Reading

California Governor Gavin Newsom signed seven bills on September 21 requiring data centers to pay for upgrades to local power grids and water systems.

The package requires the California Public Utilities Commission to create a new rate classification for data centers, and new projects must disclose estimated water use, energy efficiency and drought planning to local governments.

The actual cost impact cannot be quantified until the CPUC writes the rate rules. According to The Verge, Newsom said the laws are meant to stop data centers from passing utility costs onto residents.

The Canadian province of British Columbia sued OpenAI in a California court on September 21, alleging the company failed to report violent ChatGPT activity by the perpetrator of a school shooting.

The case relates to the February shooting in Tumbler Ridge that killed eight people. Attorney General Niki Sharma said the suit seeks accountability and change; OpenAI had previously apologised for not flagging the account.

The suit was just filed. The specific claims and any damages are not yet visible in court documents, and whether OpenAI owes any reporting duty will be for the court to decide.

Amazon has blocked Meta's Muse AI agent from making purchases on its store.

Starting Sunday night, September 20, users trying to buy goods through Muse received an error stating that continued access by an unauthorized AI agent violates Amazon's Conditions of Use. TechCrunch reports the block was first spotted by GeekWire.

Amazon is under no legal obligation to open its doors to Muse, and if the agent places a bad order, the platform handles the fallout. Muse has one of the lower hallucination rates among AI models, but it is still far from zero.

A new NBER conference paper estimates the US natural rate of unemployment has risen by about 0.1-0.2 percentage points since LLMs were introduced, with what the authors call considerable uncertainty.

The paper, by Hie Joo Ahn and Nicholas A. Carollo, combines CPS and JOLTS data with AI exposure measures from OpenAI and adoption measures from Lightcast. Workers with high AI exposure and adoption have seen larger declines in job-finding and job-switching rates than other groups, while within-job activity switching has increased noticeably.

The authors argue LLM-driven reallocation has operated mainly through within-firm task reorganization rather than mass layoffs. Note this is an unreviewed conference paper, and the magnitude estimate is sensitive to model specification.

Read from OpenAI's own charts, multi-agent parallelism mainly buys speed, not compute savings.

Writing on LessWrong, Toby Ord analyzed the GPT 5.6 launch-page charts: at equal performance, a 4-agent swarm uses roughly twice the total tokens of a single agent, and a 16-agent swarm roughly twice that again. His derived parallelizability parameter lambda falls between about 0.48 and 0.68, depending on the task.

That means scaling the swarm buys capability more expensively than lengthening a single agent's chain of thought; the payoff is speed, roughly 2x speed for 2x cost. Note this is a blogger's reading of vendor charts, with regressions run by Claude Opus 5, not an independent measurement.

On September 21, OpenAI announced a new independent Advisory Group on Mathematics and Artificial Intelligence, hosted at the Institute for Advanced Study in Princeton, with nine prominent mathematicians as initial members. It is meant to assess the significance of new results and coordinate their release.

Members are unpaid, can speak publicly and control their own membership. But the announcement states the group will not advise on how OpenAI paces its internal mathematics research, and IAS stressed that decision-making responsibility rests entirely with the company.

OpenAI also claims its internal model has resolved more than 100 open problems, a self-reported figure with no independent verification. The group follows an open letter by twenty-five Fields Medalists criticizing labs for racing to publish famous solutions; only one advisory member signed that letter.

Inspur announced the Yuannao SD200 Ultra supernode AI server at AICC2026 on September 21, built on domestic AI chips. The company says a single node can host the 2.8-trillion-parameter Kimi K3 model and supports frontier models up to 10 trillion parameters.

The system tightly couples 128 domestic AI chips with 8TB of unified-addressable memory and 64TB of system memory. Inspur claims token generation latency below 5.85ms, equivalent to 170 tokens/s per user and five times the industry average, plus a 3.5x reduction in AllReduce communication time.

Inspur also launched the HC2000 compute unit the same day, claiming 10x token throughput per unit of investment. Note that all performance figures are Inspur's own claims; the 'industry average' baseline is unspecified, and real-world results await third-party testing.

XPeng began pushing the new second-generation VLA release, XOS 6.3.0, on September 22, with single-Turing Max vehicles upgraded to the second-generation VLA in the same rollout.

The company says the release introduces the Infini-VLA long-horizon architecture, which it claims remembers the previous 30 seconds of driving context, and puts the X-Foresight predictive world model in a production car for the first time, claiming it can forecast surrounding road users' behavior up to 6 seconds ahead.

All capability descriptions are XPeng's own claims with no independent testing. For delivered Turing Max owners this is an OTA they can receive now; for anyone tracking the driver-assistance race, the numbers that matter will come from user testing and incident data, not launch copy.

Linear says its test suites nearly quadrupled this year, yet pull request CI wait fell from over 6 minutes to just over 5, with runner time per test roughly halved.

The work had two layers: moving to third-party runners with faster CPUs and better caching made like-for-like jobs 34% faster on average, and switching to the native TypeScript compiler cut median typecheck time by 73%. The other layer saves machine time: linting without type information, trimming small jobs off the critical path, and cutting per-shard setup by roughly 44%.

All figures are Linear's own measurements with no third-party verification, and the findings come from one TypeScript monorepo. Still, the claim that AI coding has made CI the bottleneck, plus the concrete optimization list, is directly useful to engineering teams facing the same surge in agent-submitted code.

NVIDIA launched DSX Ready on September 21, a qualification program that labels power and cooling products as fitting its AI factory reference designs.

Two categories launch first: battery energy storage systems, with Hitachi Energy, LG Energy Solution and Tesla qualified, and cooling distribution units, with LG Electronics, LiquidStack and Vertiv qualified. More categories will follow.

Note the limits: CDUs go through a self-qualification suite where partners run the tests themselves and submit data for NVIDIA review, and NVIDIA states that passing does not replace site-level engineering or imply site-level stability.

AWS released the open-source Strands Harness on September 21, an agent framework that runs locally or on any cloud, including Google Cloud, Azure and Cloudflare.

It ships with read, write, edit, shell and web search tools, manages its own context window, and keeps memory across sessions via session IDs. It can run on Anthropic, OpenAI, Amazon Bedrock and Google models, or a local Ollama model, and installs via pip or npm.

AWS says the agent is 26% more efficient than agents built on other frameworks, and cost 77% less than Claude Code on the same tasks using Anthropic's Fable 5 model. These are AWS's own first-party benchmarks with no independent reproduction yet.