Sources

Introducing GPT‑Live OpenAI finally upgraded the model used by ChatGPT voice mode! I’ve had preview access for a few weeks in the iPhone app, and the new model is very impressive. It also has the ability to spin off harder tasks to GPT-5.5: For questions that require web search, deeper reasoning, or more complex work, it delegates to our latest frontier model behind the scenes and brings the result back into the conversation when it’s ready. While it works, GPT‑Live can keep talking with you an…

Rewriting Bun in Rust simonwillison.net

Rewriting Bun in Rust Jarred Sumner has been promising this blog post ( since May 9th ) about his Zig to Rust rewrite of Bun for significantly longer than it took him to finish the rewrite. Honestly, it was worth the wait. This is a detailed description of an extremely sophisticated piece of agentic engineering, featuring dynamic workflows, trial runs, adversarial review and all sorts of other interesting tricks. Jarred spends the first half of the post praising Zig for getting Bun this far. Th…

NVIDIA Nemotron Achieves Benchmark-Leading Performance With LangChain Deep Agents Harness blogs.nvidia.com

NVIDIA Nemotron 3 Ultra is offering leading performance at lower cost than top closed models with the largest and most widely adopted AI agent orchestration platform. LangChain tuned its Deep Agents harness for NVIDIA Nemotron 3 Ultra, achieving the highest accuracy among open models, while completing more tasks at higher throughput and running at 10x […]

Hackers can use 9 of the most popular AI tools to assemble massive botnets arstechnica.com

Attackers exploit LLMs’ refusal to say ‘I don’t know’ by registering package names the models hallucinate, then waiting for developers to install them. The technique, dubbed HalluSquatting, works against nine widely used AI coding assistants and can seed massive botnets through poisoned dependencies.

Quoting Kenton Varda simonwillison.net

Varda argues the generated descriptions restate what any reviewer sees in the diff while skipping the high-level intent that makes a change reviewable. His moratorium covers PRs, commits, and tickets, framing AI summaries as actively worse than useless for code review.

Why AI Infrastructure must evolve for Agent Experience — Akshat Bubna, Modal CTO latent.space

Two years after Latent Space’s first Modal interview, cofounder Akshat Bubna returns to explain why ‘Agent Experience’ is finally clicking and what infrastructure changes it demands. The conversation covers lessons from running an agent-first cloud where machines, not developers, are the primary users.

The foundational elements of AI architecture that IT leaders need to scale technologyreview.com

As agentic systems multiply use cases, IT leaders face bets that must still pay off six months out. The piece returns to foundational architecture choices—data, compute, and governance layers—that survive the shift from single-model deployments to agent fleets.

My thoughts on Fable bensbites.com

Tossell uses new coding agent Fable to sketch the features he expects from any serious harness: persistent context, clear handoffs between tools, and control over when the agent pauses for human input. The post doubles as a checklist for evaluating rival harnesses.

How AI could enable autonomous robot workers in workplaces—and maybe homes arstechnica.com

Founders and researchers from Boston Dynamics, Agility, and Physical Intelligence describe how learned policies are replacing hand-coded behaviors, pushing robots toward general-purpose autonomy. Warehouses and factories come first; homes remain the hardest deployment because environments and tasks refuse to standardize.

References

MarkTechPost marktechpost.com

OpenAI Releases GPT-Live and GPT-Live-1-mini, full-duplex voice models that delegate deeper reasoning to GPT-5.5… the voice-native model snapshots conversation state and hands the task to GPT-5.5 in the background while continuing to talk.

VentureBeat venturebeat.com

GPT-Live can make interaction decisions multiple times per second… using transitional filler phrases like ‘let me just check that for you’ to hide the thinking time of the underlying engine, and hands off data to a visual UI displaying weather or stock cards as it speaks.

Hacker News discussion news.ycombinator.com

Users report ‘audio hallucinations’ where the model produces unintended sounds like shuffling papers, door knocking, background music, or intense breathing and groaning; VAD frequently stutters because it misidentifies its own vocal output as a user interruption.

Hacker News comment thread news.ycombinator.com

Many power users express a desire for a ‘Star Trek computer’ style — a dry, efficient utility — rather than the ‘creepy’ and ‘overly personable’ companion persona OpenAI continues to favor; the ‘AI girlfriend’ approach feels like a transparent engagement play.

UN report via People’s Daily Digital peopledaily.digital

A landmark July 2026 United Nations report characterized the rise of AI companions as an urgent public health challenge, warning of ‘sycophantic’ AI behavior that reinforces users’ distorted realities rather than challenging harmful thought patterns.

PCWorld pcworld.com

GPT-Live and Gemini Live both achieve first-audio-chunk delivery in the 300–600ms range, but Gemini 3.5 Live leads on ComplexFuncBench Audio with 90.8% success in multi-step function calling like booking a flight while checking a calendar.

fenado.ai — unsafe audit comparison fenado.ai

A formal audit pinned the number at 13,365 unsafe blocks… critics compared this to the high-performance Rust project uv, which maintains comparable complexity with only 73 unsafe blocks.

bun.com — Bun’s own unsafe audit / Phase B roadmap bun.com

~9,300 of these blocks (69%) can be refactored into safe Rust without performance loss or with minor redesigns; ~4% of the total Rust codebase is currently unsafe, concentrated in legacy Zig-style state machines (33.9%) and FFI boundaries (29.8%).

Hacker News discussion / Jarred Sumner quote via Kelley thread news.ycombinator.com

Bun’s rejected upstream patch stated: ‘We do not currently plan to upstream this, as Zig has a strict ban on LLM-authored contributions.’ Kelley called AI-assisted contributions ‘invariably garbage’ with ‘negative value.’

byteiota.com — post-merge bug report byteiota.com

A security hardening PR landed eight days after the merge to address 36 findings, including AI-generated fixes that accidentally introduced new use-after-free bugs while attempting to solve SSL leaks… bun upgrade displayed progress in raw bytes rather than mebibytes due to an incorrect integer formatter.

Anthropic acquisition announcement (Dec 2, 2025) anthropic.com

Anthropic acquired Oven, the startup behind Bun, coinciding with Claude Code reaching a $1B revenue run-rate. Because Claude Code ships as a Bun executable, ‘If Bun breaks, Claude Code breaks.’

getdx.com — Google LLM migration study getdx.com

Google reported that its LLM-assisted migrations reduced developer effort by 50%, with AI generating over 70% of the final code changes across 39 distinct large-scale projects.

Artificial Analysis artificialanalysis.ai

Nemotron 3 Ultra scores 47.7 on the Intelligence Index — the highest of any US open-weight model — but trails Chinese leader Kimi K2.6; its standout attribute is speed, exceeding 400 tokens/sec, roughly 5x faster than peers in its class.

LangChain blog (NemoClaw Deep Agents launch) langchain.com

Nemotron 3 Ultra reached a 0.86 aggregate score on the Deep Agents benchmark at $4.48 per run, versus the closest closed-source competitor at $43.48 — gains driven by harness tuning of tool descriptions, memory, and middleware rather than any weight update.

Towards AI — ‘Choosing Your Agent Harness’ pub.towardsai.net

Claude Agent SDK runs the agent inside the sandbox; Deep Agents keeps a pluggable backend so agents can run outside and control the sandbox remotely, swap virtual filesystems, and register harness profiles across 100+ model providers.

Apertis.ai model page apertis.ai

Fireworks lists Nemotron 3 Ultra at $0.60/$2.40 per 1M input/output tokens; Together AI charges $0.60/$3.60. Self-hosting the 550B/55B-active MoE realistically requires an 8×H100 cluster, which can erase the token-price advantage for smaller teams.

The Nuanced Perspective (Substack) — How to Choose Your AI Agent Stack thenuancedperspective.substack.com

The 2026 decision is no longer which model is smartest but which harness — sandbox topology, memory model, and tool-registration philosophy — best matches the workload; model-agnostic harnesses win where lock-in cost is high, first-party SDKs win where execution fidelity matters most.

Medium — Deep Agents from LangChain/LangGraph medium.com

Deep Agents are a LangGraph harness that decompose tasks with a write_todos tool, delegate to isolated sub-agents, and manage state via a virtual filesystem — moving beyond the think-act-observe loop toward Claude Code-style planning.

Sources

References

Jack Sun, writing.