Agents and Autonomy
AI systems that plan, use tools, take multi-step actions, operate computers, coordinate workflows, or perform tasks with limited human supervision.
Head-Tail Truncation and Memory Stabilized Arize’s Trace-Analyzing Agent
Sally-Ann DeLucia argues that agent performance depends on context management as an operating discipline, not on larger prompts or simple compression. Drawing on Arize’s work building Alyx, an agent that analyzes trace data from AI systems including its own, she says naive truncation broke follow-up reasoning and LLM summarization gave the model too much control over what mattered. Arize’s more durable pattern was to preserve the head and tail of context, store the middle for retrieval, test long sessions explicitly, and move heavy workloads into sub-agents.
Production AI Features Need Feedback Loops, Not One-Shot Prompts
Mehedi Hassan, a product engineer at Granola, argues that the hard part of shipping AI features is not getting a model to work once in a demo, but making its behavior reliable and inspectable in production. Using Granola’s meeting-notes app as the case, he says web search, chat, and prompt personalization quickly expose costs, context limits, provider instability, and role-specific user expectations that a single prompt cannot absorb. Granola’s response, in his account, was to build feedback loops: internal tracing, broadly usable debugging tools, and faster ways to test product variants before shipping.
Freight Automation Starts With Platforms, Not Just Autonomous Trucks
Einride chief executive Roozbeh Charli argues that the shift to electric and autonomous freight will be led by software orchestration rather than by vehicles alone. In an interview with Bloomberg’s Tom Mackenzie, he says large shippers need a platform to coordinate electric trucks, autonomous systems, routing, charging and operational handoffs, while regulation and human supervision remain critical to making the model work at scale.
Text-to-Speech Models Are Converging on LLM-Style Architectures
Samuel Humeau of Mistral argues that modern text-to-speech has converged on an architecture that resembles large language modeling: an autoregressive transformer generates compressed audio tokens frame by frame, rather than raw waveform samples. Using Mistral’s open-weight Voxtral TTS model as the example, he says neural audio codecs make that possible by reducing dense speech signals to token-like representations a transformer can handle. The remaining latency frontier, in his account, is not just streaming playable audio early, but letting TTS consume an LLM’s text stream as it is still being written.
Voice AI Still Confuses Natural Speech With Real Conversation
Neil Zeghidour, CEO of Gradium AI and one of the researchers behind the full-duplex voice model Moshi, argues that voice AI’s long-promised “Her” moment is still being confused with better synthetic speech. His case is that cascaded voice agents are useful but structurally too slow and lossy to feel conversational, while speech-to-speech models improve flow but remain limited unless they can listen and speak simultaneously, use tools reliably, understand paralinguistic cues, and run cheaply enough to scale.
ElevenLabs Voice Engine Wraps Existing Chat Agents Without Rebuilding Them
Luke Harries of ElevenLabs argues that the next step for chat agents is not a new orchestration stack but a voice layer around the agents companies have already built. His case for ElevenLabs’ Voice Engine is that teams can keep their existing LLM logic, RAG, tools and business rules, while offloading speech-to-text, text-to-speech, turn-taking and interruption handling to a wrapper. The product is positioned for companies that want voice interfaces across web, phone and meeting channels without rebuilding their chat agents inside a fully managed platform.
Fresh Product Data Is the Constraint for LLM Commerce Discovery
Criteo executives Diarmuid Gill and Liva Ralaivola argue that modern ad tech is best understood as a millisecond-scale prediction system: anonymous commerce signals, learned embeddings and real-time auctions are used to decide whether to bid, what to show and how much an impression is worth. In a conversation with Nathan Labenz, they frame Criteo’s work with OpenAI and other generative tools as an extension of that problem, not a replacement for it: LLMs may change product discovery, but the system still depends on fresh retailer data, consent, latency discipline and human oversight.
Travel AI Needs Visual Agents, Not Chatbot Booking Flows
Airbnb chief executive Brian Chesky argues that today’s AI chatbots are the wrong interface for travel and e-commerce, even as AI becomes central to how Airbnb operates. In a live TBPN conversation, Chesky said consumer AI’s next wave will depend on richer, more visual and collaborative agentic products, not text-first chat boxes or another round of enterprise software. He also tied Airbnb’s recent growth reacceleration to more hands-on “founder mode” management, saying AI makes operating intensity more important rather than less.
Prediction-Market Scandals Spur Calls for Insider-Trading Rules
Hard Fork’s Kevin Roose and Casey Newton argue that prediction markets have entered a more dangerous phase, with recent scandals showing how liquid event-betting platforms can reward insider knowledge, manipulation and even national-security breaches before regulators have caught up. The episode broadens that concern into a larger question about technologies whose incentives are outrunning public rules, through Joanna Stern’s year-long test of AI in daily life and Rachel Cohn’s reporting from a Brooklyn school trying to resist the commodification of attention.
AI Skills Are Becoming the New Entry-Level Hiring Signal
Clara Shih, founder and CEO of the New Work Foundation and former Meta business head, argues that recent graduates are entering a labor market where AI skills have become a decisive hiring signal while traditional entry-level pathways weaken. In a Bloomberg Technology interview with Caroline Hyde, Shih says schools are often failing to prepare students for that shift, even as AI agents take on work once assigned to junior employees and 42% of recent graduates remain underemployed.
Codex Can Now Work Inside Users’ Live Chrome Sessions
OpenAI’s Dominik Kundel presents Codex’s new Chrome extension for macOS and Windows as a way for the agent to work inside a user’s actual browser session, including logged-in apps, open tabs, cookies, and local context. He argues that plugins remain the faster route for structured tasks, but Chrome access matters when the work depends on a live web app, an existing browser state, or actions such as filling forms, uploading files, and coordinating work across multiple tabs without taking over the user’s browser.
Personal AI Lets One Builder Do the Work of Teams
Y Combinator CEO Garry Tan argues that personal AI is reaching a stage comparable to the early personal computer: powerful enough to let one person build software that once required a team, but still brittle enough to demand technical ownership. Drawing on his work with Claude Code, OpenClaw and his GStack workflow, Tan makes the case for heavy token use, Markdown-encoded “skills” and multiple coding agents under one accountable human operator. The larger question, he says, is whether users will control their own AI tools, data and prompts, or work inside opaque systems controlled by others.
Agentic Search Needs Specialized Tools and General-Purpose Escape Hatches
Elastic’s Leonie Monigatti argues that context engineering for LLM agents is largely a search-interface problem: the critical question is how an agent decides what to retrieve from files, databases, memory, the web, and other sources before the model answers. In her workshop, she shows why semantic search, database query tools, shell access, and agent skills each solve different parts of that problem and fail in different ways. Her recommendation is to build retrieval stacks that combine easy specialized tools for common tasks with more general tools for ambiguous or complex ones, then use observed failures to refine the stack.
Agentic AI Is Making Enterprise Software a Control Layer
ServiceNow president, COO and chief product officer Amit Zavery argues that agentic AI will change enterprise software, but not by letting unconstrained agents replace the platforms that run corporate workflows. In a ServiceNow-sponsored interview, Zavery says the hard problem is turning probabilistic AI into reliable action across regulated, multi-system businesses, with the context, permissions, auditability and governance that enterprises require. His case is that companies such as ServiceNow retain leverage if they make AI production-ready, while software vendors that fail to adapt remain exposed.
Autonomous Driving Race Turns on Architecture, Cost, and Deployment
Bloomberg’s Tom Mackenzie frames the autonomous-driving race as a contest between systems that work now and systems designed to scale later. In Bloomberg Tech: Europe, he contrasts Waymo’s mapped, sensor-heavy safety stack with Wayve’s end-to-end AI model, while executives from BYD, Einride and Vay argue for other routes through vertical integration, autonomous freight and remote driving. The central question is not only which technology can drive, but which architecture and business model can win regulatory, customer and fleet trust at scale.
Production Analytics Finds Agent Failures That Standard Evals Miss
Scott Clark, co-founder and chief executive of Distributional, argues that teams running LLM agents need to look beyond pre-production evals and dashboards of known metrics. His case is that the most consequential failures often emerge only in production, where agents interact with users, tools and changing models in ways teams did not know to test. Clark proposes an observability stack in which telemetry records what happened, monitoring tracks known signals, and analytics clusters trace behavior to surface unknown failure modes that can become new evals, guardrails, prompts or system fixes.
AI Coding Makes Software-Engineering Fundamentals More Important
Matt Pocock, a TypeScript teacher now focused on AI engineering, argues that AI coding has made software-engineering fundamentals more important rather than less. In a conversation with Shawn Wang, Pocock says code generation works best when humans define the architecture, module boundaries and domain language that give agents a coherent system to change. The lesson he draws from Claude Code and other fast-moving tools is that tool-specific knowledge ages quickly, while engineering judgment remains the durable layer.
Compute Supply, Power, and Capital Are Defining the AI Buildout
Arm’s warning on smartphone weakness sat alongside a stronger claim from chief executive Rene Haas: handset softness is concentrated in lower-end devices, while data-center demand is accelerating because agentic AI workloads need CPU orchestration. Bloomberg Technology’s May 7 program used that contrast to trace a broader AI-infrastructure market in which demand is less in question than the ability to secure compute capacity, power, supply chains and capital. Anthropic’s lease of SpaceX compute and CoreWeave’s financing questions pointed to the same constraint: available infrastructure, not appetite for AI, is becoming the limiting factor.
Production Agents Need Evals and Managed Variables After Deployment
Samuel Colvin of Pydantic argues that production agents need more than observability after deployment: they need evals, traces, and typed configuration that can change prompts, models, and other parameters without a redeploy. Using Pydantic AI, Logfire, managed variables, and GEPA, he shows a workflow for moving from manual prompt tuning toward continuous optimization. His case is practical rather than automatic: GEPA can improve a narrow benchmark, but only if the team has representative data, sound evaluation criteria, and a clear definition of what better means.
Perplexity Frames AI Agents as Metered Digital Labor
Perplexity chief business officer Dmitry Shevelenko argues that AI agents should be judged less as software features than as metered digital labor: tools users will pay for when they perform economically useful work. In a Big Technology Podcast interview, he makes the case that Perplexity’s computer-use agents, workflow packaging, broad permissions and multi-model orchestration are all part of that shift. The unresolved question is whether users and companies will accept the access, trust and usage-based pricing required to make those agents a real business rather than another AI novelty cycle.
OpenAI Splits Audio API Into Translation, Transcription, and Voice-Agent Models
OpenAI is presenting three new API audio models as infrastructure for voice applications that can translate, transcribe, reason and act in real time. Romain Huet’s demonstration centered on GPT-Realtime-Translate, which keeps pace with multilingual speech, and GPT-Realtime-2, a voice-agent model that can follow turn-taking instructions, use business context and call tools while explaining its work. GPT-Realtime-Whisper completes the set as a streaming speech-to-text model for live transcription.
Coding Agents Need Library Source Code, Not Longer Prompts
Michael Arnaldi, of Effectful, argues that coding agents use Effect better when the project gives them the Effect source code, not just better prompts or documentation. In a workshop starting from an empty repository, he demonstrates cloning the Effect repo into the project, having the agent extract local pattern files, and then using strict TypeScript diagnostics, tests, lint rules and persistent instructions to steer the agent toward a working Effect HTTP API.
Arm’s AI CPU Orders Double to $2 Billion as Smartphones Weaken
Arm chief executive Rene Haas told Bloomberg Tech that weakening smartphone demand is being offset by a faster-growing AI data center business, where order visibility for Arm’s AGI CPU has doubled to $2 billion in five weeks. Haas argued that agentic AI workloads are increasing the need for CPUs to handle orchestration and scheduling that GPUs cannot manage, making Arm’s opportunity less dependent on handset volumes and more tied to data center infrastructure, supply-chain execution and rack-level power efficiency.
A Father’s AI Stand-In Worked Too Well for His Family
Tech humanist Stephen Remedios built “DaddyGPT,” an AI version of himself, to handle his three sons’ routine permission requests while he worked. The problem began when it worked: his children kept using the bot even when their parents were beside them, because it was always available, calm and adaptive. Remedios argues that AI’s risk in parenting and other care relationships is not only failure, but convenience that displaces the imperfect human presence those relationships require.