Anthropic pledges 5% of its shares to a public benefit fund that grows with the company

A public benefit fund whose stake scales with Anthropic's value

In an announcement dated October 18, Anthropic said it will commit a portion of its shares to the Anthropic Public Benefit Fund. Subject to stockholder approval, 5% of the shares are committed at the outset. The board may add 2% for every 50% increase in the company's value above a set baseline. In Anthropic's words, the more the company grows, the larger the share that can be dedicated expressly to the public good.

Anthropic stresses that most of its mission-driven work stays inside the company. That includes AI safety and interpretability research, applied biomedical research, and discounts on Claude for nonprofits and scientific institutions. It says these activities carry both public benefit and for-profit value, so they remain in-house. The Fund is meant for what the company structure does not easily allow, such as large donations to traditional nonprofits or direct distributions to specific populations. Labor market impacts are the example given. Anthropic notes it has already committed $200 million to studying and addressing the issue, and says that given the pace of AI progress, bolder approaches may be needed.

On structure, Anthropic says it chose a wholly-owned LLC over a 501(c)(3) or another format to keep the ability to adjust between forms of giving. It states that none of the Anthropic equity held by the LLC will be used for lobbying or political giving, in the US or abroad. It also contrasts the design with the common practice of offering escalating compensation awards to executives, saying it is directing that equity to the public benefit instead. It places the fund alongside earlier commitments: the co-founders' pledge to donate 80% of their equity, and an equity donation-match program in which approximately half of employees participate.

Sources: Investing in the public benefit — Anthropic

Anthropic on work, employment and expectations

The fund announcement names labor market instability as one of the risks it wants to address. Among the day's items are three Anthropic pieces whose headlines concern the same broad ground: "How employment is changing at Anthropic", "The AI expectations gap", and "Working at the frontier shapes expectations for the future".

The material available to Claude News for these three contains only the headlines. We therefore do not summarize their findings or draw conclusions from them, and readers should consult the originals.

Sources: How employment is changing at Anthropic — Anthropic (via Google News) · The AI expectations gap — Anthropic (via Google News) · Working at the frontier shapes expectations for the future — Anthropic (via Google News)

Technical work and evaluation, listed by title

Anthropic also published pieces headlined "Designing more efficient quantum circuits with Claude" and "Claude reveals ECDSA curve origins". It also posted "An update on independent evaluators". The first two point to Claude being applied to specialized technical problems, and the third to the outside assessment of Anthropic's models.

As with the previous section, the items supplied carry only titles, with no body text. We make no claims about what these pieces found, what methods they used or what they conclude.

Sources: Designing more efficient quantum circuits with Claude — Anthropic (via Google News) · Claude reveals ECDSA curve origins — Anthropic (via Google News) · An update on independent evaluators — Anthropic

Developer tooling: Claude Code releases and API changes

Claude Code 2.1.302 adds the CLAUDE_CODE_AUTO_MODE_MODEL variable, which sets the model used by the client-side auto mode classifier. It does not change the model for server-side auto mode. The release also adds an Idle compaction row to /config for turning idle compaction off or back on, and the notices shown after an idle compaction now say how to turn it off. /design-sync and /design-login now show a notice that standalone Claude Design (claude.ai/design) closes by December 14, 2026. Version 2.1.301 adds CLAUDE_CODE_RETRY_WATCHDOG_MAX_USAGE_LIMIT_WAIT_MS, so unattended retry mode ends a request at a usage limit that resets too far away instead of waiting. It also adds OpenTelemetry tracing guidance to the claude-api skill, including setting up trace destinations with the ant CLI, and the % bracket-matching motion in vim mode.

The API release notes of October 15 record three changes. Fast mode for Claude Opus 5.5 is out of research preview, so access no longer needs to be requested. Requests no longer require the fast-mode-2026-02-01 beta header, though those that still send it keep working. Fast mode for Opus 5.5 is also available on Claude in Microsoft Foundry for deployments hosted on Azure. Separately, the Compliance API now returns the text each organization member saved as Instructions for Claude in claude.ai and as Global instructions in Cowork. This is in beta for Claude Enterprise organizations, through a GET call on /v1/compliance/organizations/{org_uuid}/user_preferences.

Sources: Claude Code 2.1.302 — Claude Code (GitHub) · Claude Code 2.1.301 — Claude Code (GitHub) · Claude API release notes, October 15, 2026 — Anthropic (release notes)