare versions and roll back to the best-performing agent version from a sparse starting point.
Interwhen Verifies AI Agent Actions Before They Become Irreversible
Microsoft Research’s Amit Sharma presents Interwhen as a framework for moving AI agents from post-hoc checking to verified execution while they are still acting. The open-source library uses LLMs to turn natural-language instructions, policies, and partial responses into smaller verifiable properties, then applies symbolic or model-based verifiers to tool calls and intermediate behavior. Sharma argues that this lets agents continue normally when checks pass but interrupts them when a verifier detects a violation, addressing risks that final-output review may catch too late.
GitHub Agentic Workflows Turn Actions Into AI-Run Development Processes
Microsoft Research’s Peli Halleux and Yash Lara present GitHub Agentic Workflows as a move from AI-assisted coding to repository-level process automation. Their argument is that agents should be embedded inside GitHub Actions to research, plan, assign, and open pull requests under human review, rather than operate as unconstrained swarms. The system’s promised scale depends on orchestration, sandboxing, limited permissions, and Microsoft-hosted models on Azure.
MagenticLite Brings Full Agent Workflows to Small Language Models
Microsoft Research is presenting MagenticLite as a full-stack agentic system designed to make small language models usable for multi-step work across a browser and local files. Weili Shi, Harkirat Behl and Hussein Mozannar argue that the capability comes from specializing the stack rather than relying on frontier-scale models: MagenticBrain handles planning, coding and delegation, while Fara 1.5 controls the browser. The release also emphasizes user oversight, with the agent pausing for credentials, approvals or other points where the user needs to take control.
An Event-Sourced Agent Harness Separates State Replay From Side Effects
Jonas Templestein of Iterate argues that an agent harness can be reduced to an append-only event stream plus processors: synchronous reducers to derive state, and post-append hooks to perform side effects. His design puts model chunks, tool calls, errors, schedules, subscriptions and even processor deployment into the log, so a restarted agent can replay state without replaying old LLM calls. The larger claim is that agents and third-party services can compose by reading and appending to the same durable stream, with bounded waits and circuit breakers replacing tighter, blocking plugin interfaces.
GPT-Realtime-2 Turns Voice Agents Into Tool-Using Reasoning Systems
OpenAI’s Build Hour on GPT-Realtime-2 presented the new realtime voice release as a shift from conversational voice interfaces toward tool-using, stateful agents. Teri Yu and Erika Kettleson argued that GPT-realtime-2’s larger context window, stronger instruction following, parallel tool calling and controllable speech behavior let developers build voice systems that can operate apps, reason across workflows and know when not to speak. Sierra’s Ken Murphy and Soham Ray added that production voice agents still depend on the surrounding system: guardrails, tuned turn-taking, tracing, redaction, evaluations and customer-specific workflows.
Agents Can Now Fine-Tune Open Models Through Prompted Workflows
Merve Noyan argues that open models have moved from downloadable artifacts into an operational stack for selection, serving, inspection, training and deployment. In her Hugging Face presentation, she makes the case that access to model weights now matters because developers can quantize, fine-tune and run models locally or at the edge, while Hub benchmarks, inference providers, traces, MCP and Skills let agents act directly on those workflows. Her strongest example is a coding agent that can size hardware, choose infrastructure and launch a fine-tuning job from a prompt.
ElevenCreative Adds Templates for Reusable AI Creative Workflows
ElevenLabs is introducing Templates in ElevenCreative, a feature that turns its node-based Flows into reusable creative workflows with defined inputs and outputs. The company presents the tool as a way to run repeatable production tasks — such as product shots, mockups, style transfers, character sheets, and thumbnail translation — without rebuilding the workflow each time. Users can run templates from a gallery or publish their own, choosing which variables others can edit, what asset is returned, and whether access is private, workspace-only, or public.
Continuous Agents Need Stateful Compute, Not Traditional CI/CD
Madison Faulkner and Hugo Santos of Namespace argue that traditional CI/CD is organized around human-paced pull requests, and starts to fail when autonomous agents generate continuous, overlapping streams of code. Their proposed replacement keeps validation inside a stateful agent loop, uses caching and orchestration to avoid cold starts, and moves completed work into a pre-merge layer where humans review intent and outcome rather than every diff. The underlying CI functions remain, but the pull request stops being the system’s basic unit of work.
Agent Workflows Route Conversations Through Specialized Subagents
ElevenLabs is introducing Workflows, a visual editor for its Agents Platform that lets builders design routed conversation flows instead of placing all business logic inside one agent prompt. The company argues that specialized subagents, each with their own instructions, tools, knowledge bases and model choices, give teams more control over cost, latency and accuracy. The product is positioned as a way to combine AI interpretation with predefined actions, verification steps and human handoffs on the same design surface.
Compute Allocation Is Anthropic’s Core Constraint as Claude Revenue Surges
Anthropic CFO Krishna Rao argues that the company’s rise is best understood through compute: a scarce capital asset that must be bought years ahead and constantly reallocated across model training, customer demand, internal automation and future products. In an interview with Patrick O’Shaughnessy, Rao says ordinary forecasting and software-margin frameworks break down when model capability, adoption and revenue compound together, leaving Anthropic to manage growth through scenarios rather than point estimates.
Korean AI Dividend Proposal Triggers Semiconductor Stock Selloff
A South Korean policy chief’s proposal to return part of AI-related gains to citizens jolted the country’s chip market, with Samsung and SK Hynix closing down around 5% after Kim Yong-beom argued that profits from the AI infrastructure era should be shared more broadly. Bloomberg reported that the presidential office later described Kim’s post as personal opinion, while the same program pointed to related pressure points in the AI boom: CME’s plan with Silicon Data for compute futures and Nvidia CEO Jensen Huang’s absence from Trump’s China delegation as approval for Blackwell sales looked unlikely.