AI Agents

Sep
11
Shopify banner showing React Native splitting into Swift and Kotlin, with Shop startup improvements of 23% on iOS and 50% on Android.

Shopify is leaving React Native after agents changed its cost model

Shopify rebuilt Shop in 12 weeks, backed by native specialists and strict parity checks. Maintaining two apps remains part of the bet.
3 min read
Sep
04
GPT-6 Astra news banner showing an agent workflow flagged and stopped with HTTP 403 misalignment_policy_violation, with no general resume path.

GPT-6 Astra, OpenAI's first Critical cyber model, can stop your API job without asking

OpenAI says GPT-6 Astra can find unknown vulnerabilities and develop exploits with the right tools and access. Its API can terminate flagged agent work with a 403.
4 min read
Sep
02
Claude Fable 5.1 banner showing cache reads falling from $1 to $0.25 per million tokens and three breaking changes involving forced tool use, thinking blocks, and edited context.

Claude Fable 5.1 makes caching cheaper and breaks three API behaviors

Cache reads now cost $0.25 per million tokens, but forced tool use and preserved thinking require code changes before some agent workloads can migrate.
4 min read
Aug
24
MCP roadmap banner contrasting July’s shipped scaling and authorization foundations with proposed identity, discovery and long-running-work features.

MCP’s new roadmap names the right gaps, but the fixes are still proposals

Stateless scaling and several authorization foundations shipped in July. Workload identity, progressive discovery and server-pushed events have no delivery dates, and the groups responsible for identity and discovery are still forming.
4 min read
Aug
13
News banner for Grok 4.6 shows xAI’s $2/$6-per-million input/output-token prices, while retries, tool calls, human review time, and total cost per accepted task remain unmeasured.

Grok 4.6 arrives with long-running coding agents in focus

xAI positions Grok 4.6 as a frontier competitor with $2/$6 token pricing. Repository-level tests will show whether that translates into cheaper completed tasks.
2 min read
Aug
10
Claude Code auto mode banner comparing Anthropic’s reported 89% dangerous-command detection with 13.6% for human review, while noting that the false-positive rate remains undisclosed.

Claude Code is automating approvals because humans stopped paying attention

Claude Code will default to auto mode as Anthropic argues that automated command review is safer than permission prompts developers stop reading.
3 min read
Aug
06
Meta Muse Code banner showing an append-only event log resuming after interruption, alongside Meta’s demo of more than 1,000 tool calls over runs lasting up to 24 hours.

Meta’s Muse Code bets that a coding agent needs durable state, not just a stronger model

Meta's Muse Code beta combines persistent background agents with an append-only event log, betting durable state matters more than another benchmark chart.
2 min read
Jul
31
OpenAI GPT-5.6 price comparison: Luna falls 80% from $1 to $0.20 per million input tokens, Terra falls 20% from $2.50 to $2, and Sol is unchanged.

OpenAI made GPT-5.6 Luna five times cheaper. That does not make every model cheaper

OpenAI cut GPT-5.6 Luna API prices by 80% and Terra by 20%, lowering costs for agent workloads while leaving Sol and subscription prices unchanged.
3 min read
Jul
16
Cursor logo beside a warning that a repository-root git.exe can run without a prompt on Windows, plus a seven-month disclosure timeline.

Cursor can run a repository's `git.exe`. It says the repo is the trust boundary

Mindgard reported Cursor's git.exe code execution flaw in December 2025. Seven months later, Cursor calls it out of scope and points to Workspace Trust.
4 min read
Jun
06
Anthropic warns AI could start improving itself, developers react

Anthropic warns AI could start improving itself, developers react

Anthropic says Claude now authors most of the code merged into its own codebase, raising a question about what happens when AI starts helping build the next AI.
3 min read