On this day: Sep 29

A path has opened for agent runtimes to move from vendor products to foundation-owned standard infrastructure. The key is defining the harness, model and sandbox as swappable slots: if this separation takes hold, the cos

The first major commercial product to sell an agent not as "a session that runs when asked" but as "a persona resident in the account". The billing unit shifts from tokens and seats to the number of agents, and always-on

An approach that retrieves and adapts skills from repositories instead of learning them anew for each harness, turning skills from outputs tied to a model and harness into transferable assets. Read together with ClawScan

A mid-tier update released alongside Anthropic's Sonnet 5.5 in the same week. That two labs simultaneously touted "nearly the same work at 1/5 of the frontier price" shows the competitive axis of the second half of 2026

Disaggregated serving has come down from a research claim to an operational guide. The key is using p99 latency rather than average throughput as the yardstick, since agent loops are directly exposed to tail latency. It

The key is that speed has been split off into a separately billed tier. When latency becomes the bottleneck of agent loops, tokens per second become a product separate from quality, and investment in inference infrastruc

The first case of a competing lab putting numbers on what leaks out when open weights are released without safety measures. The debate in the open-weight lineage (axis ⑨) shifts from a "performance gap" to a "safeguard g

Context management becomes a built-in API feature

Names four patterns: compaction, structured note-taking, sub-agents and just-in-time retrieval

Puts fine-grained sparse attention into a commercial API and cuts long-context costs

Distributed training moves beyond pretraining demos into the post-training path of real product models

Passed one year after the SB 1047 veto, scaled back to focus on transparency. It went on to compel each lab to publish its framework

The strongest positive evidence that FP4 training actually works. However, it is the vendor's own report, and no independent reproduction has been confirmed

Anthropic said it observed the model maintaining focus for more than 30 hours on complex multi-step tasks; released under ASL-3 protections together with the Claude Agent SDK

The first setback for legislation regulating frontier models

The finding that initial tokens act as attention sinks; 2 years later gpt-oss adopted learnable sinks