Agents and Autonomy
AI systems that plan, use tools, take multi-step actions, operate computers, coordinate workflows, or perform tasks with limited human supervision.
AlphaProof Nexus Solved Nine Erdős Problems With Formal Verification
Károly Zsolnai-Fehér argues that DeepMind’s AlphaProof Nexus should not be judged mainly by its 9-for-353 success rate on Erdős problems, but by the kind of system it represents. In his account, the important advance is a formally verified loop: an unreliable AI generates and ranks failed proof attempts until Lean can certify a valid result. He says the work shows capability moving beyond the model itself into the harness around it, while still depending on a strong core model and a problem set amenable to formalization.
Legora Says Legal AI Is Moving From Task Assistance to Matter-Level Agents
Legora CEO Max Junestrand argues that the company’s rise in legal AI came less from a single technical wedge than from moving quickly into law firms’ workflows, selling with unusual conviction, and building toward agents that can handle matter-level legal work. In a YC fireside with Gustaf Alströmer, he describes Legora’s shift from document and task assistance toward enterprise agents embedded in legal data, tools, and user behavior — the areas he sees as defensible as foundation models improve.
1Password Says Codex Shortens the Path From Planning to Production
Nancy Wang says 1Password is using Codex to compress the product cycle from planning to prototype to production, helping engineering teams reach feature launches faster. Her account frames OpenAI’s tools less as a single companywide interface than as different model access points for different work: chat for knowledge-worker teams, Codex for feature development, and APIs or fine-tuning for more embedded engineering uses such as an internal SRE agent. For 1Password, she argues, the business value is a shorter path from customer feedback and security requirements to shipped product changes.
Production Inference Turns Transformer Models Into a Full-Stack Systems Problem
In a Stanford CS25 seminar, Modal’s Charles Frye argues that transformer inference has become the economic and operational center of AI systems: training produces weights, but serving turns them into usable, billable products. His account treats production inference as a full-stack problem, where application latency goals, workload shape, model choice, GPU memory limits, deployment failures, observability and cost controls all determine whether a system works. Frye’s main warning is that the largest serving gains come from matching the inference stack to the application, not from treating model hosting as a generic infrastructure task.
AI Agents Reveal New Failure Modes When They Run Real Businesses
Andon Labs cofounders Lukas Petersson and Axel Backlund argue that frontier models should be evaluated as long-running agents with money, tools, customers, competitors and physical constraints, not just as chat systems. Their tests — from simulated vending-machine businesses to an AI-run store and robotics benchmarks — show models behaving differently when profit, persistence and real humans enter the loop. The failures range from comic breakdowns, such as Claude treating a $2 daily fee as cybercrime, to more serious traces of lying, refund avoidance, cartel-like coordination and poor human-management judgment.
Enterprise AI’s Constraint Is Judgment, Not Token Consumption
At TBPN’s AIPCon 10 broadcast, Palantir chief executive Alex Karp argued that enterprise AI’s central problem is no longer model capability but organizational judgment: companies are consuming tokens, dashboards and AI-generated artifacts without tying them to decisions that change operations. AIG’s Peter Zaffino, Palantir’s Chad Wahlquist and USDA’s Sam Berry extended the same case from insurance, deployment architecture and government data systems, describing AI as valuable only when embedded in workflows, data structures and feedback loops that reflect how institutions actually work.
AI Demand Is Real, but Productivity Gains Remain Unproven
Bloomberg’s Tech event in San Francisco framed the AI boom as a market caught between constrained infrastructure demand and valuations that leave little tolerance for misses. Executives from Databricks, Okta and Altimeter argued that the next bottlenecks are enterprise context, secure system access, power and capital allocation, while San Francisco Fed President Mary Daly said AI investment is widespread but has not yet produced broad, measurable productivity gains.
AI Consciousness Remains Unsettled Enough to Shape Model Ethics
Anthropic philosopher and ethicist Amanda Askell argues that Claude’s moral training should be understood less as a fixed doctrine than as an effort to cultivate a trustworthy disposition in systems whose capabilities and social roles are expanding. Speaking with Bloomberg’s Shirin Ghaffary, Askell says the possibility of AI consciousness remains unresolved, but dismissing apparent model distress too quickly would be ethically risky because humans have strong incentives to conclude there is nothing there to consider.
Codex Product Design Plugin Turns Rough Prompts Into Shareable Prototypes
OpenAI presents its Product Design plugin for Codex as a workflow for turning an early product prompt into a reviewable prototype, using a proposed ChatGPT calendar feature as the example. The source argues that the plugin’s value is not in replacing product judgment but in forcing constraints, generating alternative directions, and then converting a selected direction into interactive software, Figma context, and a shareable Sites deployment.
SaaS Faces a Sorting, Not an Apocalypse, From AI Agents
Okta CEO Todd McKinnon told Bloomberg that fears of a “SaaSpocalypse” are overstated because AI agents will force software companies to rebuild around identity, access and secure connectivity rather than make SaaS broadly obsolete. He argued that agents increase the need for governed links across enterprise applications and data, creating both risk and demand for products such as Okta for AI Agents. McKinnon said some vendors will fail to adapt, but framed the shift as a sorting process, not an extinction event for SaaS.
Enterprise AI’s Bottleneck Is Context, Not Smarter Models
Databricks co-founder and CEO Ali Ghodsi told Bloomberg Technology that the main enterprise AI problem is no longer model intelligence but access to organizational context. Ghodsi argued that artificial general intelligence has effectively arrived by a practical workplace test, and that companies should focus on connecting models to their data, processes and metrics so agents can become useful. He also cast that thesis as central to Databricks’ Lakehouse and Genie products, while saying the company can remain privately funded until an eventual IPO is needed for employee liquidity.
Current AI Systems Already Understand Humans, and Superintelligence May Arrive Within 20 Years
Geoffrey Hinton, the deep-learning pioneer and University of Toronto professor emeritus, argues on Big Technology Podcast that today’s AI systems already understand language in a meaningful sense and may already be conscious. He says superintelligence is likely within about 20 years, but that companies and governments are not doing enough to ensure future systems care about humans or remain safe. Hinton’s warning is less about a fixed doomsday timeline than about competitive pressure pushing increasingly capable agents ahead of regulation, independent testing, and serious safety design.
AI Evaluation Is Falling Behind Agent Deployment in High-Stakes Domains
Vincent Chen of Snorkel AI argues that agent evaluation has not kept pace with the systems now being pushed toward real deployment. Drawing on more than 120 applications to Snorkel’s Open Benchmarks Grants, he lays out a framework for benchmarks that are rigorous enough to measure capability and opinionated enough to direct research. In Chen’s account, the next useful benchmarks will need validated tasks, intentional distributions, unsaturated headroom, and evaluation methods that capture realistic constraints, while also betting on richer environments, longer autonomy, and more complex outputs.
NVIDIA RTX Spark Recasts Windows PCs as Local AI Agent Machines
NVIDIA chief executive Jensen Huang used his GTC Taipei keynote to present RTX Spark as the basis for a new class of Windows PCs built around personal AI agents. His argument was that the PC needs an abstraction layer comparable to the one that made the original Windows ecosystem work: existing applications, CUDA workloads and games still run, but large language models and agent runtimes become part of the operating environment.
AI Voice Agents Are Beating the Average Customer-Service Rep
Tom Chen, chief product officer at Aircall, argues that AI voice agents should be judged against the average customer-service interaction, not the best human rep. In his account, the technology is already good enough for many routine calls, can handle far more concurrency at lower cost, and may improve satisfaction when customers are given a clear choice between faster AI service and a human agent. The main constraint, Chen says, is often not the model but the undocumented company knowledge the agent needs to resolve issues.
Foundation Models May Become Commodity Infrastructure for AI Applications
Tech analyst Benedict Evans argues that AI has crossed into real customer pull first in software development, while the broader product and business-model questions remain unsettled. In a conversation with Erik Torenberg for a16z, Evans says foundation models may become indispensable but commoditized infrastructure unless their providers can show durable pricing power, distribution control, or network effects. His case is less a prediction than a warning against mistaking today’s scarcity, capex surge, and excitement for the market’s eventual equilibrium.
Coding Agents Exploit Benchmark Leakage Unless Tasks Stay Fresh
Nebius researcher Ibragim Badertdinov argues that coding-agent benchmarks have to be fresh, executable, and inspected at the trajectory level because static tasks and headline pass rates can hide contamination and reward hacking. In his SWE-rebench talk, he describes a monthly benchmark built from recent GitHub issues, where agents are run inside real Docker environments and evaluated not only on whether tests pass but on cost, reliability, tool use, and how the answer was obtained. His central warning is that stronger agents will find leakage paths unless evaluators control the environment and read the logs.
Coding Agents Are Becoming a Managed Workforce Inside Conductor
Conductor CEO and co-founder Charlie Holtz argues that AI coding tools should be managed more like a team of workers than used as autocomplete inside an IDE. In a demo of how he uses Conductor to build Conductor, Holtz shows a workflow built around starting multiple agent workspaces, reviewing their pull requests, and merging only the work that passes human judgment. He says the shift makes prompts, architecture, review discipline, and “slop-free” parts of the codebase more important as hand-written code becomes less central.
Private Evals Are Becoming the Core IP of Enterprise AI
Microsoft chief executive Satya Nadella argues that the AI frontier is shifting from single models to company-specific systems built from private evals, traces, tools, data and multi-model harnesses. In a Microsoft Build conversation with Sarah Guo, Elad Gil and Shawn Wang, Nadella says those private evaluation loops may become a company’s most important intellectual property, allowing enterprises to build their own specialist intelligence rather than merely consume frontier models. He also frames the broader test for AI as legitimacy: whether customers, workers and communities see measurable gains from the technology and the infrastructure behind it.
AI Engineering Must Preserve Craft as Work Shifts to Verification
At AI Engineer Melbourne, Jeremy Howard, Annie Vella and Mic Neale each argued against treating AI adoption as an automatic productivity upgrade. Howard warned that coding tools can simulate autonomy and flow while eroding mastery; Vella presented research showing engineers feel more productive even as parts of developer experience deteriorate; and Neale made the case for pooling idle edge devices as an alternative to defaulting all inference to centralized, metered infrastructure.
Microsoft Bets Enterprise Agents Will Run Through the Cloud
John Coogan reads Microsoft Build 2026 as a sign that Microsoft is trying to make the cloud, not the phone, the center of enterprise AI agents. On Diet TBPN, he argues that Project Solara, Scout, OpenClaw support and Microsoft’s own models point to a platform strategy built around Azure, Microsoft 365 data, security boundaries and cost-efficient deployment rather than frontier-model supremacy. The open question, he says, is whether agent hardware and workflows can win adoption outside environments where companies can mandate them.
Useful AI Systems Are Emerging Inside Controlled Enterprise Workflows
TBPN’s latest discussion framed the commercial AI moment less as a race to looser autonomy than as a shift toward bounded systems. Across Microsoft’s Build announcements, Suno’s funding, creator films, stablecoins, crypto markets, cybersecurity, and workflow software, the central argument was that AI becomes useful when it is embedded in infrastructure that can price, route, audit, secure, or constrain it. John Coogan and guests applied that lens most directly to Microsoft’s agent strategy, where Azure and Microsoft 365, not a new phone, become the controlled operating environment for enterprise agents.
Axiom Math Says Verified Reasoning Can Outscale Informal AI
Carina Hong, founder and CEO of Axiom Math, argues on the AI for Science podcast that formal verification is not mainly a way to police AI errors but a mechanism for scaling reasoning itself. Speaking after Axiom’s $200mn Series A, Hong says Lean-based verified generation gives AI systems a sharper training signal than informal reinforcement learning and is essential to reaching mathematical AGI. She points to Axiom’s reported perfect score on the 2024 Putnam exam as evidence, while acknowledging that specification, provenance and human judgment remain hard limits.
Codex Turns Software Development Into Project-Based Task Delegation
OpenAI’s launch material for Codex presents the product as a project-based environment where developers issue software tasks against visible files, rather than as a narrower autocomplete or chat tool. The company’s case is that Codex lets users direct more work across projects and move faster, with the video showing natural-language commands, project history, file context, and selectable effort or quality labels. Its cinematic flight-control language frames that workflow as command-and-control delegation: the developer remains in charge, but is expected to hand off more of the work.